Every product team reaches the point where people outside the business start using what you've built for real work.
We've recently reached that point with UnaGo. The feedback hasn't come through formal surveys or structured reports, but through messages, conversations and observations from people using the platform to solve genuine problems. That's been far more valuable because it shows how people behave when they're doing real work, with real expectations.
It worked first time
One of our earliest testers, working in education and language learning, sent us a message that simply said:
Everything worked! It worked perfectly.
It was a short message but reflected something we've noticed: most people approach a new AI tool expecting to spend time working around problems. They expect prompts to be rewritten, settings to be adjusted and outputs to need fixing before they're useful.
When something works on the first attempt it comes as a surprise. That first experience builds trust quickly. It encourages people to explore what else the platform can do, and it sets the standard we need to maintain.
Honest feedback is the most useful
Not every comment has been positive, and we wouldn't expect it to be.
A motion designer testing one of our creative workflows described the result as "almost perfect". The issue was specific: the workflow occasionally generated extra static images that needed cropping before the final assets were ready. They also pointed out that part of the behaviour came from the underlying AI model rather than UnaGo itself, which was an important distinction.
That sort of feedback is incredibly useful because it tells us exactly where to focus. It helps us separate issues in our own platform from the behaviour of the AI models we're orchestrating, and it highlights the small improvements that make the biggest difference in daily use. It's not just a complaint. It's a shared investigation into what happened, why it happened and how the workflow can be improved.
When users surprise themselves
A senior motion designer had been building an agent that could understand a design brief and automatically create multiple resized versions of a creative asset. After getting everything working, they sent us a message saying:
I can't believe I actually pulled this off.
The important part wasn't only that the workflow worked. It was that the user surprised themselves.
The best software doesn't replace people's skills. It gives them the confidence to tackle work they might previously have handed to someone else or decided wasn't worth attempting. It shifts the experience from "I need someone technical to build this for me" to "I can do this myself".
The conversation didn't end there. Having completed the workflow, they immediately suggested the next improvement: adding safe-zone templates so titles remained correctly positioned across different aspect ratios. One win led straight to the next practical improvement. That's exactly how good products evolve.
What early feedback really tells you
Reading between the lines of these early conversations, a few themes emerged clearly:
- Speed of trust matters. Users form opinions fast. When something works on the first real test, it builds confidence that compounds. When it doesn't, that doubt is hard to shake. Getting the core experience right, even before everything is polished, is worth prioritising above all else.
- Users become collaborators when they feel heard. Every person who flagged an issue also stayed engaged. They didn't churn, they kept testing, kept reporting, kept pushing. That only happens when people believe their input is going somewhere. Responsiveness isn't just good manners; it's a retention strategy.
- The edge cases are the roadmap. The artefacts, the cropping issues, the safe-zone gaps, none of these are blockers. They're a prioritised list of what to build next, handed to us by the people who care most about the product working well.
- Emotional signals are data. "I can't believe I pulled this off" is a product insight, not just a nice quote. It tells you that the capability felt out of reach before, and now it doesn't. That's a positioning signal, a marketing signal and a product-market-fit signal all at once.
Building in public, learning in real time
We're at an early stage, and we're not pretending otherwise. What we're building at UnaGo.ai, AI agents that handle real creative, operational and analytical work, is genuinely hard. The workflows are complex, the edge cases are plentiful, the expectations are high.
But the feedback we're getting tells us we're on the right track. Not because everything is perfect, but because the people using it are engaged, specific and coming back.
UnaGo.ai is an AI execution platform that helps teams build intelligent agents, automate complex workflows and ship faster.
If you're interested in being part of our early access programme, get in touch.