Field note
We handed our competitor research to an AI
Six competitors reviewed by an AI assistant: what it found, the mistakes we had to fix, and a ready-to-copy request to do the same for your own company.
· 7 min read
Our competitor analysis had been sitting on the to-do list for months. Not for lack of interest, but because it means a full day of open tabs and copy-pasting into a spreadsheet. And by the time you finish, part of it is already out of date.
We spend our days telling clients that this kind of task can be handed to an AI assistant. It was time to do it for ourselves.
The test
We used NitroxBrain, our own assistant. Claude or ChatGPT would do the job too, as long as web search is switched on.
The request fit in a few lines:
My company: Nitrox Consulting, an AI consultancy in Montpellier (training, support, custom solutions) for small and mid-sized companies. Find 5 direct competitors in France, then add Limova.ai. For each one: offer, public prices, target clients, homepage promise, customer reviews, recent news. Build a comparison table and suggest 3 opportunities for us. Cite a source for every piece of information. If you can’t find something, write “not found”: don’t make anything up.
That last sentence matters more than all the others. We’ll come back to it.
What came out
The assistant picked two Montpellier firms, Foxpilot and Liftia, and three national ones: Masteria, Mister IA and Stema Partners. Plus Limova, which we had asked for. Data collected on 14 September 2026.
| Company | Where | What they sell | Published prices | Reviews |
|---|---|---|---|---|
| Foxpilot | Montpellier | Training, AI assessment, no-code automations (Make, n8n), consultants on assignment | None, except the Bpifrance assessment: €10,000 excl. VAT, 40% publicly funded | 95% satisfaction, no review count |
| Liftia | 30+ cities, incl. Montpellier | Certified training (Qualiopi), consulting, custom AI assistants | Not published | 4.9/5 on their site, 4.2/5 on Trustpilot (21 reviews) |
| Masteria | Lyon | Consulting, custom builds, certified training | Training at €1,980 excl. VAT per day, consulting on quote | 98% satisfaction |
| Mister IA | Paris | Consulting, training for companies and individuals | Training from €2,000, executive coaching from €5,000 (excl. VAT) | “5/5 on Google, 1,800+ reviews” |
| Stema Partners | Paris | Assessment, automation, business apps, training | Assessment from €10,000 excl. VAT | 4.9/5 on training |
| Limova | Nice | Subscription to 9 specialised AI assistants (phone, social media, SEO, prospecting…) | Unreadable on their site. €69.90 to €79.90/month excl. VAT according to third-party sites | 4.9/5 on Trustpilot (1,205 reviews) |
Everyone targets small and mid-sized companies, with nuances. Foxpilot also works with local authorities in the Occitanie region, Mister IA trains individuals through the French personal training account (CPF), Stema targets companies with 50 to 500 employees, and Masteria also works with large groups, as far as Switzerland and Belgium. Liftia and Limova speak more to freelancers and very small businesses.
The promises look a lot alike. Foxpilot opens with “Your teams lose hours on repetitive tasks. We’re changing that.”, Stema promises “from strategy to execution”, and Masteria calls itself “the AI consultancy that trains your teams”.
On the news front, the big story is Mister IA’s €10 million funding round in May 2026, with 120 employees announced. Limova announced a collaboration with OpenAI; on Trustpilot, its customers mostly complain about setup time and support that is sometimes slow.
What we take away for ourselves
First finding, slightly awkward: every consultancy on the list highlights Qualiopi certification and the public funding it unlocks (training funds, Bpifrance). We don’t mention it. For a small company, what’s left to pay often matters more than the sticker price.
Second finding, more encouraging: none of the five consultancies highlights an in-house assistant used every day, for real. They sell training, automations or consultants. Limova sells software, without custom support. With NitroxBrain, we sit between the two, and that is probably where we should put the emphasis.
Third finding: trust. Competitors show ratings, client logos, figures. Our site has no reviews and no price range. Three testimonials and a budget indication would be enough to close the gap.
Where we had to go back over it
The table above didn’t come out like that. It had to be reread, and we found quite a lot.
First, the assistant’s own mistakes. In its first summary, Montpellier ended up in the wrong region (Provence instead of Occitanie). Digging through the company register, it also came across an older Nitrox Consulting, created in 2001 and struck off in 2015, whereas the current company was registered in 2023. Read too quickly, that became “company closed”. It also mistranslated an activity code and a legal form.
Then, the inconsistencies we found in other companies’ pages while checking. On a single page, Mister IA claims 600, 750 and “more than 1,000” client companies. Stema claims 200 people trained on one page and 5,000 on another, gives two different price ranges for the same type of project, and still mentions Bpifrance funding “up to 25%” when it rose to 40% on 17 June 2026. Liftia shows “200 reviews” on its Montpellier page, “86 Google reviews” on its homepage, and 4.2/5 from 21 reviews on Trustpilot. As for Limova, its prices can’t be read on its own site, and third-party sites give two different price lists.
One more trap: Liftia has a page dedicated to Montpellier, which suggests a local agency. In reality, the company works in more than thirty cities.
The “cite your sources” rule didn’t prevent any of these mistakes. But it did make every one of them checkable in a single click, and that is the whole point.
The bottom line
The longest part wasn’t the research, it was the checking.
AI is very good at the tedious part: searching, sorting, comparing. Checking and deciding stay on our side, and that’s exactly how it should be. It’s the same rule as in the nine tasks that are easy to hand to AI: the assistant prepares, a person checks.
The request, ready to copy
Here is the generic version, improved with what we learned along the way. Replace what’s in brackets, paste it into an assistant that can search the web (NitroxBrain, Claude, ChatGPT…), and set aside time to check the result.
My company: [name], which sells [your products or services] to [your clients], in [city or region].
Find 5 direct competitors in [country or region]. Also add [the competitors you already know].
For each one, search the web for: what they sell, their public prices, their target clients, their homepage promise, their customer reviews (rating, number of reviews and source site) and their news from the last 12 months.
Present everything in a comparison table, then suggest 3 concrete opportunities for my company.
Rules: cite a link for every piece of information and give its date. If you can’t find something, write “not found”: don’t make anything up. Flag figures that contradict each other from one page to another, and check that a competitor presented as local really is.
Run the exercise again every quarter. Prices and public funding change fast, as we saw with Bpifrance.
Sources
- Nitrox Consulting: nitroxconsulting.com, French company register, societe.com
- Foxpilot: foxpilot.io, Bpifrance Data & AI assessment
- Liftia: liftia.ai, Montpellier page, Trustpilot
- Masteria: consulting page
- Mister IA: training page, FrenchWeb
- Stema Partners: AI agency page, homepage
- Limova: plans, Trustpilot, Entreprisma, AI-First, Formalive