OUR VISION
The Agent Experience
For a decade, the best teams obsessed over how users experience their products. The next decade belongs to those who obsess over how agents experience their products.
Some of the best software products we know today came from teams obsessed with building useful, intuitive products with an amazing user experience (UX). Companies such as Airbnb, Uber, and Spotify revolutionized entire industries solely by building a good UX.
But what has worked for the past 10 years won’t continue working for the next 10. Every day, we see more and more traffic to products and services coming not from human users, but from AI agents.
Agents are quickly becoming the primary users, consumers, and contributors of software products.
At AgentCat, our belief is that the next wave of revolutionary products will come from teams obsessed with building products for agents.
We call this building a good Agent Experience (AX): a new term for how easy and intuitive a product is for agents to navigate. And sometimes, AX and UX are at odds.
Good UX and good AX pull in different directions
GOOD UX
Humans have limited attention spans and easily get distracted.
Humans care about information density and typography to communicate structure.
Humans get overwhelmed with walls of text or confusing buttons.
GOOD AX
Agents don’t have eyes and don’t see on their own.
Agents don’t care about typography, buttons, or visual hierarchy.
Agents prefer structured data, clear tools and documentation, and no fluff.
Already today, companies are adding llms.txt to their products and turning web pages into Markdown, making it easier for agents to parse information.
MCP has quickly gained wide adoption with thousands of companies from startup to F500 building an agent-first experience into their products. And standards like WebMCP are building on top of that momentum to bring the power of tools and function calling to the entire web.
MCP is quickly becoming the backbone for companies rearchitecting their products around a good AX.
Agents don’t see. They prefer structured data, clear tools, and no fluff.
Why we’re building AgentCat
Existing analytics and observability products don’t help teams understand how agents use their products. They provide logs and traces, and track button clicks and form inputs.
Those are all great signals if your primary users are humans.
Teams building agent-first products need an agent-first solution. Our background is in cybersecurity and building AI tools that augment what’s possible.
Now, we’re focused on empowering the next wave of products built for agents.
AgentCat is the analytics and observability platform to help you understand exactly what agents are doing in your product, what they’re trying to achieve, and where they’re getting stuck.
What they’re doing
Replay every agent session, end to end.
What they’re trying to achieve
Classify intent and goals automatically.
Where they’re getting stuck
Surface errors and drop-off points.
Built fully open-source to help teams ship Claude Connectors and ChatGPT Apps. A single platform to improve your AX and ship the next generation of products built for agents.