A Windsor-area automotive supplier
A tier-two automotive supplier with three plants in Southwestern Ontario, serving the cross-border Windsor–Detroit corridor. Around 350 employees, mid-eight-figure revenue. The leadership team had been approached by a US-based AI vendor offering to automate the request-for-quote process, pitched at $420k for the first year plus a multi-year licence. The team had no in-house AI experience and no obvious way to evaluate the claim.
The vendor demo had been polished and persuasive, run on the vendor's own data with the vendor's own RFQ template. The leadership team could see the pitch worked beautifully in the vendor's room. What they could not see, and could not get an honest answer on, was whether it would work on their actual RFQ documents, which were a mix of PDF email attachments, scanned faxes, and internal Excel sheets accumulated across three plant managers' idiosyncratic conventions. The clock was running: the vendor wanted a signed letter of intent inside two weeks.
We ran a half-day AI Demo built on twelve of the client's real RFQ documents, picked at random from the most recent quarter. The prototype was deliberately simple, an off-the-shelf model with a thin custom wrapper. The session put the working prototype in front of the leadership team and the three plant managers together. We walked through what it handled cleanly, where it failed quietly, and what production-grade software would actually need to cover the gaps. A second engagement two weeks later was a 90-minute AI Webinar for the wider engineering team, framing what AI does and does not do for this class of problem.
The leadership team declined the $420k vendor contract. Their in-house engineer subsequently built a more modest internal tool, using the demo prototype as a starting point, for roughly $20k of effort, focused on the three specific document types that drive 80% of their RFQ volume. The tool has been in production for three months. Of the two further questions the leadership team has had about AI investments since, both went through the same demo-first pattern before any vendor signature.
Published April 18, 2026. Sector and figures are generalised where necessary to keep the client comfortable being represented here. The structure of what we did and what changed is exactly as it ran.
Tell us the situation on a free first call. We will say which engagement family fits and what it would cost. If none of our offers fit, we will tell you that too.
Schedule a free consultation