{
    "componentChunkName": "component---src-templates-blog-post-js",
    "path": "/en/blog/2026-10-11-why-ai-does-not-design-the-structure/",
    "result": {"data":{"markdownRemark":{"html":"<p>Generative AI models can read a drawing, summarise a calculation note or suggest a plausible value. It is tempting to ask them directly: “can this tower take one more antenna?”. That is precisely what we do not do. <strong>At TriDyme, AI does not design the structure.</strong> It prepares the data, runs the validated calculation engine and checks the result.</p>\n<h2>A structural calculation carries liability</h2>\n<p>A calculation note is not an opinion. It demonstrates that a structure can withstand the loads set by the standards, and it commits the engineer who signs it. It must therefore be:</p>\n<ul>\n<li><strong>reproducible</strong>: the same inputs always give the same result;</li>\n<li><strong>verifiable</strong>: an inspection body must be able to redo the calculation and find the same values;</li>\n<li><strong>compliant</strong>: with the Eurocodes and their national annexes, for example EN 1993-3-1 for towers.</li>\n</ul>\n<p>A language model has none of these properties. Its answer can change from one run to the next, it can produce a convincing but wrong figure, and it cannot explain in a checkable way how it got there.</p>\n<p>That is why calculations on the platform are run by a <strong>deterministic engine</strong>, whose calculations have been validated by an inspection body. The resulting deliverables carry engineering liability and are insured.</p>\n<h2>What AI does: everything around the calculation</h2>\n<p>In an engineering file, the calculation itself takes a small share of the time. The rest is finding, reading, reconciling and checking data. That is where AI is useful.</p>\n<p><strong>Before the calculation, it prepares the data.</strong> It reads archive documents (calculation notes, as-built files, drawings), extracts the technical data, reconciles it with the surveyed equipment and flags discrepancies. Every proposed value carries its source: document and page.</p>\n<p><strong>It runs the engine.</strong> Once the inputs are approved, the calculation is run by the validated calculation engine, on one site or thousands.</p>\n<p><strong>After the calculation, it checks and explains.</strong> When a model does not match its documents, an agent compares it with the latest as-built file and calculation note, identifies the problem and <strong>proposes the input data to correct — never a calculated value</strong>.</p>\n<h2>A human in the loop, at every binding step</h2>\n<p>AI proposals never change the reference data directly. They are placed on a <strong>branch</strong>, like a proposed code change: an engineer reviews the diff, then merges or rejects it. Every decision is logged: who approved, when, on which data, with which calculation result.</p>\n<p>This approach has two benefits:</p>\n<ul>\n<li>the engineer stays in control of what goes into a calculation note;</li>\n<li>the tedious work of collecting and reconciling data is automated, so a whole portfolio can be processed rather than one file at a time.</li>\n</ul>\n<h2>The right division of roles</h2>\n<table>\n<thead>\n<tr>\n<th>Step</th>\n<th>Who does it</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>Read and structure documents</td>\n<td>AI, with a confidence level per data point</td>\n</tr>\n<tr>\n<td>Reconcile documents with installed equipment</td>\n<td>AI, every discrepancy sourced</td>\n</tr>\n<tr>\n<td>Approve the input data</td>\n<td>An engineer</td>\n</tr>\n<tr>\n<td>Calculate</td>\n<td>The validated deterministic engine</td>\n</tr>\n<tr>\n<td>Check the consistency of the result</td>\n<td>AI, then the engineer</td>\n</tr>\n<tr>\n<td>Sign the deliverable</td>\n<td>An engineer</td>\n</tr>\n</tbody>\n</table>\n<h2>Key takeaways</h2>\n<ul>\n<li>A structural calculation must be reproducible, verifiable and compliant: that is the job of a deterministic engine, not a language model.</li>\n<li>AI brings its efficiency where time is really lost: reading, reconciling and checking data.</li>\n<li>Every AI proposal is sourced, reviewed and approved by an engineer before it is used.</li>\n</ul>\n<p><strong>Want to see how it works on your data?</strong> Discover <a href=\"/plateforme/\">the platform</a> and <a href=\"/solutions/agents-ia/\">our AI agents</a>, or <a href=\"/contact/?subject=demo\">talk to an expert</a>.</p>","excerpt":"Generative AI models can read a drawing, summarise a calculation note or suggest a plausible value. It is tempting to ask them directly…","frontmatter":{"title":"Why AI does not design the structure (and what it does instead)","language":"en-GB","date":"11 octobre 2026","author":"TriDyme Team","description":"A calculation note carries engineering liability. At TriDyme, AI prepares the data, runs the validated calculation engine and checks the result — but never replaces the calculation.","featuredImage":{"publicURL":"/static/277a8c96546d448336657f7fe27af91b/blog-ia-calcul-structure.jpg"}},"fields":{"slug":"/blog/2026-10-11-why-ai-does-not-design-the-structure/"}},"relatedPosts":{"edges":[{"node":{"id":"4f60bf0c-be40-50f2-b64c-138f23fa539a","frontmatter":{"title":"Des agents IA qui travaillent comme des collaborateurs","date":"11 octobre 2026","featuredImage":{"publicURL":"/static/4afbfb9fcc669c9f8c3fc530ae631106/blog-agents-ia-autonomes.jpg"}},"fields":{"slug":"/blog/2026-10-11-agents-ia-collaborateurs/"}}},{"node":{"id":"a47b3e6e-8367-515e-9fef-3d44fa34225e","frontmatter":{"title":"AI agents that work like team members","date":"11 octobre 2026","featuredImage":{"publicURL":"/static/4afbfb9fcc669c9f8c3fc530ae631106/blog-agents-ia-autonomes.jpg"}},"fields":{"slug":"/blog/2026-10-11-ai-agents-like-team-members/"}}},{"node":{"id":"725ea80a-e400-58d0-adc6-3678ecfe7042","frontmatter":{"title":"Pourquoi l’IA ne calcule pas la structure (et ce qu’elle fait à la place)","date":"11 octobre 2026","featuredImage":{"publicURL":"/static/277a8c96546d448336657f7fe27af91b/blog-ia-calcul-structure.jpg"}},"fields":{"slug":"/blog/2026-10-11-pourquoi-ia-ne-calcule-pas-la-structure/"}}}]}},"pageContext":{"slug":"/blog/2026-10-11-why-ai-does-not-design-the-structure/","locale":"en"}},
    "staticQueryHashes": ["2349346863"]}