My background and platform shifts
Platform shifts are how I separate different eras in software. I’ve worked through three shifts: iPhone, cloud and now AI. Each one changed how software products get discovered and successfully used. With AI, both discovery and usage are moving from people to agents. That makes this shift the biggest one yet.
iPhone
The iPhone didn’t just bring apps. All websites had to be rebuilt for mobile. This brought work for digital agencies for a decade.
On the app side, I remember debates and years of development around whether HTML5 could be used to build mobile applications for both iOS and Android. It turned out native apps were the way to go.
In Finland and globally, there’s hardly a better example of a winner in this shift than Supercell. The iPhone and the App Store made it possible, and Supercell executed. With their full focus on a native, high-quality iPhone game, the gaming experience was nothing short of magical. The product was good, and their monetization worked as well.
Cloud
Then came the cloud, and AWS was the big story. Every startup I knew set up its databases there.
I joined a startup called Aiven that offered databases as a managed service in the cloud, including AWS. It turned out a lot of people wanted someone to take care of their Postgres for them, and Aiven quickly grew into a unicorn.
This shift had its own platform lessons. As a technology provider, you are always competing and collaborating with the hyperscalers: AWS, Google and Microsoft. Every channel has its own challenges and opportunities. Part of Aiven’s growth came from recognizing the paths where we had the best fit and focusing our efforts on the right partnerships.
AI
AI has always interested me. Still, I’m one of the ChatGPT-awakened. I got hooked on LLMs almost as soon as the first announcement tweet.
Shortly after, I joined an AI usage analysis startup and founded Symposium AI, an AI developer community and event series that quickly grew into a network of the top 500 AI engineers in the Helsinki region.
I started building my own AI products, and one lesson came quickly:
People don’t want to use your software. They want their agents to use it.
Why Agent Disco
I recognized the same reaction in myself when trying new tools. Commanding workflows across tools from one view makes an immense difference to productivity compared with switching between UIs. If my agent couldn’t access a tool, I wouldn’t use it at all.
So instead of fighting it, I shifted my full focus to MCPs and CLIs, and to figuring out the best ways for agents to find new tools. I became an early publisher on ChatGPT and other MCP directories. Like Supercell with iPhones and Aiven in the cloud, this shift already has its early winners, such as Supabase and Vercel. I researched why some were winning, implemented what I learned in my own products and somewhat naturally ended up helping others make the same shift and grow their agent discovery and usage.
This led me to found Agent Disco in the fall of 2026.
What I believe
Agents will have an ever larger impact on the entire software industry. They will choose and use the best tools for the tasks their users set. It’s extremely interesting to research and find the levers companies have to find their place and win in this market. Beyond the technical research, it’s often also an operational change for companies. That’s why we’ve started not only by giving technical insights but also by offering a strategic service that helps with execution. I’ve seen this create much more value, much faster, than a dashboard solution.
How I work
Even though the tech evolves, my way of looking at growth and business operations has not changed through the platform shifts.
- 1
Understand
Trace the work to a business result.
- 2
Find
Locate the constraint holding it back.
- 3
Act
Make a focused change with your team.
- 4
Learn
Measure. Pick the next one.