I want to write down what we are actually about at computerlove.tech. Not because anything changed, but because being vocal about a vision is how it starts to live in other people's minds, and I would rather have people know exactly what we stand for than have them guess.
So here it is: our vision is to enable knowledge workers to securely delegate work to AI agents.
That sentence is short, but every word in it is doing work, so let me unpack it. Between that vision and reality we have identified two bottlenecks, and everything we do as a company is ideally addressing one of the two.
The first bottleneck is skill
We have spent the last year teaching developers how to develop software using AI agents, around 600 of them by now, and a year is forever in this industry. What we see every single week is the same struggle. The hard part is not prompting, and it is not the tools. The hard part is the delegation itself.
Developers are used to being individual contributors. They do the work with their own hands, and their instincts are tuned for that. What agents demand of them is something else entirely: clarifying the task, planning it, specifying it, structuring the work, and caring about the environment around the agent so it is set up to do great work. These are leadership activities. A developer working with agents is going from individual contributor to leader, a leader of agents, and that shift is genuinely weird for most people. It is a new way of working, and nobody learns it by reading about it.
And now this way of working is arriving for everyone. With tools like Claude Cowork, delegating work to agents is becoming available to all knowledge workers, not just developers. They will have to learn the same techniques the developers are learning now, but they are not as technical, so they cannot build their own scaffolding and wire everything together the way developers can. Which leads directly to the second bottleneck.
The second bottleneck is infrastructure
Agents are most effective when they can reach the tools you actually do your job with: your email, your calendar, the CRM, Slack or Teams, and all the systems you touch on a normal working day. Most of that software was built for humans clicking around in interfaces. It is not legible to agents, and there is a huge amount of infrastructure work to be done before it is, in connectivity, in making systems readable and operable for agents, in everything around it.
But access cuts both ways. The same access that gives a knowledge worker superpowers can also let an agent do things it was never supposed to do. My favorite example is maybe a little sci-fi: HAL 9000 from 2001: A Space Odyssey is basically an AI agent with access to the spaceship's systems, including the life support of the crew in hibernation, and we all know how that goes. The realistic examples are less cinematic but the same shape. An agent accidentally sends an email containing information that should have stayed internal. A customer support agent has access to data across users, and suddenly one customer can query another customer's data. A lot of this is good old permissioning, deciding what you grant an agent access to and what you don't, but agents introduce a whole new class of security questions on top of it.
This balance between superpowers and security is exactly the concern we hear from leaders, again and again. If you want delegation to actually happen inside an enterprise, security is not a feature you bolt on afterwards. It is the thing that makes the rest possible. That is why the word securely sits in the middle of our vision and not in the footnotes.
What we are doing about it
Our strategy has two legs, one for each bottleneck.
The first leg is upskilling. We teach workshops for software developers working with tools like Claude Code, Codex and Cursor, and for general knowledge workers working with tools like Claude Cowork. Teaching makes us a living, so we do not have to raise funding, which is nice, but it does more than that. Every workshop is also discovery, because we learn an enormous amount about the market every time we are out talking to people about where they actually struggle. And honestly, we are just passionate about helping people learn this technology.
The second leg is building product in the AI agent infrastructure space. Our current thesis is a single gateway that enterprise employees and AI agents connect to, and that gives every employee and every agent secure, governed access to exactly the systems they have been granted, with fine-grained control on every single endpoint. I say thesis deliberately. We have not built it out in the industry yet, and we are humble about this being a new space. We know the vision and we know what we want to enable, but we would rather discover the right solution by implementing real things for real customers than settle on whatever we imagined from the office. That is what we are most eager to do right now: get out there and start solving real technical problems for real organizations.
Why us
You could say the space of governed agent access is getting crowded, and on paper it is. But we keep meeting companies that struggle with exactly this, and if there were a clear solution they could just go and pick, they would have picked it. They haven't.
We also think we bring a perspective that is hard to copy. We have been teaching this way of working for a year, to hundreds of developers, and we have watched up close what people actually struggle with in organizations when they try to delegate work to agents. Now the same shift is reaching every knowledge worker, and those workers will need the infrastructure built for them.
And there is one more thing our customers keep asking for: a neutral third party. They do not want to lock themselves to Anthropic, OpenAI, Microsoft or any of the hyperscalers. They want a solution that is independent of any one tool and any one agent, from someone whose only interest is making the delegation work safely. We can be exactly that.
What I want you to take away
I want this picture to exist in your head: computerlove.tech is the company that wants to enable knowledge workers to securely delegate work to AI agents, and we currently believe that the most effective way we can help companies get there now is by upskilling developers and knowledge workers, and by building the infrastructure that makes delegation safe. If that picture is useful to someone in your network, pass it on.
And if you are working in this space yourself, whether you are standing in front of these problems inside your organization or building things to solve them, reach out and connect. We want to talk to the people who are wrestling with the same questions, because that is how we make progress: by solving real problems together.