Cut report production from three weeks to minutes
Move two of four. I turned hand-made reports, our biggest drain on expert time, into an instant product feature, without losing the quality customers loved.
- Company
- Hive HR
- Year
- 2025
- Role
- Product Manager (solo)
- Team
- Data team, engineering, services team
- Product strategy
- AI in product
- Change management
At a glance
- The problem
- Customers loved our reports, but experts made each one by hand. It took up to three weeks, and the business couldn't grow past their capacity.
- My role
- Product owner. I led the product direction, the platform build and the coordination across four teams, in close partnership with the data team.
- The result
- Reports in minutes, quality held or improved, and expert time freed for higher-value work.
The problem
Our reports were a key reason customers chose us. But every one was made by hand, needed specialist tools, and took up to three weeks. The team making them was burning out.
Leadership had decided to stop making reports by hand, and the data team had automated parts of the process. But speeding up pieces of a manual workflow doesn’t change the model. I took ownership of turning reporting into a real product feature.
What I did
I made four bets:
- Reports should be instant, and customers could run them any time, rather than waiting weeks.
- AI could write some of the commentary, carefully. The numbers stayed calculated, never generated. AI added plain-English context on top.
- Analysis of written comments belonged in the product, not behind a paid service.
- Product first, consultancy where it adds value. Standard reports became part of the product. Bespoke work stayed available, priced separately.
Every team had a fair worry. Data didn’t want their methods running unchecked. Services worried about losing touch with customers. Engineering was wary of code only the data team understood. Leadership knew a loss of trust would sink the whole strategy.
So I named the risks openly and agreed guardrails: phased releases, clear ownership and small pilot groups. We led with the value in our messaging, not the AI.
The result
- to produce a report, down from three weeks
- Minutes
- or improved report quality, with easier sharing
- Held
- gross profit, the point of the strategy
- Software-led
Experts got their time back for coaching customers. Customers got independence, with consultancy still there when it helped.
It also showed us the next problem: a report in two minutes still has to be read and understood. That’s why Auto Insights exists.
Described the reporting as “phenomenal”, and said it had “blown my mind”. A university customer