I started a business with AI and no tech background. Here is what still fails to compensate.

This talk-show article is based on an interview with Tim Desoto, a 49-year-old founder and CEO, who lives in San Francisco. The following has been edited for length and clarity.

I have been working on my startup, an AI-powered shopping platform, since the end of the year 2024. With all the progress in the tech industry, people can think of AI as a hammer, treat everything else as a nail, and in my experience, that doesn’t work.

I don’t have a technical background, and since starting my business, I’ve learned a lot about where to optimize AI, and where not to. I always try to be flexible about when to transition from AI to human intervention, and vice versa.

Over time, it became clear to me where and where not to use AI for my business.

I rely on my network and social networks to help me decide which AI tools to use

I am lucky to live in San Francisco. I go to conferences to hear what people are using, and I’ve attended some developer conferences. My ears always perk up when I hear that a new tool or version of something is working for someone in my network.

Currently, there is a strong focus on agency workflows. OpenClaw created buzz as an open-source autonomous agent project, and Moltbook expanded that attention by making agent-to-agent visible in the social community. Claude Cowork is also gaining traction, especially among teams looking for a business-ready travel agent with clear security.

Beyond the agents themselves, the focus is moving from “what can agents do?” that “how do we run them reliably and safely at scale?”

Whether I’m looking at X, LinkedIn, or other platforms, there’s a lot of work being done to share these updates.

I use AI for all standard timekeeping tasks

My paid stack includes business models such as Claude Max, Gemini Ultra, and ChatGPT Business, as well as AI-powered development and production products such as Cursor, Figma Make, Notion AI, Superhuman Ask AI, and Necessary.

The images of Gemini have been amazing. The latest new models have really improved everything. I’ve seen quick reaction, stable thinking, and strong multimodal abilities, especially in graphic design. I was impressed with how well the images held up during processing. I’ve already seen improvements in the responses it provides for real-time information.

I define what tools I’m using based on what has just happened, and I use them either made or as I think I can use them in my current flow. For example, I use Needed, an AI website builder, to create slide decks.

My ‘AI conveyor belt’ helps me test different AI models

Usually, I start by writing a prompt, then go multimodal, talking out loud to the model. I will talk back and forth with it about my strategy and try to get the agent to push back because I know that other AI models tend to be more consistent.

Once I get an output I’m happy with, I use another model to get a different perspective. For long-form analysis and structured views, I lean towards Claude and Gemini. Gemini’s inline source links are particularly useful for analysis and in-depth research. For organized thinking and organized writing, I mainly use ChatGPT and Claude.

Sometimes, I push a document out to multiple models at the same time and see what comes back at the same time. For design analysis and multimodal work, I use both Gemini and ChatGPT to create early-stage concepts, mockups, and visual inputs. Some models are better than others for certain tasks, but I always get a well-rounded view by feeding content to multiple models.

The process can take as little as 15 minutes, or more advanced options can take several hours to several days, depending on the complexity.

I came to love human designers

When I was vibe coding the alpha version of my product, I would hit places where 30 or 40% would be wrong. I didn’t know what was the problem. I would have multiple screens running the code to check it, and I would continue to use the AI ​​against the AI ​​until I could get to about 95% confidence.

I contracted a few developers to help move my product forward. Now I have a product that grows faster, is more stable, and the scale grows.

As much as I can do with AI, it’s amazing what technical people can do with AI tools that technology can’t.

The personal perspective has not changed in my business

I reached out to many informal advisors and friends in the area who could help at the beginning of my process, but recently I developed having advisors to pull things off. That was a big feather in my head.

These consultants have their own expertise to draw on, and they know a lot of smart people in this area working on projects that have helped us see potential blind spots. It helped me connect with potential partners in ways that I think are more difficult as a solo entrepreneur.

I feel like I have a clearer view now of what I can trust in AI, compared to what I thought at the beginning of my journey. While issues such as guesswork and compromise can be mitigated, long-term control of length and taste still requires human attention.

AI can create opportunities, but choosing the right direction remains a human responsibility.

Do you have a story to share about running an AI-powered business? Contact this reporter, Agnes Applegate, at [email protected].