1(self.webpackChunk_N_E=self.webpackChunk_N_E||[]).push([[3831],{1065:(e,t,n)=>{"use strict";n.d(t,{x:()=>s});var a=n(95155),i=n(12115),r=n(53999);function s(e){let{children:t,speed:n=.5,direction:s="up",className:o,disabled:l=!1}=e,d=function(){let{speed:e=.5,direction:t="up",disabled:n=!1}=arguments.length>0&&void 0!==arguments[0]?arguments[0]:{},a=(0,i.useRef)(null);return(0,i.useEffect)(()=>{if(n||!a.current)return;let i=a.current,r=!1,s=()=>{if(window.innerWidth<768){i.style.transform="translate3d(0, 0px, 0)";return}i.getBoundingClientRect();let n=window.pageYOffset*e,a="up"===t?n:-n;i.style.transform="translate3d(0, ".concat(a,"px, 0)"),r=!1},o=()=>{r||(requestAnimationFrame(s),r=!0)},l=()=>{s()};return window.addEventListener("scroll",o,{passive:!0}),window.addEventListener("resize",l,{passive:!0}),s(),()=>{window.removeEventListener("scroll",o),window.removeEventListener("resize",l)}},[e,t,n]),a}({speed:n,direction:s,disabled:l});return(0,a.jsx)("div",{ref:d,className:(0,r.cn)("will-change-transform",o),children:t})}},24506:(e,t,n)=>{"use strict";n.d(t,{ue:()=>r,zX:()=>i});let a=[{slug:"datasheet-havoc-engineering-data-extraction",title:"Datasheet Havoc: Why Engineering Data Extraction Is Still Manual â and What AI Changes",excerpt:"Across process plants and capital projects, vast amounts of critical engineering data live inside datasheets. Here is why extraction remains manual, and how modern AI changes the equation.",date:"August 13, 2026",readTime:"6 min read",author:"Mohammed Sami Ahmed",category:"Engineering Document Intelligence",keywords:["engineering data extraction","datasheet extraction","process engineering AI","CFIHOS","P&ID data extraction","tag list extraction","engineering document intelligence","HAZOP packages"],image:"/Images/datasheet-havoc-cover.jpg",gradient:"from-blue-600 to-indigo-600",featured:!0,content:"\nAcross process plants and capital projects, a huge amount of engineering information lives inside data-sheets.\n\nDesign pressure. Operating temperature. Materials of construction. Pump duty. Motor rating. Instrument ranges. Valve specifications. Equipment tags. Vendor model numbers.\n\nAnd very often, somebody has to read the datasheet and manually transfer the information into:\n\n- Equipment registers\n- Instrument indexes\n- Valve lists\n- Handover databases\n- CMMS or EAM systems\n- HAZOP packages\n- Engineering reviews\n\nA single datasheet may only take a few minutes to review.\n\nBut across hundreds or thousands of equipment items, instruments and vendor documents, that becomes a substantial engineering workload, in addition to misses due to human error and fatigue.\n\n---\n\n## Why Isn't the Information Already Structured?\n\nIn theory, it often starts that way. An owner or EPC issues a standard datasheet template or information requirement, and the vendor fills it in.\n\nThe problem is what happens next.\n\nAs the document moves through the supply chain:\n\n- Vendors may use their **own version of the template**\n- Multiple **vendor-issued revisions** are created over time\n- Documents are converted to PDF, and sometimes scanned\n- Rows, fields, or sections are added, removed, or rearranged\n- Units and terminology vary between suppliers\n- The same information may appear in more than one place\n\nBy the time the datasheet comes back, the engineering meaning is usually still there, but the original structure may no longer be consistent enough for software to reliably read it field-by-field.\n\n\n\nThis is one reason the process industry has invested heavily in standardising information exchange. CFIHOS, for example, was created to improve handover between operators, contractors, equipment manufacturers, and suppliers, covering both engineering data and documents so information can be found, exchanged, and reused more consistently across the facility lifecycle. [1][2]\n\n---\n\n## Industry Has Been Trying to Solve This for Years\n\nThe desired end state is clear:\n\n- A supplier receives defined information requirements.\n- The supplier returns structured information against those requirements.\n- That information moves cleanly into the owner's engineering, maintenance and asset-information systems.\
1n\nStandards such as **CFIHOS** help define the common language needed for that exchange, including equipment classes, properties, units and relationships. [3]\n\nThat is important progress.\n\nBut real projects still contain a mixture of structured data, spreadsheets, PDFs, scans and historical files from many different suppliers and systems.\n\nSo even when the destination is structured, somebody still has to bridge the gap between the document and the database.\n\n---\n\n## Traditional Automation Helped â But Mainly Where Documents Were Predictable\n\nEarlier document-extraction systems generally relied on templates, fixed positions, text recognition and rules.\n\nIf every vendor returned exactly the same datasheet layout, this could work very well.\n\nFor example:\n> *âDesign Pressure is always in this box on page 2.â*\n\nThe difficulty starts when the next vendor moves the field, changes the table, uses a different unit, adds a page or sends a scan.\n\nThen the extraction logic needs to adapt again.\n\nThis made automation useful for tightly controlled document sets, but much harder across the variety of documents found on real industrial projects.\n\nThe persistence of the problem is visible even in current engineering-information platforms.\n\nHexagon, for example, describes using AI in HxGN SDx2 to automate tag extraction from varied engineering sources and match those tags against existing databases, specifically to reduce manual work and errors. [4]\n\n---\n\n## The Problem Is Bigger Than Simply Reading Text\n\nSuppose a system successfully reads:\n\n```\nDesign Pressure: 15 barg\n```\n\nThat is useful. But engineering data extraction has another layer. What if another vendor writes:\n\n```\n15 bar(g) or 1.5 MPa(g)\n```\n\nThe system needs to understand that these represent equivalent engineering information.\n\nThat is partly a normalisation problem â exactly the kind of issue standards such as CFIHOS address through common classes, properties and units [3]. In practice, CFIHOS adoption is inconsistent, and **it offers little help with the vast installed base of existing operating assets whose documentation predates** the standard entirely and was never going to be retrofitted into it.\n\nBut then there is an even harder problem.\n\nWhat if the latest P&ID shows a different equipment tag than the datasheet? What if revision C changed a value but the equipment register still shows revision B? What if two pages of the same vendor package disagree with each other?\n\nFor a solution to be useful in building a **tag list, preparing a HAZOP, troubleshooting a live plant, or working through an MOC**, the requirement is no longer simply:\n\n> *âCan software read this PDF?â*\n\nIt becomes:\n\n> *âCan software understand the engineering information well enough to know what it means and where it came from?â*\n\n---\n\n## This Is Where Modern AI Becomes Interesting\n\nModern multimodal AI can interpret documents much more flexibly than traditional fixed-template extraction.\n\nInstead of relying entirely on a field appearing at a predefined coordinate, the system can use the surrounding text, table structure, document, and project context to determine what a value represents.\n\nThat makes it possible to work across changing layouts and terminology without creating a new extraction rule for every document format.\n\n**But replacing OCR with an AI model does not automatically make the information trustworthy.**\n\nFor engineering use, a useful system should be able to answer:\n\n- **What did you extract?**\n- **Where exactly did you find it?**\n- **How did you interpret the unit or terminology?**\n- **Does anything elsewhere in the project appear to contradict it?**\n- **What could you not determine confidently enough to populate automatically?**\n\nThat changes the workflow.\n\nInstead of an engineer manually reading every field of every datasheet, AI can perform the first pass.\n\nThe engineer reviews the values that need judgment: low-confidence fields, conflicts, missing information, and unusual cases.\n\n**The objective is not to remove engineering review.**\n\n**It is to make that review far more focused.**\n\n---\n\n## The Bigger Opportunity Is Source Traceability and Reuse\n\nThere is another important implication.\n\nOnce a datasheet has been understood and validated, why should that engineering information be extracted again for the next task?\n\nThe same equipment data may later support:\n\n- An asset register\n- Vendor-document review\n- P&ID checks\n- HAZOP preparation\n- Modification reviews\n- Handover\n- Troubleshooting in an operating plant\n\n\n\nThe extraction itself becomes infrastru
