BENGALURU — Bhavin Turakhia has started six companies in a little over two decades, and he funded most of them out of his own pocket before any venture capitalist would take the meeting. His newest bet is $30 million of that same money, aimed at a target most founders would consider suicidal: Microsoft Office.
The company is called Neo, and its premise is narrower than it sounds. Turakhia is not arguing that Word, Excel and Outlook are bad products. He is arguing that bolting a chatbot onto software designed two decades before generative AI existed, which is what Microsoft has done with Copilot and what Google has done with its Workspace AI features, cannot produce software that actually thinks alongside the person using it. Neo is built model-agnostic from the ground up, according to the company, meaning a business can swap the AI model underneath it without switching the software itself, and its core unit is not a document or a spreadsheet but an AI agent that a knowledge worker delegates tasks to.
Turakhia’s pattern is what makes the bet worth taking seriously rather than dismissing as another AI-wrapper pitch. He co-founded Directi at 18 with roughly $25,000 in savings, sold four of its subsidiaries to Endurance International Group for $160 million in 2014, then built Titan, a business email service that drew a $30 million investment from Automattic at a $300 million valuation. Zeta, his banking-technology company, took $300 million from Sodexo and later $250 million from SoftBank’s Vision Fund, reaching unicorn status. In every case, he financed the company himself long enough to prove the idea worked before outside money arrived. Neo follows the same sequence: self-funded first, VC pitch later, if at all.
What Neo does not yet have is a public product. The Bengaluru-based startup employs about 45 people, 18 of them engineers, and plans to begin rolling the software out to mid-sized businesses in the coming months, starting with knowledge workers in technology, consulting and professional services. That is a small team and an unproven release schedule to go up against a company that reported more than $270 billion in annual revenue last year and has spent three years and billions of dollars integrating AI into the exact software Neo is trying to replace.
The skepticism Neo has to overcome is not really about the technology. Building an AI-agent-centric document and spreadsheet tool is achievable for a team with Turakhia’s engineering track record. The harder problem is the one every enterprise-software challenger faces: getting a mid-sized business to migrate years of documents, templates, macros and institutional habit off Office, a product most employees already know how to use without training, onto something unfamiliar that asks them to trust an AI agent with real work. Microsoft’s moat was never really the software. It was the decades of files already sitting inside it.
Neo’s model-agnostic pitch is also a hedge against a specific risk in the AI enterprise market: vendor lock-in anxiety. Businesses that have watched AI model quality and pricing shift dramatically between providers over the past two years have grown wary of building workflows around a single model. By designing Neo to run on whichever model a client prefers, Turakhia is betting that flexibility matters more to a mid-sized business’s IT department than picking the single best-performing model available today. Whether that flexibility survives contact with actual enterprise procurement, where switching costs and support contracts often matter more than architecture, is untested.
India’s broader technology sector has drawn a wave of AI-infrastructure investment this year, from Amazon’s $48 billion AWS commitment to chip partnerships between global labs and Indian manufacturers, but Neo is a rarer kind of story: a founder personally underwriting a direct assault on the software nearly a billion office workers use daily, without first raising a funding round to share the risk. If Neo fails, Turakhia absorbs the loss alone. If it works, he owns considerably more of the outcome than a venture-backed founder would.
Turakhia has not said publicly what he considers success within what timeframe, or what happens to Neo’s 45 employees if enterprise adoption stalls the way most challengers to Office eventually do. Google Workspace, the most durable Office competitor to date, has taken more than a decade to reach roughly a third of Office’s market share, and it arrived with Google’s distribution and cloud infrastructure behind it. Neo has neither. What it has is a founder who has been right about unglamorous enterprise software before, and a thesis that has not yet been tested against a single paying mid-sized customer.
