By Chris Burgess

Why I Stopped Telling People What Data to Use

When I started building Intuifi, I assumed the biggest problem companies faced was fragmentation. I thought they were struggling because their customer feedback was spread across different tools and conversations.

My thinking was simple: bring all of your customer feedback into one place, and you would get a more complete view instead of just hearing from the loudest customers.

But speaking to early users like Kobai completely challenged that assumption. Centralising data was not the problem they were most concerned about. They already had a source they considered valuable. The problem was making sense of it.

They suggested starting with their GitHub backlog.

That backlog was useful because it was not just a record of engineering work. People across the company could add ideas to it, whether they came from sales, engineering, or elsewhere. It contained customer signals, but it was not exclusively customer feedback.

The problem was making sense of what was already there

Kobai's backlog contained hundreds of records. The information was valuable, but turning it into something useful for decision-making meant manually organising it, grouping similar ideas and working out which themes had enough evidence behind them.

That changed how I understood where Intuifi could be useful. I had assumed I could define which data companies should analyse first. In reality, that depended on where they already had information worth understanding.

You may already have a sense of which issues are recurring in your department, which are becoming more serious and what is going well. The difficult part is establishing how much evidence sits behind that judgement.

To do that manually, you have to find the relevant records, work out how often each issue appears, and check whether the examples actually support the conclusion you have reached. Reviewing a small number of records is manageable. The problem becomes harder when the information is spread across hundreds of tickets. Finding out if a problem is genuinely widespread means searching through the source material, grouping similar comments, and keeping track of the records behind each conclusion.

Manual review can still produce a useful picture, but it can take a significant amount of time, and the level of support behind each conclusion may vary.

Testing it on what you already know

That experience helped me see Intuifi differently. The value was not dependent on aggregating everything first. It could start with analysing one source the company already considered important. Intuifi identifies the themes within that material and keeps each finding connected to the source records that support it. This allows you to distinguish a widespread problem from one that only felt widespread because a few examples were particularly memorable.

A useful way to test the analysis is to start with material you already understand. You can compare the themes Intuifi identifies with your own conclusions and inspect the records behind each finding.

That is exactly what Katy Parkhill, Head of Engineering at Kobai, did when she tested Intuifi against a backlog she had already analysed manually. As she told me:

"When I first joined the company, I needed to go through the entire backlog of over 400 tickets to decide what mattered and what didn't. The themes Intuifi found matched very closely with my own analysis, work that had taken almost six weeks was done in five minutes."

A clearer picture for the department

Once you are satisfied that the output reflects the material you know, you can use Intuifi to examine information you have not had time to review in the same depth. It may identify patterns that are difficult to see record by record, or show that the evidence behind an assumed problem is weaker than expected.

The immediate benefit is a clearer view of your own area of the business. You can investigate recurring problems earlier, decide what deserves attention, and avoid over-prioritising issues that are not reflected across the wider evidence.

Crucially, the judgement still sits with you. The analysis can show what is appearing in the data, but it cannot decide what the business should do about it. Your role is still to provide context, weigh the implications, and choose the response.

The difference is that you can spend less time assembling the picture and more time acting on it.

A useful place to start is with a source your team already understands. Run it through Intuifi, compare the findings with your own view and inspect the evidence behind them. You can then judge for yourself how much of the manual work it removes.