1cture for engineering work, done by AI reviewed by Humans.\n\nThat is where the value can extend well beyond the original document-processing task.\n\nAt [Novek AI](https://novek.ai), this is the problem we are working on.\n\nNot simply pulling text from engineering documents.\n\nThe goal is to turn those documents into engineering information that is:\n\n**Source-traceable, reviewable, correctable, and reusable.**\n\n- If it extracts a design pressure, pump duty, or instrument range, the engineer should be able to point exactly where that value came from throughout the lifecycle.\n- If two documents disagree, the system should surface the conflict rather than silently choosing one.\n- If it cannot determine something reliably, it should ask for review rather than guess.\n- And once that information has been verified, the same understanding can be reused across the workflows that depend on it.\n\nThe opportunity is not simply to make document reading more automated.\n\nFor many engineering teams, the more immediate opportunity is much simpler:\n\n**Stop making engineers repeatedly read and re-type information that already exists inside their documents.**\n\nThen make that extracted information trustworthy and traceable enough to reuse.\n\n---\n\n### References\n\n1. **[CFIHOS â About](https://www.jip36-cfihos.org/about/)** â CFIHOS describes its purpose as providing practical, standardised information-handover specifications for industrial facilities across operators, contractors, equipment manufacturers and suppliers. It explicitly organises both data and documents in a structured way.\n2. **[IOGP â CFIHOS Explained](https://www.iogp.org/blog/standards/cfihos-explained/)** â IOGP explains that CFIHOS is intended to make information handover quicker, easier and safer across operators, contractors, equipment manufacturers and suppliers.\n3. **[CFIHOS â Reference Data Library Development Guide](https://www.jip36-cfihos.org/wp-content/uploads/2022/08/CFIHOS-RDL-Development-Guide.pdf)** â The CFIHOS RDL provides the common reference structure used to standardise engineering information across the supply chain, including equipment-related data definitions.\n4. **[Hexagon â The Role of Artificial Intelligence in HxGN SDx2](https://aliresources.hexagon.com/smart-digital-reality/ai-driven-asset-management-transforming-operations-with-hxgn-sdx2)** â Hexagon describes using AI to extract tags from varied engineering sources and match them against existing databases, reducing manual work and errors.\n"},{slug:"why-ai-is-failing-process-industry",title:"Why AI Is Failing to Deliver ROI in the Process Industry â and What the Path Forward Actually Looks Like",excerpt:"The process industry is spending billions on operational AI. ROI remains elusive. The reason isn't the technology â it's that AI is being applied to the wrong layer. Here is what the path forward actually looks like.",date:"July 9, 2026",readTime:"10 min read",author:"Novek AI",category:"Engineering Document Intelligence",keywords:["EPC AI ROI","process industry AI","engineering document intelligence","P&ID extraction","RFI delays EPC","AI manufacturing ROI","engineering automation EPC","document AI process industry"],image:"/Images/Chat_bot_illustration.png",gradient:"from-blue-600 to-cyan-500",featured:!0,content:"\n## The Investment Surge and the Accountability Gap\n\nAcross manufacturing, energy, and construction, AI adoption has accelerated sharply over the last three years. Boards are applying pressure. Budgets are being allocated. Pilots are launching across procurement, operations, quality control, and predictive maintenance.\n\nYet the 2026 Grant Thornton AI Impact Survey of 950 senior business leaders across ten industries reveals a widening accountability gap. Despite significant AI investment, the majority of organizations cannot demonstrate that their AI systems are working, cannot measure the impact of individual initiatives, and are not prepared to respond when something fails.\n\n- **78%** of business executives lack full confidence they could pass an independent AI governance audit within 90 days\n- **22%** of operations leaders report having a fully developed and implemented AI strategy â despite AI already running in their operations\n\nThe survey's conclusion is pointed: organizations are expanding AI across more pilots, use cases, and functions â but without consistent measurement, feedback loops, or clarity on where value is created. In manufacturing specifically, companies are responding to board pressure by deploying AI solutions before choosing use cases â a pattern Grant Thornton describes as \"putting a square peg in a round hole.\"\n\n*Source: [Grant Thornton AI Impact Survey, 2026 â n=950](https://www.grantthornton.com/insights/survey-reports/2026/ai-impact-survey)*\n\n---\n\n## Why the Process Industry Is a Harder Case Than Most\n\nThe generic AI adoption challenges above are acute across all industries. In the process and EPC industry, they are compounded by a structural characteristic that most AI deployment frameworks do n
1ot account for: **a single project is never owned, designed, and executed by a single organization.**\n\nA refinery expansion, a petrochemical plant, an LNG terminal â each involves an asset owner who commissions the project, an engineering contractor who designs it, procurement firms who source equipment across global supply chains, and field construction contractors who build it. Within each of those layers, there are further subcontractors, specialist vendors, and discipline engineers, each working from their own view of the same underlying project documentation.\n\nThis fragmentation has a specific consequence for AI strategy: **you cannot impose a unified data platform across organizations that do not share ownership, systems, or incentives.** The push toward a fully data-driven, platform-integrated EPC industry is real â but it is a decade-long infrastructure transition, not a deployment decision.\n\nWhat exists today, in every project, at every organizational boundary, is documents. Piping and instrumentation diagrams. Equipment datasheets. Instrument specifications. Vendor drawings. Inspection reports. These documents are the connective tissue across the entire project lifecycle â from FEED through commissioning â and they are almost entirely read by humans, manually, one at a time.\n\n---\n\n## The Document Layer: Where Every Task Starts\n\nEvery substantive task in an EPC project begins with an engineer reading a document. Not navigating a digital twin. Not querying a structured database. Reading a document â often a PDF, often a scanned drawing, often one created by a different organization using different conventions and symbology.\n\nThis is not a temporary condition that will resolve as digitization matures. It reflects the contractual, legal, and multi-organizational reality of how capital projects are delivered. Documents are the audit trail. Documents are the handoff mechanism. Documents are what every organization can send to every other organization regardless of which software stack they use.\n\nThe research bears this out. Knowledge workers across industries spend an average of **9.3 hours per week** searching and gathering information â equivalent to hiring five employees and having one spend the entire week looking for answers instead of working.\n\n*Source: [McKinsey Global Institute, The Social Economy, 2012](https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-social-economy)*\n\nFor engineers working from multi-organization, multi-revision document sets, this figure is likely conservative.\n\n### The RFI as the clearest expression of the document problem\n\n| Metric | Figure |\n|---|---|\n| RFIs per $1M of project value | 9.9 |\n| Average days to close an RFI | 9.7 |\n| Cost per RFI to process | ~$1,080 |\n| RFIs that receive no response | 21.9% |\n\n*Source: [Navigant Construction Forum â Impact and Control of RFIs on Construction Projects, 2013, n=1,362 projects, 1M+ RFIs](https://www.procore.com/library/rfi-construction)*\n\nA Request for Information is, almost always, a question that could be answered by someone reading the right document â but the person asking either cannot find it, cannot interpret it c
