---
title: "How engineering software wins customers through AI search"
description: "Engineering software gets recommended by AI when its workflows, file formats, plans and limits are stated plainly and confirmed by sources engineers trust."
canonical: "https://underneath.agency/resources/engineering-software-ai-search"
published: 2026-10-07
updated: 2026-10-08
publisher: "Underneath (https://underneath.agency/agent)"
entity: "https://underneath.agency/.well-known/entity.json"
---
Guide · AI search

# When engineers ask AI which CAD, simulation or PLM tool to use, will they hear our name?

They will if an assistant can find and verify what your software does for a specific workflow, what it costs to start, which files it reads and where it falls short. Engineers already use AI in their own work, software buyers increasingly ask chatbots for recommendations, and a wave of acquisitions has many teams asking what to use next.

This guide is for companies that sell computer-aided design (CAD), simulation and analysis (CAE) and product lifecycle management (PLM) software to engineers. It is not about engineering services firms, which we cover separately, or software for running factories. The revenue here comes from a trial or free plan that turns into paid seats, then an enterprise agreement.

## The short version

1. The category is large and growing: [CIMdata](https://www.industrialmachinerydigest.com/software/quality-management-software/cimdata-publishes-executive-plm-market-report) puts the broader PLM market, which includes CAD, simulation and PLM platforms, at $88.3 billion in 2025, up 9.9%.
2. The leaders sell recurring subscriptions: [Autodesk](https://adsknews.autodesk.com/en/?p=55945) reported $7,206 million in fiscal 2026 revenue, up 18%, with its manufacturing products at $1,379 million, and [Dassault Systèmes](https://www.3ds.com/assets/invest/2026-02/dassault-systemes-25q4-earnings_pr_va.pdf) said recurring revenue was 82% of its 2025 software revenue.
3. Consolidation is reshaping choices: Synopsys completed its acquisition of Ansys in July 2025, and Siemens completed its purchase of Altair for an enterprise value of about $10 billion.
4. Engineers are early AI adopters but cautious: in [Digital Engineering 24/7’s 2025 reader survey](https://www.digitalengineering247.com/article/engineering-technology-outlook-2026/features), 33% already used AI and 56% said CAD assistants would benefit design work most.
5. Starting is cheap, so the shortlist matters: [Onshape](https://www.onshape.com/en/pricing) offers a free plan for non-commercial use and up to 6 months free on its Professional plan for qualified users, so engineers can try several tools before anyone talks to sales.

## Who chooses engineering software, and what is a customer worth?

Engineers pick the tool they want; managers, CAD administrators, IT and resellers decide how many seats and on what terms.

The buying group usually includes the design or simulation engineer who will use the software, an engineering manager who owns the budget, a CAD or PLM administrator who worries about data and file compatibility, IT and procurement, and often a reseller. Autodesk, for example, calls its channel partners Solution Providers. In Digital Engineering 24/7’s survey of 194 readers, the largest group (26%) described their role as product or system design engineering.

The money sits in recurring revenue and expansion. Public results show the scale:

| Company and period | Figure |
|---|---|
| Autodesk, fiscal 2026 | $7,206 million revenue, up 18% |
| Autodesk manufacturing products, fiscal 2026 | $1,379 million, up 16% |
| Dassault Systèmes, 2025 | €6.24 billion total revenue |
| Dassault Industrial Innovation (CATIA, SIMULIA, ENOVIA), 2025 | €3.13 billion, 56% of software revenue |
| Dassault Mainstream Innovation (including SOLIDWORKS), 2025 | €1.43 billion |

At the entry level, the seat price is public for some tools. Onshape lists its Standard plan at $1,500 and its Professional plan at $2,500 per user per year. A team of engineers on an annual plan, renewed for years and expanded into data management or simulation, is worth far more than the first seat. That expansion path, not the first sale, is where AI visibility pays back.

## Where does AI already sit in how engineers pick tools?

At the start: engineers and software buyers ask chatbots for options, then test them hands-on.

There is no public survey of how engineers use AI to choose CAD or simulation software. Two pieces of evidence come close.

