Tender Tech Talks 8: product development, customer feedback and AI | Brainial

Building AI for tenders means constantly chasing moving expectations

Product owners build the best solution for their users. But how do you do that when users’ expectations change almost quicker than you can build your tool? Teuntje Peeters is our product owner at Brainial. In our latest Tender Tech Talks, she explains how she ensures Brainial’s features stay perfectly matched to our client’s expectations. We’ve gathered her most important insights for you in this blog, so sit back and read on!

Last year's breakthrough is this year's baseline

The field of AI moves quicker than any other sector. And... So do clients’ expectations about AI solutions, Teuntje explains. “Last year, we launched a new AI proposal writer. It was a huge success, and customers instantly used it intensively. But now, a year later, that magic’s gone. It’s become table stakes – the bare minimum for an AI tool.

This rapid shift presents a unique challenge for product development. People now work with ChatGPT, Gemini, and Copilot daily, which has completely reshaped what they expect from AI. What once impressed them now feels obvious. The rules change while you're still building, and keeping up requires constant adaptation and a deep understanding of where the market is heading.”

Start small, validate, then scale

Those expectations translate into requests customers make. But not all of them make it into the final product. Teuntje receives ideas from all directions: customers, the customer success team, sales, and her own observations. Her approach? Start with an MVP and test it with select customers across different industries before scaling up.

"For example, predicting a question is easy, but predicting a question that actually makes sense is much harder," Teuntje explains. Customers want to reuse questions from previous tenders when the context is similar, which requires deeper research and validation. She also monitors usage statistics and support tickets to determine what needs improvement and what might no longer be relevant.  

Verifiability is non-negotiable

Beyond smart prioritization, there's one principle Teuntje never compromises on: verifiability. Users need to know exactly where information comes from. This wasn't always obvious, Teuntje notes. "In the beginning, customers were anxious. Would AI take over their work? Could they trust it?"

That traceability is crucial because we're not at a point where people can hand over their work completely to an agent. Brainial pre-fills information so users only need to verify it, which only works if they know where to check. On data security, Teuntje is equally clear: every customer has their own isolated environment, data never leaks between clients, and Brainial never trains on customer data.

What's next?

So, what's on Brainial's roadmap? "An improved task management system for seamless team collaboration, evaluation functionality that automatically extracts award criteria from tenders, and smart ways to reuse high-scoring content in future bids. These features build on the feedback loop that makes Brainial truly tender-specific rather than just another generic AI tool.”

Want to hear more from our tender specialists? Follow our (Dutch) Tender Tech Talks on Spotify and never miss an episode again!  

‍