1onfidently, or cannot do so quickly enough to avoid disrupting the field schedule. A crew standing idle for three days while an RFI works its way through the system represents lost productivity, potential out-of-sequence work, and material substitutions made under deadline pressure that have to be revisited later.\n\n---\n\n## The Chain Reaction From Faster Document Reading\n\nWhat happens when an engineer can find the right answer in 30 seconds instead of 30 minutes â with a direct link back to the exact region of the drawing or specification it came from?\n\n**RFI cycle time drops.** A question that previously required formal submission, routing, and a 7â10 day wait gets answered at the point of need. Idle crew time is eliminated or dramatically reduced.\n\n**Deadline-driven shortcuts decrease.** Engineers who cannot find the right specification under time pressure make judgment calls. Better information access under the same deadlines means fewer field decisions made without documentation.\n\n**Safety and reliability improve.** Many near-misses and reliability issues in process plants trace back to a decision made with incomplete information. Reducing the friction of finding the correct specification reduces the frequency of this class of error.\n\n**Auditability improves without additional effort.** When every answer is traceable to a source document and region, the audit trail is a byproduct of the workflow rather than a separate documentation burden.\n\n**The efficiency gain compounds across every organization on the project.** Because documents are the universal handoff mechanism across organizational boundaries, improving document readability benefits every party â owner, engineer, contractor, and subcontractor â without requiring any of them to change their systems.\n\n---\n\n## Why This Approach Does Not Require Trusting AI to Make Decisions\n\nThe approach described here is not AI autonomy. It is AI-assisted search with human verification â using AI the way a highly capable analyst who has read every document on the project would be used: to surface the relevant passage, identify the relevant drawing region, and hand it to the engineer for a decision.\n\nThis matters for adoption in a regulatory and safety-critical industry. The process industry does not need AI to be right by itself. It needs AI to reduce the time it takes a qualified engineer to be right. Every extracted answer should point back to its source â the exact drawing, the exact revision, the exact region â so that the engineer can confirm it in seconds rather than re-derive it from scratch.\n\n**Traceability is not a feature in this model. It is the mechanism by which the entire workflow remains human-accountable** â which is the precondition for adoption in any environment where a wrong answer has safety or legal consequences.\n\n---\n\n## The Technology Precondition That Now Exists\n\nTemplate-based and rule-based document extraction systems for engineering drawings have been attempted repeatedly since the early 2000s. They worked in controlled conditions and failed in production: custom symbol sets, varying drawing conventions, scan quality, and project-specific tag formats each required re-training or re-templating â and a wrong extraction still required a human to re-derive the answer from scratch.\n\nThe shift that makes the current moment different is not incremental improvement in object detection. It is the availability of models that can **reason about an unfamiliar symbol from context** â inferring relationships rather than only matching memorized patterns:\n\n- Which line connects to which equipment\n- Whether an instrument reads the process line or carries only a signal\n- Whether nearby text is a tag, a size, a specification class, or an unrelated label\n\nThis capability generalizes across drawing conventions and symbol sets in a way that template-matching cannot. Combined with source traceability â pointing every extracted fact back to the exact region of the drawing it came from â it addresses both of the failure modes that stalled earlier systems: generalization and correction cost.\n\n---\n\n## What Novek-AI Is Building\n\nNovek-AI is building relationship-aware extraction for engineering drawings â P&IDs, datasheets, and instrument specifications â with full source traceability. Our extraction model reasons about relationships between entities on a drawing, not just the entities themselves, and every output is linked back to its exact source location so that engineering review is a confirmation rather than a re-derivation.\n\nWe are publishing our accuracy results against ground truth, segmented by industry, scan quality, drawing density, and entity type â including where we are weaker â at [novek.ai/accuracy](https://novek.ai/accuracy).\n\n---\n\n## References\n\n1. Grant Thornton, *AI Impact Survey 2026*, n=950 senior business leaders across 10 industries. [grantthornton.com/insights/survey-reports/2026/ai-impact-survey](https://www.grantthornton.com/insights/survey-reports/2026/ai-impact-survey)\n2. McKinsey Global Institute, *The Social Economy: Unlocking value and productivity through social technologies*, 2012. [mckinsey.com](https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-social-economy)\n3. Navigant Construction Forum, *Impact and Control of RFIs on Construction Projects*, 2013. Study of 1,362 projects containing over 1 million RFIs. Note: cost figures are from 2013 and likely conservative by today's standards. [procore.com/library/rfi-construction](https://www.procore.com/library/rfi-construction)\n "},{slug:"ai-instrument-list-extraction-guide",title:"AI Instrument List Extraction: How Engineering Teams Are Cutting Instrument Index Time by 80%",excerpt:"Manual instrument list extraction from P&IDs takes days. Here is how AI instrument list extraction works, what it produces, and why EPC teams are adopting it.",date:"April 12, 2026",readTime:"9 min read",author:"Novek AI",category:"AI Instrument List Extraction",keywords:["AI instrument list extraction","instrument list extraction","instrument index from P&ID","instrument list automation","EPC instrument register","AI engineering document extraction"],image:"/Images/Chat_bot_illustration.png",gradient:"from-violet-500 to-blue-600",featured:!0,content:"\n## What Is AI Instrument List Extraction?\n\nAn instrument list â also called an instrument index or instrument register â is one of the core deliverables EPC firms produce from a P&ID package. It captures every instrument in the design: its tag, service description, type, line number,
1connected equipment, and datasheet reference.\n\nManually, this means reading every P&ID drawing, identifying instrumentation symbols, reading the nearby text, and keying data into a spreadsheet row by row. On a package with 80â100 drawings, that is a week of engineering time â for a single output.\n\n**AI instrument list extraction** replaces that manual reading cycle with an automated pipeline that reads drawing notation, identifies instrument symbols, extracts associated attributes, and structures them into a verified, source-linked instrument index.\n\n---\n\n## What an Instrument List Contains (And Why Each Field Is Hard to Extract Manually)\n\n| Field | Why It Is Hard Manually |\n|---|---|\n| Tag number (e.g. FT-1012) | Has to be read from a drawing callout adjacent to a specific symbol |\n| Instrument type | Inferred from the ISA-5.1 symbol shape, not always spelled out |\n| Service description | Often abbreviated in the callout, requires interpretation |\n| Line / equipment reference | Linked to nearby piping notation â easy to miss |\n| P&ID reference | Has to be tracked per-row across dozens of drawings |\n| Datasheet reference | Cross-document lookup â manual cross-referencing is error-prone |\n\nEach field multiplied across 400â600 instruments is where manual effort compounds into days of work.\n\n---\n\n## How AI Instrument List Extraction Works in Novek\n\nNovek's approach to AI instrument list extraction is visual-first, not OCR-first.\n\n**Step 1 â Drawing ingestion**\nThe P&ID package (typically PDF) is loaded as a drawing set, not converted to plain text. Each drawing is processed page by page.\n\n**Step 2 â Symbol recognition**\nA visual recognition layer identifies ISA-5.1 instrument symbols by shape and position. Flow transmitters, pressure indicators, control valves, and relays are identified individually.\n\n**Step 3 â Tag and attribute parsing**\nAdjacent text is parsed together with the symbol â tag number, service notes, line references, and class designations are extracted as linked attributes, not floating text strings.\n\n**Step 4 â Cross-drawing consolidation**\nInstruments that appear on multiple drawings (e.g. on a P&ID and a loop diagram) are deduplicated and their references merged.