First, engineers are using AI in their work. Digital Engineering 24/7 found 33% of its readers already used AI, generative AI or machine learning, and 29% planned to within two years. Asked where AI would help most, they chose CAD assistants (56%), AI-supported simulation (48%) and generative CAD (45%).

Second, software buyers in general are using chatbots to find products. In [G2’s 2026 Buyer Behavior Report](https://sell.g2.com/2026-buyer-behavior-report), more than 80% of buyers had sourced software recommendations from an AI chatbot in the last two years. That survey covers all business software, not engineering tools, so we treat it as a direction, not a measurement for this category.

Simulation and PLM are already widespread among engineers, which keeps switching questions alive. In the same Digital Engineering 24/7 survey, 48% of readers used simulation software and 30% used PLM.

## Which questions do engineers ask AI about software?

Workflow questions, comparison questions and “what now” questions after a vendor changes hands.

The questions below are ours, written to show the pattern. They were not captured from any assistant or user.

| Situation | Example question |
|---|---|
| Choosing a first tool | “Best CAD for a three-person hardware startup doing sheet metal and small assemblies” |
| Comparing | “Fusion vs SOLIDWORKS vs Onshape for a small team that needs data management” |
| Specialist analysis | “CFD software for electronics cooling that can run in the cloud” |
| Regulated work | “PLM for a medical device company that needs FDA design controls” |
| After an acquisition | “What changes for Altair license holders now that Siemens owns Altair?” |
| Replacing a tool | “Alternatives to our current FEA package that read our existing models” |
| On a budget | “Cheapest professional CAD that exports STEP and has a free trial” |

The last row matters more than it looks. In [our prompt phrasing study](https://underneath.agency/research/ai-prompt-phrasing-study), asking the same question again kept the same first brand 68.0% of the time, but adding “on a tight budget” kept it only 15.3%. Engineers who add a constraint such as a budget, team size or file format can get a different shortlist, so your facts need to cover those constraints.

Acquisitions create a wave of these questions. [Synopsys completed its acquisition of Ansys](https://www.airframer.com/news/release/synopsys-completes-acquisition-of-ansys) in July 2025 and said it was positioned to win in an expanded $31 billion addressable market. [Siemens completed its acquisition of Altair](https://www.digitalengineering247.com/article/siemens-completes-acquisition-of-altair/news) for about $10 billion. Customers of both are asking about roadmaps, licensing and alternatives.

## How does an AI answer become seats and an enterprise agreement?

Through a shortlist, a free plan or trial, a team pilot, a purchase and then expansion.

1. **Shortlisted.** An engineer asks for tools that fit a workflow. The answer usually offers a handful of CAD or simulation tools.
2. **Tried.** Engineers download a trial, use a free plan or start an education license. Onshape, for example, offers professional-grade CAD free of charge to students and educators and no-cost licenses to qualified startups.
3. **Tested on real work.** The team checks a benchmark part, file import, data management and performance on its hardware or in the cloud.
4. **Bought.** Seats are purchased directly or through a reseller.
5. **Expanded.** More seats, data management, simulation and PLM follow, sometimes under an enterprise agreement.

An assistant can help with the shortlist and the trial; your product, documentation and resellers win the last three. A tool that never makes the shortlist never gets the trial.

## Why does an assistant recommend one engineering tool over another?

It recommends tools whose facts it can find and confirm; the platforms document their searches, not their picks.

**What the platforms document.** According to [OpenAI](https://help.openai.com/en/articles/9237897-chatgpt-search), ChatGPT search turns a question into one or more targeted queries for its search providers, and only sites that let OAI-SearchBot in are eligible to appear. [According to Google](https://blog.google/products/search/ai-mode-search/), AI Mode fans a question out into multiple related searches on its subtopics, a technique the company calls “query fan-out.” Neither publishes how a CAD or simulation tool gets chosen.

