Yasser Elsaid

Yasser Elsaid is the founder and CEO of Chatbase, a platform for customer-facing AI agents. His work has taken Chatbase from answering questions about uploaded documents toward agents that act through business systems. He has kept the company bootstrapped, emphasizing control over its direction and building software customers can depend on.[3][5][4]

Elsaid describes Chatbase’s opportunity as connecting general language models to information specific to a person or business. His bet was that the surrounding software would gain value as the models improved: “if the model improved and I built the harness around it that made it able to do these cool new things, then I was winning, and my customers were winning too.” This was a bet on extending model capabilities with data and software, rather than treating their initial limitations as permanent.[5]

In his announcement of the new Chatbase platform, originally published February 4, 2025 and updated July 13, 2026, Elsaid described the expansion into agents that both answer questions and take actions. Connections to Stripe and Cal.com let agents handle tasks such as changing subscriptions and booking appointments. The announcement also said Chatbase had begun incorporating OpenAI reasoning models, including o3-mini, for complex, multi-step issues.[3]

Elsaid favors keeping closely related services together as Chatbase grows. Supabase’s customer case study describes the Chatbase team’s migration of embeddings from Pinecone into Postgres using pgvector. A concrete benefit was deletion consistency: removing a document or agent could cascade to its embeddings within the database, instead of requiring separate service calls, retries, and handling for cases where only one deletion succeeded.[4]

Consolidation also required managing competing workloads. Supabase reports that the Chatbase and Supabase teams jointly diagnosed heavy analytical queries interfering with transactional traffic. Their plan combined query improvements, increased disk IOPS and throughput, and a read replica to isolate analytics. Elsaid also advises choosing tools that coding models know well: “So you want to choose tools that LLMs are very proficient at, that are always updated, and that they know how to use.” He recommends Supabase on that basis.[4]

In his EO interview, Elsaid reported that Chatbase had reached $10 million in annual recurring revenue without outside funding. He explained his financing choice through control: “The biggest advantage is control: you only listen to your customers and the team, and that's it.” He argues that AI assistance with coding, support, and marketing makes smaller teams more viable. But he also warns bootstrapped founders against remaining overly cautious once revenue becomes dependable, advocating experiments and hiring strong people even when immediate returns are uncertain.[5]

Elsaid says some early customers left because they could not discover or use capabilities Chatbase already offered. The team redesigned onboarding around different customer personas, surfaced relevant features during setup, and offered calls with team members. He favors an intuitive self-service foundation with sales assistance layered on top, especially for larger customers. In his view, self-service forces a product to work without someone guiding every step, while direct help can address demand that the product alone does not serve.[5]

His stated decision-making framework is to revisit choices when models, competitors, or market conditions change. He argues that founders should encourage teams to reverse bad decisions rather than defend them out of pride: “I think it's also the job of the founder to instill a culture where going back on a bad decision is encouraged when you're building in such a fast-moving industry.”[5]

Before Chatbase, Elsaid recalls studying computer science in Canada and interning at Tesla and Facebook. Although those internships were initially exciting, he found the career path too structured for what he wanted to build. He credits smaller independent projects with giving him more useful preparation for Chatbase: repeated practice turning an idea into something people would pay to use.[5]

Elsaid recalls developing the idea while working on AI projects in 2022, before ChatGPT launched. His first Chatbase example was uploading a textbook and chatting with it. He says a paying customer arrived thirty minutes after his launch tweet, prompting him to focus on the product. With his savings going mainly toward serving the models, he promoted Chatbase organically through Twitter, LinkedIn, and Reddit. He reports reaching $1 million in annual recurring revenue 117 days after that first tweet, which helped persuade him to build a larger company rather than keep Chatbase as a side business.[5]