Learning Systems: Turning Feedback Into Fuel

One of the biggest shifts in our innovation work this year has been rethinking what “feedback” actually means in a large customer organisation. We don’t run endless pilots or lab experiments. Most of our learning doesn’t come from slick journey tests or carefully moderated interviews. It comes from something far more honest: data, live chat transcripts, operational friction, and the unfiltered things customers tell us when they need help.

The insights are there - thousands of them - but they’re scattered. Data lives in one place. Contact drivers live in another. Operations sees one side of the story, product teams see another, and live chat verbatim sits quietly in the corner holding the truth no one has time to read.

For a long time, each team translated feedback through its own lens. Analysts looked at trends. Service teams looked at pain. Product looked at journeys. Everyone was right, but nothing connected.

So we started treating learning as something we build deliberately, not something we collect passively. Instead of creating more dashboards or documents, we focused on behaviour: capture what we now understand that we didn’t before, and make sure it’s visible beyond the team that found it.

The moment this started to click wasn’t a glamorous pilot. It was a spike in contacts on a specific issue that didn’t line up with what the data suggested should be happening. Live chat transcripts told a very different story - customers describing workarounds, expectations, and pain points in language none of our dashboards were designed to surface. Once we shared those insights with other teams, it reshaped the way they viewed their own problem space.

It wasn’t a breakthrough. It was a realisation: learning happens when someone discovers something useful, and someone else actually uses it.

Since then, we’ve treated feedback as a shared organisational asset. Not as “noise,” not as tickets, not as commentary - but as evidence that needs to flow. When someone shares a short summary of what they’ve learned, it isn’t documentation; it’s propagation. It stops the knowledge evaporating when teams change shape, and it means the next decision doesn’t start from scratch.

This doesn’t require complex tooling. It requires rhythm and discipline. Capture the learning. Tag it. Share it. Connect it to what’s next. That’s the whole system.

In a customer environment as wide, busy, and interconnected as streaming, insights surface every single day - in data patterns, in operational anomalies, in the raw words customers use when something doesn’t work. The trick isn’t to catalogue everything; it’s to make sure every discovery leaves a trace the organisation can build on.

That’s how teams get smarter without having to run more experiments than they need to.