Strategic data, faster
Instead of digging through reports and dashboards, the manager just asks.
B2B · AI · HR Technology
TOTVS · 2026
Here I share how I helped a brand-new AI product find its direction, and how that turned into a clear roadmap and teams working together.


My Role
In this project, my work was:
Overview
Let me start with what Copilot is. It's an AI that lives inside HR platforms and helps managers in their day-to-day. Think of a manager who needs to understand how their team is doing, or make a decision about someone, and doesn't have time to go through a bunch of reports. That's where Copilot comes in.
When I joined, the product was very new and a lot was still open: who it should serve first, which problem it solves best and how it becomes a business. So my work started right there: helping model the business, understanding the opportunities and looking at what the data was telling us.
Instead of digging through reports and dashboards, the manager just asks.
The answer comes with context, right when the decision needs to be made.
More information about each person on the team, so the manager can lead people, not just processes.
The Challenge
The question that guided me
How do we turn a manager's first try into a habit they don't want to give up?
People were already trying Copilot. What I needed to understand was why they weren't staying, and help the whole company agree on what the product had to become to earn a place in a manager's routine.
What the data showed
When I looked at the product data, one thing stood out right away: people were finding Copilot and trying it, but they weren't coming back.
Activation
Managers find Copilot and give it a try. The curiosity is there.
Retention
That first use doesn't become a habit. They try it once and don't come back.
Our reading
We believe what we deliver today is still too shallow to become part of a manager's day.
That changed the question for us. The problem wasn't getting managers to try Copilot. It was giving them a reason to come back.
Discovery
A new product can't find its way inside just one team. So I led a discovery with business, product, technology and sales, so we could decide together where Copilot should go.
Viability
The business model, the positioning and what makes the product sustainable.
Value
The managers' real problems and where Copilot actually helps.
Feasibility
What we can really do with AI and the data inside the platforms.
Market
What customers ask for, what makes them hesitate and what sells.
Method
First, everyone looked at the same diagnosis: activation without retention.
Then each area shared where it saw Copilot creating real value.
Together we decided what the product is, who it serves first and what it shouldn't try to be.
Finally, that direction became a short roadmap with clear priorities, not a wish list of features.
Impact
This discovery had a direct impact on business decisions and on the product roadmap. In practice, we came out with:
A vision everyone shares, which helps us decide and say no to what doesn't fit.
Fewer bets, but deeper ones, focused on the value that brings managers back.
Business, product, technology and sales speaking the same language and rowing in the same direction.
Copilot is still at the beginning of its story, and honestly, that's the most exciting part. Instead of piling on features to chase usage, we now have a shared north and a roadmap focused on depth: delivering value strong enough to bring managers back.
For me, this project made one thing clear: in a new product, my work goes way beyond screens. It's about reading the signals, bringing the right people together and helping the company decide what to build, and what not to.