\n\n**Step 5 â Source-linked output**\nEvery row in the instrument list is linked to the drawing page and coordinate it came from. Engineers can click any tag and see the source.\n\n---\n\n## What the Output Looks Like\n\nThe instrument list Novek produces maps directly to what an EPC engineer would build manually:\n\n- Tag number\n- Description / service\n- Instrument type (from ISA classification)\n- Process connection / line number \n- Connected equipment tag\n- P&ID sheet reference\n- Datasheet reference (where available)\n\nThe output can be exported to Excel in the client's project template, not a generic format that needs reformatting downstream.\n\n---\n\n## Where AI Instrument List Extraction Fits in the Wider EPC Workflow\n\nInstrument list extraction rarely happens in isolation. It typically feeds:\n\n- **Instrument datasheets** â the instrument list defines what datasheets need to be produced or collected\n- **Instrument index** â a higher-level register referencing the full instrument list\n- **Hook-up drawings** â the instrument list informs how instruments connect to process piping\n- **Cost estimates** â instrument counts directly feed procurement budgets\n\nWhen the instrument list is produced fast and accurately, every downstream step moves earlier. When it is slow or error-prone, everything waiting on it slips.\n\n---\n\n## Common Questions From Engineering Teams\n\n**Can AI extract all instrument types, including specialty instruments?**\nNovek extracts from ISA-5.1 notation, which covers the full instrumentation vocabulary. Specialty instruments that deviate from standard symbology can be flagged for manual review rather than silently dropped.\n\n**What happens with instruments that span multiple P&ID sheets?**\nCross-drawing references are captured automatically. The instrument list entry includes all drawing references where that tag appears.\n\n**How is the output checked for accuracy?**\nEvery extracted row stays linked to its source drawing location. Engineers can sample-check by clicking a row and seeing the exact location on the source P&ID page â no manual cross-referencing required.\n\n**Can the instrument list be exported in our project template?**\nYes. Output format follows the client's column structure, not a generic export that requires reformatting.\n\n---\n\n## Why This Matters for EPC Delivery Timelines\n\nInstrument list extraction is early in the project deliverables lifecycle. When it happens manually, the team spends the first weeks of a project doing largely administrative data collection before any engineering value-ad
1d begins.\n\nAI instrument list extraction compresses that window significantly. That frees engineering time for analysis, review, and client coordination â the work that actually requires engineering judgement.\n\nFor teams running multiple concurrent projects, the compounding effect is significant: the same headcount can service more packages per quarter without additional resource.\n "},{slug:"ai-tag-list-extraction-epc",title:"AI Tag List Extraction for EPC Firms: What It Is, How It Works, and What Teams Get Out of It",excerpt:"Tag list extraction is one of the most time-consuming early-stage EPC tasks. Here is how AI tag list extraction works, what a correct tag list requires, and why engineering teams are switching.",date:"April 10, 2026",readTime:"8 min read",author:"Novek AI",category:"AI Tag List Extraction",keywords:["AI tag list extraction","tag list extraction","tag register from P&ID","EPC tag register automation","P&ID tag extraction","engineering tag list"],image:"/Images/Properties_Illustration.png",gradient:"from-blue-500 to-cyan-600",featured:!0,content:"\n## What Is a Tag List in Engineering?\n\nA tag list â also called a tag register or equipment register â is a structured inventory of every tagged item in a design: instruments, valves, equipment, piping lines, and other engineered items. Each tag is a unique identifier that links the physical item to its P&ID location, datasheet, purchase order, and maintenance record.\n\nOn any EPC project, the tag list is foundational. Everything from procurement to commissioning traces back to it. A wrong tag, a missed tag, or a duplicate tag creates rework that propagates through the entire project.\n\n---\n\n## Why Manual Tag List Extraction Is Slow and Error-Prone\n\nThe conventional way to build a tag list is to read each P&ID drawing, identify every tagged item, and key it into a spreadsheet.\n\nThe problem is not that this is hard â it is that it is slow, repetitive, and the kind of work where human error is almost inevitable:\n\n- Tags are embedded in drawing callouts that require careful reading\n- The same tag may appear on multiple drawings (a process P&ID and a utility P&ID, for example)\n- Abbreviations, drawing conventions, and revision states vary across the package\n- Someone still has to cross-check the final list against the drawing set\n\nOn a 100-drawing package, a thorough manual tag list takes a senior engineer several days. On a larger package, it is weeks. And any document revision restarts part of that process.\n\n---\n\n## What AI Tag List Extraction Changes\n\n**AI tag list extraction** automates the reading and extraction step so that the engineering team receives a structured, source-linked tag list without the manual reading cycle.\n\nNovek's extraction pipeline:\n\n1. Reads each P&ID drawing visually â symbols, callouts, line designations, and equipment references\n2. Identifies tagged items by type: instruments (ISA-5.1), valves, equipment, piping lines\n3. Extracts the tag number, description, referenced line, connected equipment, and drawing location\n4. Deduplicates tags that appear across multiple drawings and merges their references\n5. Produces a structured tag list with every row linked to its source drawing page\n\nThe result is a tag list that would have taken days to build manually â completed in minutes, with every entry traceable.\n\n---\n\n## What a Correct Tag List Requires\n\nSpeed alone is not enough. A usable tag list has to be:\n\n**Complete** â no tags missed across the full package\n\n**Accurate** â tag numbers, types, and references correctly read from the drawing\n\n**Deduplicated** â the same tag appearing on multiple drawings appears once in the list, with all drawing references captured\n\n**Source-linked** â every row should be verifiable against the source drawing without re-reading the entire document\n\n**Template-consistent** â the output should map to the project's tag register format, not a generic export\n\nNovek's AI tag list extraction is designed to meet all five. Engineers can open any row, click the source reference, and verify the tag against the exact P&ID location.\n\n---\n\n## Tag List vs. Tag Register vs. Equipment List: What Is the Difference?\n\nThese terms are often used interchangeably on projects but have distinct meanings:\n\n| Term | Typically contains |\n|---|---|\n| Tag list | All tagged items across all disciplines |\n| Tag register | A curated, controlled version of the tag list, often maintained as a living document |\n| Equipment list | A subset â only major equipment (vessels, pumps, c
1ompressors) |\n| Instrument list | A subset â only instrumentation |\n\nNovek can produce any of these as standalone outputs or as subsets of the full tag extraction run.\n\n---\n\n## What Teams Actually Get Out of AI Tag List Extraction\n\n**Faster project starts** â tag lists are typically needed early for procurement and datasheets. Getting them in hours instead of days accelerates everything that depends on them.\n\n**Fewer spreadsheet errors** â the extraction is systematic. The same reading logic is applied to every drawing, every time.\n\n**Faster QA** â source links mean engineers can sample-check the extracted output in minutes rather than re-reading drawings.\n\n**More packages per team** â when tag extraction takes minutes instead of days, the same team can service more concurrent projects without adding headcount.\n\n---\n\n## Frequently Asked Questions\n\n**Does AI tag list extraction work on scanned P&IDs?**\nNovek works with standard PDF-format engineering drawings. Heavily degraded or low-resolution scans may affect accuracy, but standard project PDFs work well.\n\n**What happens when a drawing revision comes in?**\nRevisions can be re-run through the extraction pipeline. Changed tags, added tags, and removed tags are flagged as differences against the previous extraction.\n\n**Can we customise the output columns?