As companies increasingly turn to artificial intelligence to reduce staffing costs and accelerate software development, Chatbase founder Yasser Elsaid is taking a different approach, planning to double his startup’s workforce over the next six months after concluding that AI-assisted coding has made a larger team more valuable, not less.
Chatbase, a San Francisco-based startup that helps businesses build AI agents for customer service and support, currently employs 40 people. Elsaid said the company intends to expand that headcount to about 80, reversing his earlier preference for operating with a lean workforce.
“I might have been wrong about building with a lean team in the age of AI,” Elsaid wrote in a post on X this month.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
In comments to Business Insider, he described his previous position as “pretty wrong,” arguing that founders risk making poor decisions when they organize their businesses around technological capabilities that have not yet materialized.
“I think it’s a big mistake for founders to live in a future that doesn’t exist yet,” he said.
The reversal highlights a growing distinction in the debate over AI and employment. While more capable models are allowing companies to automate coding tasks and potentially operate with fewer employees, some founders are finding that lower development costs create an opportunity to pursue more ambitious products, expand customer relationships and increase the amount of work a company can undertake.
Elsaid previously embraced the lean-team approach. In a LinkedIn post in January, he said Chatbase had maintained a workforce of 17 full-time employees for much of its first three years, with most of its recruitment concentrated in engineering.
The availability of more capable coding models, including Anthropic’s Claude Opus 5.5, initially reinforced his belief that the company could accomplish more with fewer people. But he subsequently concluded that this logic overlooked an important distinction: making software easier to produce does not automatically make it easier to identify what customers need or determine which products are worth building.
A startup that uses AI only to maintain its existing development plans may benefit from lower costs and faster delivery. A company that uses the same technology to pursue more complex projects could instead need additional engineers, product specialists, and customer-facing employees.
“If building is easier, build harder things,” Elsaid said. “Most moats in software come from someone doing something hard for longer, and being right. That hasn’t gone away.”
His argument challenges the assumption that improvements in coding models must translate directly into smaller engineering teams. If AI reduces the time needed to write code, companies may be able to redirect more resources towards product differentiation, experimentation and customer needs rather than simply cutting headcount.
The distinction makes a lot of difference in customer-service software, where producing working code is only one part of delivering a useful product. Businesses must also understand customer workflows, integrate AI agents into existing systems, and ensure that automated interactions meet expectations.
Elsaid estimated that roughly 90% of the work performed by Chatbase’s engineers now involves product development rather than writing code. That includes researching problems, speaking with customers, forming views about what the product should do, and translating those decisions into specifications that AI agents can implement.
The remaining work still requires engineers who understand the underlying code and can ensure that different AI agents operate consistently rather than producing conflicting changes.
His account suggests that AI is changing the composition of engineering work rather than eliminating the need for engineering judgement. Code generation may be increasingly automated, but deciding what to build, evaluating whether the output is correct, and coordinating complex software development remain human responsibilities.
Hiring Plans Run Against The Industry’s Workforce Cuts
Chatbase’s expansion contrasts with a series of workforce reductions by technology companies that have cited AI as a reason they can operate with fewer employees or reorganize teams around more automated workflows.
Block cut nearly half its workforce in February, with CEO Jack Dorsey pointing to AI and the move towards smaller, flatter teams. Coinbase announced a 14% workforce reduction in May, citing engineers’ ability to complete in days work that previously took teams weeks. The cryptocurrency exchange has also been testing smaller AI-focused teams and the concept of one-person teams.
The broader movement has encouraged predictions that more capable AI tools will allow startups to generate substantial revenue with very small workforces. OpenAI CEO Sam Altman has predicted the emergence of a billion-dollar company run by one person. Cursor, meanwhile, grew from $1 million to $100 million in annual recurring revenue in less than a year with fewer than 50 employees, according to private-market research platform Sacra.
These examples demonstrate the potential for AI to increase output per employee, but they do not establish that every software business will benefit from reducing its workforce. Revenue growth, product complexity, customer acquisition costs and the demands of enterprise clients can all influence how many people a company needs.
Elsaid said Chatbase is also recruiting for customer-facing positions, particularly in sales and customer success. Enterprise customers still expect providers to understand their requirements, respond to concerns, and demonstrate why a product is worth purchasing. Those activities cannot be reduced to generating code.
The hiring plan also highlights a potential limitation of the one-person-company narrative. Even when AI tools reduce the cost of producing software, businesses still need to find customers, build trust, manage commercial relationships, and turn technical capabilities into recurring revenue.
“The models are extremely smart and capable, but not fully independent,” Elsaid said. “Maybe that changes and one day you won’t need humans in the building process at all. But that isn’t today.”
The next six months will test whether a larger workforce can help Chatbase take advantage of faster AI-assisted development and pursue a more ambitious product roadmap. The decision does not disprove the potential for AI to reduce employment elsewhere. Instead, it illustrates how the same technology can produce different staffing outcomes depending on how businesses choose to use the productivity gains.