Hamza and his brother are building Olostep, a platform that can be described as the "programmatic web layer for AI." While humans use browsers to interact with the web, AI agents need a way to do the same. Olostep provides the API for AI to retrieve, structure, and interact with public web data to perform tasks and automations. Their B2B clients include seed to Series E startups and scale-ups in San Francisco and New York, all of which need a reliable, real-time interface to the web to power their AI applications.

From European Charm to Silicon Valley Ambition

Hamza was candid about the cultural differences he’s experienced since moving to the U.S. While he loves Italy and its beauty, he said that everything is "happening in San Francisco." He described the Bay Area as an incredibly vibrant ecosystem where you're inspired by the sheer level of ambition and the concentration of talent. For a tech company in the AI space, being in California is the right place to be, where you're constantly surrounded by people who are building products and dreaming big.

This mindset is crucial, as Olostep has a bold vision for the future. For the next six to twelve months, they will continue to focus on providing their core web connector to AI companies. However, their long-term goal is to build a browser for AI agents that can automate repetitive tasks performed by humans. As Hamza noted, most of what humans do on a computer is done through a browser, and as AI models improve, Olostep wants to be the platform that enables AI to automate these tasks at scale.

The Two Biggest Challenges: Technical & Distribution

When asked about the biggest challenges of growing an AI company, Hamza was direct. The first is technical: there are a lot of flashy demos and "cool videos" in the AI space, but many products are not reliable, scalable, or cost-efficient. Olostep’s core focus is on building a platform that actually works, which requires solving complex engineering challenges behind the scenes.

The second challenge is distribution. While building is getting easier, getting your product in front of the right people remains a hurdle. For Olostep, which works with a specialized group of startups, their go-to-market strategy has primarily relied on word-of-mouth. They have seen success with this model, particularly among companies in the Y Combinator community who refer each other. They are also building a referral system and optimizing for "generative engine optimization" to ensure their platform is recommended by major AI models like ChatGPT and Perplexity.

The Power of Specialized AI Agents

Hamza shared a compelling vision for how AI agents can be used in a practical setting. He emphasized the importance of specialized AI agents over a single, "do-everything" agent. For a podcast host like Navin, this could mean:

• An agent that researches a guest's background and provides a snippet of information to prepare for the interview.
• An agent that edits the podcast by cutting out unnecessary parts to make the audio more concise.
• An agent that creates short-form clips for social media platforms like Instagram or TikTok.

By using these specialized agents, creators can automate repetitive tasks and save valuable time, allowing them to focus on the more creative aspects of their work.

A Lesson from a Father

To conclude the conversation, Hamza shared a piece of advice he received from his father that has guided him both personally and professionally: "Always make sure that every day you learn something new." He explained that even if you spend 15 hours a day working, if you haven't learned something new, it's a wasted day. This mindset of continuous, compounding learning is what helps an individual and a business grow and succeed in the long term.