**\nYes. The tag list output maps to the project's template â column names, field order, and attribute selection are configurable per project.\n\n**How long does the extraction actually take?**\nFor a typical 80â100 drawing package, extraction runs in minutes. Review and spot-check by the engineering team typically takes an additional 1â2 hours for a first run.\n "},{slug:"ai-line-list-extraction-from-pids",title:"AI Line List Extraction from P&IDs: Replacing the Spreadsheet with a Traceable Workflow",excerpt:"Building a line list manually from a P&ID package is weeks of engineering time. This post explains how AI line list extraction works and what it produces.",date:"April 7, 2026",readTime:"10 min read",author:"Novek AI",category:"AI Line List Extraction",keywords:["AI line list extraction","line list extraction","piping line list from P&ID","line list generation","line list automation EPC","piping line schedule extraction"],image:"/Images/Properties_Illustration2.png",gradient:"from-emerald-500 to-teal-600",featured:!0,content:"\n## What Is a Line List and Why Does It Take So Long?\n\nA piping line list is a structured register of every piping line in a design. Each row captures the line number, service, fluid, design pressure, design temperature, pipe class, insulation, and connected equipment. It is one of the most universally required deliverables on EPC and FEED projects.\n\nThe problem is where the data lives. Line information is distributed across P&ID callouts, title blocks, notes, and associated documents â spread across every drawing in a package. There is no single place to read it. An engineer has to work through the drawing set one line at a time, reading notation, interpreting abbreviations, and keying values into a spreadsheet.\n\nOn a 100-drawing package, that is 400â800 lines. At even conservative rates, building the line list manually means 2â4 weeks of engineering time before the team can move to the next deliverable.\n\n---\n\n## How AI Line List Extraction Works\n\n**AI line list extraction** automates the reading-and-extraction step. Rather than engineers reading drawing notation line by line, an AI pipeline reads the drawings and structures the output.\n\nHere is how Novek's pipeline works:\n\n**1. Drawing ingestion**\nThe P&ID set is loaded as a drawing package, not converted to plain text. The visual layout of each drawing is preserved.\n\n**2. Line designation recognition**\nPiping line designations follow a structured notation: line number, service abbreviation, nominal diameter, pipe class, insulation class. Novek reads this notation and parses each component.\n\n**3. Attribute extraction**\nFor each identified line, Novek extracts the associated attributes: service description, design conditions (where stated), pipe class, insulation spec, and from/to connections.\n\n**4. Cross-drawing consolidation**\nThe same piping line often appears on multiple P&ID sheets (continuing lines). Novek consolidates these appearances so the final line list has one row per line, with all drawing references captured.\n\n**5. Source-linked output**\nEvery row in the exported line list links back to the drawing and location it came from. Engineers can verify any row against its source without re-reading the drawing.\n\n---\n\n## What Fields Does the Line List Contain?\n\nA typical AI-extracted line list from Novek contains:\n\n| Field | Source |\n|---|---|\n| Line number | P&ID line designation callout |\n| Service description | Abbreviation expanded from the line designation |\n| Nominal diameter | Extracted from line designation |\n| Pipe class | Extracted from line designation |\n| Design pressure | From drawing notes or connected equipment data |\n| Design temperature | From drawing notes or connected equipment data |\n| Insulation class | From line designation or separate insulation spec |\n| From / To | From and to connections shown on the P&ID |\n| P&ID sheet reference | All sheets where the line appears |\n\nOutput format maps to the project's piping line list template â not a generic export that needs reformatting.\n\n---\n\n## Where the Line List Fits in an EPC Workflow\n\nThe line list is an early deliverable that many downstream activities depen
1d on:\n\n- **Stress analysis** â pipe stress engineers need the line list to scope and prioritise their work\n- **Piping material take-off (MTO)** â the MTO extends the line list with quantities\n- **Procurement** â pipe classes and line counts feed procurement schedules\n- **Hydrostatic test packages** â test boundaries are defined against the line list\n- **As-built documentation** â the finished line list is part of the as-built record\n\nWhen line list extraction is slow, all of these activities wait. When it is fast and accurate, teams can move faster across the project.\n\n---\n\n## The Review Step: Why Source Links Matter\n\nSpeed is not the only goal. The line list has to be right. That means the engineering team needs to be able to check the extracted output efficiently.\n\nSource links make this possible. With Novek, every row in the line list shows which drawing page it came from and where on that page. Engineers can open the source, scan the callout, and confirm the extraction â in seconds per row rather than minutes.\n\nThis is the difference between a fast extraction that still requires full re-reading of the drawing set to verify, and a fast extraction that is also fast to check.\n\n---\n\n## Common Questions From Engineering Teams\n\n**Can AI handle all P&ID line naming conventions?**\nLine designation formats vary by contractor and project. Novek can be configured to match the project's line naming convention. Where deviations exist, rows are flagged for review rather than silently misread.\n\n**What about continuing lines that span multiple drawings?**\nContinuing lines are identified and consolidated. The final line list entry captures all drawing references, not just the first sheet where the line appears.\n\n**Our P&IDs are in revision â can we re-run extraction after a revision?**\nYes. Updated drawings can be re-run and the delta â new lines, removed lines, changed attributes â is shown as a change list against the previous extraction.\n\n**How does this work with draft or in-progress P&IDs?**\nExtraction works on draft P&IDs with the same pipeline. Lower-completeness drawings produce partial line data, which is clearly indicated so engineers know what still needs to be confirmed.\n\n---\n\n## What Teams Report After Switching to AI Line List Extraction\n\nTeams using Novek for line list work consistently report the same outcomes:\n\n- Line list completion in hours, not weeks\n- Engineers spending time on review and check rather than data entry\n- Fewer errors reaching client review â mismatches are caught during extraction, not during QA\n- More projects serviced by the same team without additional headcount\n\nThe structural change is not just speed. It is that the engineering team's role shifts from doing data entry to reviewing an extracted output â which is a fundamentally higher-value use of engineering time.\n "},{slug:"what-is-ai-pid-extraction",title:"What Is AI P&ID Extraction? A Practical Guide for EPC Engineers",excerpt:"AI P&ID extraction turns piping and instrumentation diagrams into structured engineering data automatically. This guide explains what it is, how it works, and what EPC teams actually get out of it.",date:"April 4, 2026",readTime:"11 min read",author:"Novek AI",category:"AI P&ID Extraction",keywords:["AI P&ID extraction","P&ID data extraction","piping and instrumentation diagram extraction","P&ID digitization","engineering drawing extraction","tag extraction from P&ID"],image:"/Images/Chat_bot_illustration.png",gradient:"from-sky-500 to-blue-700",featured:!1,content:"\n## What Is a P&ID and Why Is Extracting Data from It Hard?\n\nA Piping and Instrumentation Diagram (P&ID) is the primary technical drawing for any process plant or industrial facility. It shows every pipe, instrument, valve, vessel, and piece of equipment in a process system, along with the control and signal connections between them. Every significant engineering deliverable downstream â tag registers, line lists, instrument indexes, MTOs â traces back to the P&ID package.\n\nThe problem is that P&IDs are complex visual documents. The useful engineering information sits inside:\n\n- Instrument symbols (ISA-5.1 notation)\n- Line designation callouts\n-
