28 Interviews Were Just the Start

From a hunch about customer feedback to a platform built on evidence from anywhere in the business.

Before I built Intuifi, I had a hypothesis. I wanted to know whether it was real, or just something I had convinced myself of. What follows is the method I used to go from a hunch to evidence, and from evidence to something that actually shaped the product.

The feedback filter

As companies scale past their first handful of customers, the person making product decisions moves further away from the people actually using the product. Feedback still gets captured, but it arrives filtered through sales, customer success, support tickets, and hallway conversations. Each of those filters is well intentioned, but none are neutral. By the time an insight reaches leadership, it has already been shaped by whoever passed it along.

To see whether this was more than my own experience, I interviewed 28 tech leaders, ranging from pre-seed startup founders to enterprise CEOs. Four themes stood out from those conversations:

• No structured process: 13 of the 28 explicitly described having no formal or consistent way of handling feedback.
• Nothing joined up: 10 described tools that did not talk to each other, or channels that never connected.
• Feedback that never got followed up: one participant had 129 unreviewed leads worth an estimated $1.8m that nobody had gone back to.
• One person's judgement decided what counted: individuals or founders were injecting their own bias into what got fed through.

The problem came up again and again, across companies of very different sizes. Not everyone saw it as urgent, though, and one CEO was openly sceptical that AI-generated insights could be trusted at all. But before writing production code, I needed an answer to two questions: could a system extract insights from scattered, unstructured information at all, and would the insights themselves be any good.

Testing against reality

I built a prototype that worked from manually exported data rather than live integrations, so I could test whether the analysis itself actually worked before building the harder infrastructure to connect it to anything live.

I ran my own interview transcripts through it first. Useful for debugging, but I already knew what I expected to find, so it only told me so much.

The first proper test came from Kobai. They gave me an export of their GitHub backlog, containing over 400 tickets of ideas, bugs, and requests. Their Head of Engineering had already analysed the backlog by hand, a process that had taken her weeks.

Running that exact data through Intuifi gave me a direct comparison against judgement that had already been checked against reality. It landed on the same conclusions she had, and surfaced one theme, around data ingestion, that she hadn't consciously grouped that way herself. You can see the actual output from that test on the sample report page.

The analysis Intuifi produced was strong enough to move the conversation from whether the tool worked to what it would take to run it live.

The realisation: it was never about feedback

That test taught me two things.

First, the pattern suggested a trust dynamic: CEOs often sit furthest from the underlying evidence, so AI output becomes more credible when someone closer to the data, like a Head of Engineering, verifies it against material they already understand.

Second, the product was never really about feedback. Kobai's GitHub backlog was not customer feedback in the conventional sense. It was messy qualitative data accumulated from across the business. The mechanism underneath did not care. It applied the same evidence-based structure, regardless of where the qualitative data came from.

This began to clarify where I saw Intuifi sitting in the market. Thematic targets customer experience teams to analyse feedback at scale. Productboard equips product managers to consolidate user requests into a roadmap. Dovetail supports UX research teams to turn transcripts into a searchable repository.

By contrast, Intuifi gives founders and executive decision-makers a clearer picture of what's actually happening across the business. Intuifi is not a feedback collation tool, it is a platform that surfaces evidence wherever it sits.

What this means for the future

I had to establish one thing first: that Intuifi could turn unstructured qualitative data into evidence leadership could genuinely trust.

Now that I have evidence the core analysis works, the next step is live integrations into the tools teams already use, so leadership always has a current view instead of one built from manual exports.