**Observed in our studies.**

- Before answering a software buyer’s question, ChatGPT ran a mean of 3.7 searches in [our hidden-searches study](https://underneath.agency/research/ai-hidden-searches-study). When one of those searches named a source, the answer cited that source 44.0% of the time, against 8.1% when it did not.
- Outside coverage was the strongest predictor in [our brand entity study](https://underneath.agency/research/brand-entity-ai-recommendations-study): every tenfold rise in the number of independent sites that named a brand in the cited pages came with 4.7 times the odds of being recommended.
- In [our study of self-promoting lists](https://underneath.agency/research/self-promoting-best-lists-study), 24.2% of AI-cited “best X” lists with an identifiable publisher ranked their own publisher first. Many “best CAD software” pages are written by vendors, and engineers know it.

**Our inference for engineering software.** The facts an engineer checks are concrete: supported workflows (part, assembly, drawing, sheet metal, CFD, FEA), file formats and kernels, operating systems and cloud options, licensing terms and what changed after an acquisition, plan limits, and training resources. Independent sources that confirm them include trade publications, user forums, review platforms, university courses and benchmark reports.

## What does an engineering software vendor lose when AI leaves it out?

Trials it never sees, from engineers who will standardize a team on whatever they tried first.

We have no measurement of trials lost to AI absence, so this reasoning is our own:

- **Trials are self-serve.** With free plans and long trials common, an engineer can try three tools in a week without contacting a vendor. A tool missing from the shortlist is not compared at all.
- **Habits last.** We infer that students and hobbyists who learn a tool on a free or education plan often carry it into work, so early AI recommendations shape which tools a generation of engineers knows.
- **Switching windows are short.** Acquisitions and pricing changes push customers to look around for a few months. A rival that is named during that window can win accounts that were not for sale before.
- **Wrong facts cost deals.** An answer that says your tool lacks a feature, cannot read a format or has no free trial removes you from evaluations. When an answer gets a file format or a plan wrong, our guide to [fixing wrong brand information in AI answers](https://underneath.agency/resources/fix-wrong-brand-information-in-ai-answers) shows how to find and correct the page behind it.

## How does GEO work for a CAD, simulation or PLM company?

Generative engine optimization (GEO) makes your software easy for AI assistants to find, describe accurately and verify.

For an engineering software company, the work usually covers:

1. **Workflow pages in plain text.** One page per job the software does, with capabilities, limits and example models, not just a feature grid.
2. **Interoperability facts.** Supported file formats, geometry kernels, import and export limits, and integrations with data management and simulation tools.
3. **Plan and licensing clarity.** What the free, trial, education and startup options include, and what changes on paid plans, stated where a crawler can read it.
4. **Honest comparison and migration pages.** How you differ from the tools engineers compare you with, and how to move models across. Our review of [whether comparison pages help B2B brands get cited](https://underneath.agency/resources/do-comparison-pages-help-b2b-ai-citations) covers what works and what does not.
5. **Independent proof.** Trade press, benchmark studies, user forums, university courses, review platforms and conference talks. Our article on [best-of lists and AI recommendations](https://underneath.agency/resources/best-of-lists-ai-recommendations) explains why independent lists carry weight.
6. **Consistent reseller and partner pages.** Resellers often publish their own pages about your product; keep product names, plans and capabilities consistent across them.
7. **Crawl access and measurement.** Allow the documented search crawlers, and ask a fixed set of workflow, comparison and budget questions across ChatGPT, Gemini, Perplexity, Copilot and Google’s AI features over time, then compare with trial signups.

Nobody can promise that an assistant will recommend your CAD, simulation or PLM tool. The goal is to make it the easiest one for an assistant, and then an engineer, to check. Developer tool companies sell to a similar technical audience, and our article on [how developer tool companies win users when developers ask AI first](https://underneath.agency/resources/developer-tools-ai-search) shows how that plays out.

## Which parts of this can’t the data show yet?

How often engineers pick tools from AI answers, and how many seats that produces, has not been measured.

- **No engineering-specific buying data.** The chatbot figure here covers all business software buyers. We found no public survey of how engineers use AI to choose CAD, simulation or PLM tools.
- **Sources have interests.** Vendors report their own results, CIMdata sells research to the PLM industry, and G2 runs a review platform.
- **No category ranking studies.** Our studies covered buyer questions across many categories, not engineering software, so their relevance to CAD, simulation and PLM buyers is our inference.
- **Seat economics are private.** Public results show revenue and some list prices, not what a typical team spends over its lifetime.