1Equipment tag bubbles\n- Process notes and annotations\n- Title block references\n\nNone of this information is in a structured format. To use it, engineers have to read it manually, interpret the notation, and key it into the appropriate deliverable. Across a package of 80â100 drawings, that reading and extraction work is measured in weeks.\n\n---\n\n## What AI P&ID Extraction Means\n\n**AI P&ID extraction** refers to using an AI system to read engineering drawings and convert the useful information inside them into structured data â automatically.\n\nThis is distinct from:\n\n- **OCR** â optical character recognition reads text characters but does not understand what they mean in engineering context\n- **Smart P&IDs** â digital authoring formats (SPF, AVEVA PDMS) are structured by design, but most legacy projects use PDF-format drawings\n- **General-purpose AI** â a general language model can read text but cannot interpret P&ID symbology or maintain the structural relationships between tags, lines, and equipment\n\nAI P&ID extraction is specifically designed for engineering drawings, with the ability to:\n\n1. Recognise instrument and equipment symbols by shape and position\n2. Parse engineering notation correctly (tag formats, line designations, pipe classes)\n3. Preserve relationships between symbols and their adjacent text\n4. Maintain traceability from each extracted value back to its source drawing\n\n---\n\n## What an AI P&ID Extraction System Produces\n\nThe outputs from P&ID extraction are the structured deliverables the engineering team needs to build:\n\n| Output | What It Contains |\n|---|---|\n| Tag register | All tagged items: instruments, valves, equipment, piping lines |\n| Line list | All piping lines with service, class, and from/to connections |\n| Instrument list | All instruments with type, tag, line, and datasheet reference |\n| Equipment list | All tagged equipment with description and P&ID reference |\n| Valve list | All valves with type, tag, size, class, and actuator details |\n| Mismatch report | Tags and attributes that conflict across drawings |\n\nThese are the deliverables the team would have produced manually. AI extraction produces them systematically, in a fraction of the time, with every value linked back to its source.\n\n---\n\n## The Five Steps in Novek's P&ID Extraction Pipeline\n\n**Step 1 â Package ingestion**\nThe drawing set is loaded as a package. Drawings are organised by sheet number and revision, with relationships between sheets tracked.\n\n**Step 2 â Visual symbol recognition**\nEach drawing is scanned visually for engineering symbols. ISA-5.1 instruments, valve types, equipment items, and line flow indicators are identified by shape and context â not by reading text alone.\n\n**Step 3 â Notation parsing**\nAdjacent text is parsed together with each symbol. Tag numbers, line designations, service abbreviations, and equipment references are extracted as structured attributes linked to their source symbol.\n\n**Step 4 â Cross-drawing consolidation**\nTags and lines that appear across multiple sheets are consolidated. A flow transmitter that appears on three P&IDs is one entry in the tag register, with all three references captured.\n\n**Step 5 â Source-linked export**\nEach extracted row links to its source drawing page and position. Engineers can verify any value against the source without re-reading the full drawing.\n\n---\n\n## Why General-Purpose AI Tools Struggle With P&IDs\n\nA common question is why tools like general-purpose LLMs (GPT, Claude, Gemini) cannot just read a P&ID and produce the same outputs.\n\nThere are three structural reasons:\n\n**Symbols are not text.** ISA-5.1 instrumentation symbols are geometric shapes. A circle with a line through it means a specific type of instrument. A rectangle with specific line types means a different type of equipment. Language models trained on text cannot reliably interpret these.\n\n**Relationships are spatial.** The relationship between a tag number and its instrument is defined by proximity on the drawing â not by any explicit linguistic connection. General AI models that convert drawings to flat text lose these relationships.\n\n**Consistency matters.** A tag register built by sampling a few drawings and extrapolating is not useful. EPC deliverables need to be complete â every drawing, every tag, every line. Systems designed for this task have completeness checking built in. General-purpose tools do not.\n\n---\n\n## A Practical Example: What Changes on a Real EPC Project\n\nConsider an EPC team receiving a 100-drawing P&ID package for a process facility expansion. The conventional workflow:\n\n1. Assign 2â3 engineers to read drawings and build tag register: 2 weeks\n2. QA check â cross-reference against drawing set: 3 days\n3. Build line list from same package: 2 more weeks\n4. Build instrument list: 1 additional week\n\nTotal: approximately 6 weeks to produce the foundational deliverables.\n\nWith AI P&ID extraction:\n\n1. Run extraction across the full package: hours\n2. Engineering review using source-linked outputs: 1â2 days\n3. Export in project template format: immediate\n\nTotal: 2â3 days for the same set of outputs, with full traceability throughout.\n\nThe engineering team's role shifts from data collection to review â which is a better use of their specialist knowledge, and a faster path to client-ready deliverables.\n\n---\n\n## Frequently Asked Questions\n\n**Does AI P&ID extraction work with PDF-format drawings?**\nYes. Novek is designed for standard PDF-format P&IDs used in most legacy and current projects. Smart P&ID formats are not required.\n\n**What about P&IDs with non-standard conventions?**\nEvery project has its own drawing conventions. Novek adapts to the project's specific notation, and any notation it encounters that falls outside its model is flagged for engineering review.\n\n**How is accuracy verified?**\nEvery extracted value stays linked to its source drawing page and position. The engineering team can verify any row by clicking through to the source â no re-reading of the full drawing required.\n\n**What happens when drawings are revised?**\nUpdated drawing revisions can be re-processed. The system shows what changed between the old and new extraction â new tags, removed tags, modified attributes.\n\n**Can the outputs be exported in our project template?**\nYes. Output format maps to the client's project template â column names, attribute selection, and field order are configurable per project.\n "},{slug:"beyond-smart-pids-how-intelligent-documents-are-reshaping-industrial-asset-management",title:"Beyond Smart P&IDs: How Intelligent Documents Are Reshaping Industrial Asset Management",excerpt:"Discover how NovekAI's intelligent documents transcend Smart P&IDs, providing contextual understanding, automated workflows, and interconnected knowledge for superior asset management.",date:"May 3, 2025",readTime:"10 min read",author:"Novek AI",category:"Industrial Asset Management",keywords:["smart P&ID","intelligent documents","industrial asset management","AI engineering"],image:"/Images/Properties_Illustration.png",gradient:"from-blue-500 to-purple-500",featured:!1,content:"\n## Bey
1ond Smart P&IDs\n\nTraditional Smart P&IDs store engineering data in structured authoring formats. Novek takes this further by making standard PDF-format P&IDs equally queryable and extractable â no authoring tool required.\n\nThis post explores how intelligent document understanding is changing what's possible in industrial asset management.\n\n### The Limitation of Smart P&ID Formats\n\nSmart P&ID authoring tools like AVEVA E3D or Hexagon SPF produce structured, database-backed drawings. But most engineering knowledge in the world is not in these formats â it is in PDF drawings, Excel datasheets, and Word procedures that pre-date or sit outside of these systems.\n\n### What Novek Does Differently\n\nNovek applies visual AI to standard PDFs, giving teams the same structured-data access without requiring everyone to work in a Smart P&ID authoring environment.\n\nThe result: engineering teams can extract tags, lines, instruments, and equipment references from their existing document libraries â no conversion, no re-authoring.\n "},{slug:"unlocking-operational-excellence-how-generative-ai-is-revolutionizing-document-management",title:"Unlocking Operational Excellence: How Generative AI is Revolutionizing Document Management in Asset-Intensive Industries",excerpt:"Discover how Generative AI is transforming document management in asset-intensive industries, enabling instant searchability, real-time assistance, and automated workflows.",date:"May 2, 2025",readTime:"12 min read",author:"Novek AI",category:"Generative AI",keywords:["generative AI","document management","asset-intensive industries","operational excellence"],image:"/Images/Chat_bot_illustration.png",gradient:"from-green-500 to-blue-500",featured:!1,content:"\n## How Generative AI Is Changing Document Management\n\nIn asset-intensive industries â oil and gas, chemicals, power, mining â critical knowledge is buried in technical documents. Generative AI is transforming how teams access and use that knowledge.\n\n### The Documentation Challenge\n\nAsset-intensive industries rely on vast documentation: P&IDs, equipment datasheets, maintenance records, and operating manuals. Historically, this information is scattered and hard to access quickly.\n\n### What Generative AI Enables\n\n- **Instant Searchability**: Scan and interpret P&IDs, datasheets, and manuals instantly\n- **Real-Time Assistance**: Query AI systems for specific procedures or troubleshooting steps\n- **Automated Workflows**: Routine extraction and compliance tasks automated\n- **Knowledge Retention**: Centralised access reduces reliance on key individuals\n\n### Real-World Impact\n\nIndustries adopting Generative AI for document management report significant gains in uptime, compliance, and engineering efficiency.\n\nNovek is purpose-built for this use case â trained on engineering documents, not general text.\n "}];function i(){return[...a].sort((e,t)=>new Date(t.date).getTime()-new Date(e.date).getTime())}function r(){let e=arguments.length>0&&void 0!==arguments[0]?arguments[0]:3;return i().filter(e=>e.featured).slice(0,e)}},42714:(e,t)=>{"use strict";Object.defineProperty(t,"__esModule",{value:!0}),Object.defineProperty(t,"setAttributesFromProps",{enumerable:!0,get:function(){return r}});let n={acceptCharset:"accept-charset",className:"class",htmlFor:"for",httpEquiv:"http-equiv",noModule:"noModule"},a=["onLoad","onReady","dangerouslySetInnerHTML","children","onError","strategy","stylesheets"];function i(e){return["async","defer","noModule"].includes(e)}function r(e,t){for(let[r,s]of Object.entries(t)){if(!t.hasOwnProperty(r)||a.includes(r)||void 0===s)continue;let o=n[r]||r.toLowerCase();"SCRIPT"===e.tagName&&i(o)?e[o]=!!s:e.setAttribute(o,String(s)),(!1===s||"SCRIPT"===e.tagName&&i(o)&&(!s||"false"===s))&&(e.setAttribute(o,""),e.removeAttribute(o))}}("function"==typeof t.default||"object"==typeof t.default&&null!==t.default)&&void 0===t.default.__esModule&&(Object.defineProperty(t.default,"__esModule",{value:!0}),Object.assign(t.default,t),e.exports=t.default)},52063:(e,t,n)=>{"use strict";n.d(t,{r:()=>i});var a=n(12115);function i(){let{threshold:e=.1,rootMargin:t="0px 0px -100px 0px",triggerOnce:n=!0}=arguments.length>0&&void 0!==arguments[0]?arguments[0]:{},[i,r]=(0,a.useState)(!1),s=(0,a.useRef)(null);return(0,a.useEffect)(()=>{let a=new IntersectionObserver(e=>{let[t]=e;t.isIntersecting?(r(!0),n&&a.unobserve(t.target)):n||r(!1)},{threshold:e,rootMargin:t});return s.current&&a.observe(s.current),()=>a.disconnect()},[e,t,n]),{ref:s,isVisible:i}}}
1,53088:(e,t,n)=>{"use strict";n.d(t,{BlogHero:()=>o});var a=n(95155),i=n(1065),r=n(52063);let s=(0,n(19946).A)("BookOpen",[["path",{d:"M12 7v14",key:"1akyts"}],["path",{d:"M3 18a1 1 0 0 1-1-1V4a1 1 0 0 1 1-1h5a4 4 0 0 1 4 4 4 4 0 0 1 4-4h5a1 1 0 0 1 1 1v13a1 1 0 0 1-1 1h-6a3 3 0 0 0-3 3 3 3 0 0 0-3-3z",key:"ruj8y"}]]);function o(){let{ref:e,isVisible:t}=(0,r.r)({threshold:.2});return(0,a.jsxs)("section",{ref:e,className:"relative pt-32 pb-16 overflow-hidden bg-background",children:[(0,a.jsx)("div",{className:"absolute inset-0",children:(0,a.jsxs)(i.x,{speed:.3,className:"absolute inset-0",children:[(0,a.jsx)("div",{className:"absolute top-20 left-20 w-96 h-96 bg-blue-400/15 dark:bg-blue-900/20 rounded-full blur-3xl animate-pulse"}),(0,a.jsx)("div",{className:"absolute bottom-10 right-20 w-80 h-80 bg-cyan-400/15 dark:bg-cyan-900/20 rounded-full blur-3xl animate-pulse delay-1000"})]})}),(0,a.jsx)("div",{className:"relative max-w-7xl mx-auto px-4 sm:px-6 lg:px-8",children:(0,a.jsxs)("div",{className:"text-center space-y-8",children:[(0,a.jsxs)("div",{className:"transition-all duration-1000 ease-out ".concat(t?"opacity-100 translate-y-0":"opacity-0 translate-y-8"),children:[(0,a.jsx)("div",{className:"flex items-center justify-center mb-6",children:(0,a.jsxs)("div",{className:"flex items-center gap-2 bg-blue-50 dark:bg-blue-900/30 px-4 py-2 rounded-full border border-blue-200 dark:border-blue-800",children:[(0,a.jsx)(s,{className:"w-4 h-4 text-blue-600 dark:text-blue-400"}),(0,a.jsx)("span",{className:"text-sm font-semibold text-blue-700 dark:text-blue-300",children:"Engineering Insights"})]})}),(0,a.jsxs)("h1",{className:"text-4xl sm:text-5xl lg:text-7xl font-bold leading-tight",children:[(0,a.jsx)("span",{className:"text-foreground",children:"The Novek"}),(0,a.jsx)("br",{}),(0,a.jsx)("span",{className:"bg-gradient-to-r from-blue-600 to-cyan-600 bg-clip-text text-transparent",children:"Blog"})]})]}),(0,a.jsx)("div",{className:"transition-all duration-1000 delay-200 ease-out ".concat(t?"opacity-100 translate-y-0":"opacity-0 translate-y-8"),children:(0,a.jsxs)("p",{className:"text-lg sm:text-xl text-muted-foreground leading-relaxed max-w-3xl mx-auto",children:["Practical guides on"," ",(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"AI P&ID extraction"}),","," ",(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"line list extraction"}),","," ",(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"tag list extraction"}),", and"," ",(0,a.jsx)("span",{className:"font-semibold text-foreground",children:"instrument list extraction"})," ","for EPC firms and engineering consultants."]})})]})})]})}},69243:(e,t,n)=>{"use strict";Object.defineProperty(t,"__esModule",{value:!0}),!function(e,t){for(var n in t)Object.defineProperty(e,n,{enumerable:!0,get:t[n]})}(t,{default:function(){return w},handleClientScriptLoad:function(){return p},initScriptLoader:function(){return f}});let a=n(88229),i=n(6966),r=n(95155),s=a._(n(47650)),o=i._(n(12115)),l=n(82830),d=n(42714),c=n(92374),u=new Map,h=new Set,g=e=>{if(s.default.preinit){e.forEach(e=>{s.default.preinit(e,{as:"style"})});return}{let t=document.head;e.forEach(e=>{let n=document.createElement("link");n.type="text/css",n.rel="stylesheet",n.href=e,t.appendChild(n)})}},m=e=>{let{src:t,id:n,onLoad:a=()=>{},onReady:i=null,dangerouslySetInnerHTML:r,children:s="",strategy:o="afterInteractive",onError:l,stylesheets:c}=e,m=n||t;if(m&&h.has(m))return;if(u.has(t)){h.add(m),u.get(t).then(a,l);return}let p=()=>{i&&i(),h.add(m)},f=document.createElement("script"),y=new Promise((e,t)=>{f.addEventListener("load",function(t){e(),a&&a.call(this,t),p()}),f.addEventListener("error",function(e){t(e)})}).catch(function(e){l&&l(e)});r?(f.innerHTML=r.__html||"",p()):s?(f.textContent="string"==typeof s?s:Array.isArray(s)?s.join(""):"",p()):t&&(f.src=t,u.set(t,y)),(0,d.setAttributesFromProps)(f,e),"w
1orker"===o&&f.setAttribute("type","text/partytown"),f.setAttribute("data-nscript",o),c&&g(c),document.body.appendChild(f)};function p(e){let{strategy:t="afterInteractive"}=e;"lazyOnload"===t?window.addEventListener("load",()=>{(0,c.requestIdleCallback)(()=>m(e))}):m(e)}function f(e){e.forEach(p),[...document.querySelectorAll('[data-nscript="beforeInteractive"]'),...document.querySelectorAll('[data-nscript="beforePageRender"]')].forEach(e=>{let t=e.id||e.getAttribute("src");h.add(t)})}function y(e){let{id:t,src:n="",onLoad:a=()=>{},onReady:i=null,strategy:d="afterInteractive",onError:u,stylesheets:g,...p}=e,{updateScripts:f,scripts:y,getIsSsr:w,appDir:b,nonce:v}=(0,o.useContext)(l.HeadManagerContext),x=(0,o.useRef)(!1);(0,o.useEffect)(()=>{let e=t||n;x.current||(i&&e&&h.has(e)&&i(),x.current=!0)},[i,t,n]);let I=(0,o.useRef)(!1);if((0,o.useEffect)(()=>{if(!I.current){if("afterInteractive"===d)m(e);else if("lazyOnload"===d)"complete"===document.readyState?(0,c.requestIdleCallback)(()=>m(e)):window.addEventListener("load",()=>{(0,c.requestIdleCallback)(()=>m(e))});I.current=!0}},[e,d]),("beforeInteractive"===d||"worker"===d)&&(f?(y[d]=(y[d]||[]).concat([{id:t,src:n,onLoad:a,onReady:i,onError:u,...p}]),f(y)):w&&w()?h.add(t||n):w&&!w()&&m(e)),b){if(g&&g.forEach(e=>{s.default.preinit(e,{as:"style"})}),"beforeInteractive"===d)return n?(s.default.preload(n,p.integrity?{as:"script",integrity:p.integrity,nonce:v,crossOrigin:p.crossOrigin}:{as:"script",nonce:v,crossOrigin:p.crossOrigin}),(0,r.jsx)("script",{nonce:v,dangerouslySetInnerHTML:{__html:"(self.__next_s=self.__next_s||[]).push("+JSON.stringify([n,{...p,id:t}])+")"}})):(p.dangerouslySetInnerHTML&&(p.children=p.dangerouslySetInnerHTML.__html,delete p.dangerouslySetInnerHTML),(0,r.jsx)("script",{nonce:v,dangerouslySetInnerHTML:{__html:"(self.__next_s=self.