## Where should an engineering software company start?

Ask assistants the workflow and comparison questions engineers ask before a trial, and see whether your tool is named.

That first check usually shows whether your software appears for its core workflows, whether its formats, plans and limits are described correctly, which publications, forums and lists the answers rely on, and which tools are recommended instead.

If your growth depends on more trials turning into team seats and enterprise agreements, [get in touch and we will test how assistants recommend your software](https://underneath.agency/contact). We put the workflow, comparison and budget questions engineers ask to the main assistants, show which rival tools, forums and lists win those answers, and set out the fixes most likely to bring more qualified trials. How the ongoing work runs for a CAD, simulation or PLM vendor, from workflow pages to reseller consistency, is laid out on our [generative engine optimization service](https://underneath.agency/services/generative-engine-optimization) page.

## Frequently asked questions

### Do engineers trust AI recommendations for software?

They use them as a starting point. Engineers test tools on real models before committing, so AI decides what gets tried, not what gets bought.

### Should an engineering software company list its plan prices publicly?

Where you can, yes. Clear plan and licensing facts let an assistant answer budget questions accurately instead of guessing or leaving you out.

### Do vendor-written “best CAD software” lists help?

They can be cited, but self-ranking lists are easy to spot. Independent reviews, forums and trade coverage are stronger confirmation.

### How should we handle questions after an acquisition?

Publish a plain page on what changes and what does not for customers: licensing, support, roadmap and file compatibility. Customers and assistants both look for it.

## Sources

- CIMdata via Industrial Machinery Digest (2026-06-04), [CIMdata Publishes Executive PLM Market Report](https://www.industrialmachinerydigest.com/software/quality-management-software/cimdata-publishes-executive-plm-market-report)
- Autodesk (2026-02-26), [Autodesk, Inc. Announces Fiscal 2026 Fourth Quarter Results](https://adsknews.autodesk.com/en/?p=55945)
- Dassault Systèmes (2026-02-11), [Q4 revenue growth of 1% with solid operating margin and EPS expansion](https://www.3ds.com/assets/invest/2026-02/dassault-systemes-25q4-earnings_pr_va.pdf)
- Synopsys via Airframer (2025-07-17), [Synopsys completes acquisition of Ansys](https://www.airframer.com/news/release/synopsys-completes-acquisition-of-ansys)
- Digital Engineering 24/7 (2025-03), [Siemens Completes Acquisition of Altair](https://www.digitalengineering247.com/article/siemens-completes-acquisition-of-altair/news)
- Digital Engineering 24/7 (2026), [Engineering Technology Outlook 2026](https://www.digitalengineering247.com/article/engineering-technology-outlook-2026/features)
- Onshape (n.d.), [Onshape pricing](https://www.onshape.com/en/pricing)
- G2 (2026), [2026 Buyer Behavior Report](https://sell.g2.com/2026-buyer-behavior-report)
- OpenAI Help Center (2026), [ChatGPT search](https://help.openai.com/en/articles/9237897-chatgpt-search)
- Google (2025-03-05), [Expanding AI Overviews and introducing AI Mode](https://blog.google/products/search/ai-mode-search/)
- Underneath (2026), [Does rewording a question change AI brand recommendations?](https://underneath.agency/research/ai-prompt-phrasing-study)
- Underneath (2026), [The hidden searches AI assistants run before they answer](https://underneath.agency/research/ai-hidden-searches-study)
- Underneath (2026), [Do Wikipedia and schema make AI assistants recommend a brand?](https://underneath.agency/research/brand-entity-ai-recommendations-study)
- Underneath (2026), [How many “best of” lists cited by AI rank their own brand first?](https://underneath.agency/research/self-promoting-best-lists-study)

---

This is the Markdown twin of https://underneath.agency/resources/engineering-software-ai-search. The HTML page is canonical. Publisher: Underneath, https://underneath.agency/agent. Site index: https://underneath.agency/llms.txt.