__next_s||[]).push("+JSON.stringify([0,{...p,id:t}])+")"}}));"afterInteractive"===d&&n&&s.default.preload(n,p.integrity?{as:"script",integrity:p.integrity,nonce:v,crossOrigin:p.crossOrigin}:{as:"script",nonce:v,crossOrigin:p.crossOrigin})}return null}Object.defineProperty(y,"__nextScript",{value:!0});let w=y;("function"==typeof t.default||"object"==typeof t.default&&null!==t.default)&&void 0===t.default.__esModule&&(Object.defineProperty(t.default,"__esModule",{value:!0}),Object.assign(t.default,t),e.exports=t.default)},81877:(e,t,n)=>{"use strict";n.d(t,{BlogFeatured:()=>g});var a=n(95155),i=n(24506),r=n(52063),s=n(69074),o=n(14186),l=n(92138),d=n(66766),c=n(6874),u=n.n(c);let h=(0,i.ue)(4);function g(){let{ref:e,isVisible:t}=(0,r.r)({threshold:.1});return 0===h.length?null:(0,a.jsx)("section",{ref:e,className:"py-20 bg-background relative overflow-hidden",children:(0,a.jsx)("div",{className:"relative max-w-7xl mx-auto px-4 sm:px-6 lg:px-8",children:(0,a.jsxs)("div",{className:"transition-all duration-1000 ease-out ".concat(t?"opacity-100 translate-y-0":"opacity-0 translate-y-8"),children:[(0,a.jsxs)("div",{className:"mb-12",children:[(0,a.jsx)("div",{className:"inline-flex rounded-full border border-blue-200 bg-blue-50 px-4 py-2 text-sm font-semibold text-blue-800 dark:border-blue-900 dark:bg-blue-950/50 dark:text-blue-200 mb-4",children:"Featured guides"}),(0,a.jsx)("h2",{className:"text-3xl font-bold text-foreground sm:text-4xl",children:"Start with the topics that matter most"}),(0,a.jsx)("p",{className:"mt-3 text-lg text-muted-foreground max-w-3xl",children:"Practical guides on AI P&ID extraction, line list extraction, tag list extraction, and instrument list extraction for EPC teams."})]}),(0,a.jsx)("div",{className:"grid md:grid-cols-2 gap-6",children:h.map((e,n)=>(0,a.jsx)(u(),{href:"/blog/".concat(e.slug),className:"group block",children:(0,a.jsxs)("article",{className:"h-full rounded-[2rem] overflow-hidden border border-border bg-background shadow-sm transition-all duration-500 hover:-translate-y-1 hover:shadow-xl ".concat(t?"opacity-100 translate-y-0":"opacity-0 translate-y-8"),style:{transitionDelay:"".concat(80*n,"ms")},children:[(0,a.jsxs)("div",{className:"relative h-44 overflow-hidden",children:[(0,a.jsx)(d.default,{src:e.image||"/Images/Chat_bot_illustration.png",alt:e.title,width:600,height:300,className:"w-full h-full object-cover transition-transform duration-500 group-hover:scale-105"}),(0,a.jsx)("div",{className:"absolute inset-0 bg-gradient-to-br ".concat(e.gradient," opacity-75")}
1),(0,a.jsx)("span",{className:"absolute top-4 left-4 rounded-full bg-white/20 backdrop-blur-sm px-3 py-1 text-xs font-semibold text-white border border-white/30",children:e.category})]}),(0,a.jsxs)("div",{className:"p-6 space-y-3",children:[(0,a.jsx)("h3",{className:"text-lg font-bold text-foreground leading-snug group-hover:text-blue-700 dark:group-hover:text-blue-300 transition-colors duration-200",children:e.title}),(0,a.jsx)("p",{className:"text-sm text-muted-foreground leading-relaxed line-clamp-2",children:e.excerpt}),(0,a.jsxs)("div",{className:"flex items-center justify-between pt-2 border-t border-border",children:[(0,a.jsxs)("div",{className:"flex items-center gap-3 text-xs text-muted-foreground",children:[(0,a.jsxs)("span",{className:"flex items-center gap-1",children:[(0,a.jsx)(s.A,{className:"w-3.5 h-3.5"}),e.date]}),(0,a.jsxs)("span",{className:"flex items-center gap-1",children:[(0,a.jsx)(o.A,{className:"w-3.5 h-3.5"}),e.readTime]})]}),(0,a.jsxs)("span",{className:"flex items-center gap-1 text-xs font-semibold text-blue-600 dark:text-blue-400 group-hover:translate-x-1 transition-transform duration-300",children:["Read ",(0,a.jsx)(l.A,{className:"w-3.5 h-3.5"})]})]})]})]})},e.slug))})]})})})}},85423:(e,t,n)=>{"use strict";n.d(t,{BlogGrid:()=>m});var a=n(95155),i=n(24506),r=n(52063),s=n(69074),o=n(14186),l=n(92138),d=n(66766),c=n(6874),u=n.n(c),h=n(12115);let g=(0,i.zX)();function m(){let{ref:e,isVisible:t}=(0,r.r)({threshold:.1}),[n,i]=(0,h.useState)(6),c=g.slice(0,n),m=n<g.length;return(0,a.jsx)("section",{ref:e,className:"py-20 bg-blue-50/30 dark:bg-blue-950/10 relative",children:(0,a.jsxs)("div",{className:"relative max-w-7xl mx-auto px-4 sm:px-6 lg:px-8",children:[(0,a.jsxs)("div",{className:"mb-12 transition-all duration-1000 ease-out ".concat(t?"opacity-100 translate-y-0":"opacity-0 translate-y-8"),children:[(0,a.jsx)("h2",{className:"text-3xl font-bold text-foreground sm:text-4xl",children:"All Articles"}),(0,a.jsxs)("p",{className:"mt-2 text-lg text-muted-foreground",children:[g.length," article",1!==g.length?"s":""," on AI-powered engineering extraction"]})]}),(0,a.jsx)("div",{className:"grid md:grid-cols-2 lg:grid-cols-3 gap-6",children:c.map((e,n)=>(0,a.jsx)(u(),{href:"/blog/".concat(e.slug),className:"group block",children:(0,a.jsxs)("article",{className:"h-full rounded-[2rem] overflow-hidden border border-border bg-background shadow-sm transition-all duration-700 hover:-translate-y-1 hover:shadow-xl ".concat(t?"opacity-100 translate-y-0":"opacity-0 translate-y-8"),style:{transitionDelay:"".concat(60*n,"ms")},children:[(0,a.jsxs)("div",{className:"relative h-40 overflow-hidden",children:[(0,a.jsx)(d.default,{src:e.image||"/Images/Chat_bot_illustration.png",alt:e.title,width:400,height:200,className:"w-full h-full object-cover transition-transform duration-500 group-hover:scale-105"}),(0,a.jsx)("div",{className:"absolute inset-0 bg-gradient-to-br ".concat(e.gradient," opacity-70")}),(0,a.jsx)("span",{className:"absolute top-3 left-3 rounded-full bg-white/20 backdrop-blur-sm px-2.5 py-1 text-xs font-semibold text-white border border-white/30",children:e.category})]}),(0,a.jsxs)("div",{className:"p-5 space-y-3",children:[(0,a.jsx)("h3",{className:"text-base font-bold text-foreground leading-snug group-hover:text-blue-700 dark:group-hover:text-blue-300 transition-colors duration-200",children:e.title}),(0,a.jsx)("p",{className:"text-sm text-muted-foreground leading-relaxed line-clamp-2",children:e.excerpt}),(0,a.jsxs)("div",{className:"flex items-center justify-between pt-2 border-t border-border",children:[(0,a.jsxs)("div",{className:"flex items-center gap-3 text-xs text-muted-foreground",children:[(0,a.jsxs)("span",{className:"flex items-center gap-1",children:[(0,a.jsx)(s.A,{className:"w-3 h-3"}),e.date]}),(0,a.jsxs)("span",{className:"flex items-center gap-1",children:[(0,a.jsx)(o.A,{className:"w-3 h-3"}),e.readTime]})]}),(0,a.jsx)(l.A,{className:"w-4 h-4 text-blue-600 dark:text-blue-400 group-hover:translate-x-1 transition-transform duration-300"})]})]})]})},e.slug))}),m&&(0,a.jsx)("div",{className:"mt-12 text-center",children:(0,a.jsxs)("button",{onClick:()=>i(e=>e+6),className:"inline-flex items-center gap-2 rounded-full border border-blue-200 bg-blue-50 px-7 py-3 text-sm font-semibold text-blue-800 transition-all duration-300 hover:bg-blue-100 dar
1k:border-blue-800 dark:bg-blue-950/40 dark:text-blue-200 dark:hover:bg-blue-900/60",children:["Load more articles",(0,a.jsx)(l.A,{className:"h-4 w-4"})]})})]})})}},92138:(e,t,n)=>{"use strict";n.d(t,{A:()=>a});let a=(0,n(19946).A)("ArrowRight",[["path",{d:"M5 12h14",key:"1ays0h"}],["path",{d:"m12 5 7 7-7 7",key:"xquz4c"}]])},92374:(e,t)=>{"use strict";Object.defineProperty(t,"__esModule",{value:!0}),!function(e,t){for(var n in t)Object.defineProperty(e,n,{enumerable:!0,get:t[n]})}(t,{cancelIdleCallback:function(){return a},requestIdleCallback:function(){return n}});let n="undefined"!=typeof self&&self.requestIdleCallback&&self.requestIdleCallback.bind(window)||function(e){let t=Date.now();return self.setTimeout(function(){e({didTimeout:!1,timeRemaining:function(){return Math.max(0,50-(Date.now()-t))}})},1)},a="undefined"!=typeof self&&self.cancelIdleCallback&&self.cancelIdleCallback.bind(window)||function(e){return clearTimeout(e)};("function"==typeof t.default||"object"==typeof t.default&&null!==t.default)&&void 0===t.default.__esModule&&(Object.defineProperty(t.default,"__esModule",{value:!0}),Object.assign(t.default,t),e.exports=t.default)},97195:(e,t,n)=>{Promise.resolve().then(n.bind(n,81877)),Promise.resolve().then(n.bind(n,85423)),Promise.resolve().then(n.bind(n,53088)),Promise.resolve().then(n.bind(n,47746)),Promise.resolve().then(n.bind(n,39464)),Promise.resolve().then(n.t.bind(n,69243,23))}},e=>{var t=t=>e(e.s=t);e.O(0,[6711,5254,4354,8809,3729,8441,1684,7358],()=>t(97195)),_N_E=e.O()}]);
Line numbers count LF bytes from the start of the resource, as the search results do. Vendor segments are library code the classifier recognised; they are stored but not indexed. Bytes are shown as Latin1 characters, one per byte.