The deal is finished. Somewhere in the agreement is a term worth remembering: a compromise, a concession, a way around a problem that will almost certainly turn up again. The lawyers have done the difficult work. Then the document goes into a database, where finding it can become another piece of difficult work. Kevin Walker has built a company around this peculiar talent of professional knowledge to disappear while remaining exactly where somebody put it.
Centari, the company he co-founded in 2023, turns transaction documents into structured, searchable information. Walker came to the problem from the negotiating side of the table. He had practiced M&A, worked inside businesses, and experienced the irritation of knowing useful precedent existed without having an easy way to retrieve it. For him, the archive contained unfinished business.
A detective with a different case
There is an earlier, rather cheaper version of this story. In elementary school, Walker tried running a neighborhood detective company. The offer was straightforward: solve a case for 25 cents. He recalled the venture years later while explaining an interest in business that predated his legal career. The pricing strategy, at least, left room for growth.
His route into entrepreneurship passed through a history degree at UCLA and a J.D. at UC Berkeley. A lifelong Californian before moving to New York, he built his professional experience in and around San Francisco. Those details place him within two overlapping worlds: the legal teams that work through transactions and the businesses trying to complete them.
The childhood detective anecdote offers a pleasing opening scene, but his adult decisions give the story its weight. Walker kept choosing work close to deals. In law firms, startups, and real estate, he returned to the same activity: helping people put an agreement together and move a business forward.
The education between jobs
At Paul Hastings, Walker practiced M&A. He then moved in-house, serving as head of legal at Kasa Living and later as deputy general counsel at Hamilton Zanze Real Estate Investments. The sequence gave him experience of legal work from both the outside adviser’s position and the client’s.
At Kasa, where he was the ninth employee, the job stretched beyond a narrowly defined legal department. He described working across sales, technology, and other operating questions. An early company has a way of expanding a job description without troubling itself to issue a revised one.
That experience matters to the business he eventually chose to build. A lawyer can recognize a problem in a document. An operator also has to consider how the problem affects the work around it. Walker had seen enough transactions to recognize repeated friction, and enough of a young business to want to address it himself.
“I've always loved building.”
Kevin Walker, Pioneers and Pathfinders, June 2025
An archive with unfinished business
Generative AI supplied the opening. Walker met technical co-founder Bryan Gilbert Davis, whose background included machine learning and data teams at Nextdoor, Meta, and Indeed. When Centari publicly announced its platform in August 2023, its initial proposition included helping lawyers find relevant forms, clauses, and deal history while drafting agreements.
That founding pairing was practical. One founder knew the work that produced the documents; the other brought experience building systems that could work with information. Their early backers included South Park Commons and Alt Capital. The announcement invited more legal teams into a pilot program.
Walker later described the arrival of ChatGPT as a moment that made pursuing the opportunity feel urgent. He wanted firms to make better use of information they already possessed. In his account, the attraction extended beyond writing assistance to organizing a firm’s own accumulated deal knowledge.
By July 2024, he was talking about a system that could use multiple AI models and accommodate different deployment needs. The broader ambition was a transactional law platform, with structured deal information supporting more of the work. The recurring object of his attention was the use case: what a lawyer could actually do with the result.
A little Moneyball at the negotiating table
Walker gave his argument a more familiar setting in an essay for the American Bar Association: baseball. His Moneyball comparison concerned the possibility of finding useful patterns in accumulated deal information. A firm’s agreements record positions that people actually negotiated. Examining that record could improve the information available for the next discussion.
The appeal is easy to grasp. An experienced lawyer remembers a great deal, but a firm’s collective history extends beyond any one person. Walker was interested in making that wider record available, including terms and relationships that take legal understanding to interpret.
There is an appealing modesty to the underlying question. Before asking a professional to invent something new, help them locate something useful they have already done. The comparison does not settle a negotiation. It gives the person conducting it a better starting point. A filing cabinet rarely gets invited to a client meeting; its contents may have something to contribute.
The cookie school of enterprise sales
Building the product left Walker with another job: persuading people to buy it. His description of Centari’s early sales effort was notably human. Large law firms shared conferences and dinners, and introductions developed through those overlapping circles. The team brought Maman cookies to conferences and became recognizable for them.
It is a useful detail in a business often described through diagrams of sales funnels. The cookies gave people a reason to remember an encounter. The harder part was arriving with software they believed could help. Walker described those early relationships turning into customers and word of mouth.
By June 2025, he was discussing the transition toward a larger sales effort, including senior account executive hires and a chief of staff. The founder could explain the problem from experience. Expanding the business meant helping other people carry that understanding into conversations of their own.
The money, and what followed
In September 2025, Centari announced funding that brought its total raised to $14 million. The Series A portion was $8.4 million, led by Sentinel Global, with participation from investors including GTMfund, South Park Commons, Alt Capital, RiverPark Ventures, and Recall Capital. The distinction matters: a running total and a new round describe different things.
The company planned to put the capital into research, development, and hiring. Its team included former lawyers as well as technical specialists. For Walker, legal experience remained part of the product-building argument: understanding a transaction’s language required understanding the work.
Then came a concrete example of adoption. In October 2025, Ropes & Gray announced it had taken up Centari across its transactional practices. The firm described using its proprietary deal and fund databases to surface market insight and make institutional knowledge accessible to its lawyers.
That announcement brought Walker’s thesis into a recognizable professional setting. A firm had an existing body of work and wanted more people to use what it contained. The archive had acquired a new assignment.
When the knowledge leaves the filing cabinet
During 2026, the product expanded toward the places that information might be used. In March, Centari introduced Intelligence and Views. Intelligence offered natural-language search and analysis across deal history; Views provided ways to visualize the resulting data. The company reported customers in the United States, United Kingdom, and Australia.
By then, Junyuan Lau was serving as chief technology officer. The current leadership reflected the continuing need to bring legal and technical work together. Walker remained focused on what reliable information could enable in negotiations, market comparisons, and client advice.
July’s External Views release pushed that argument outside the firm. Permission-controlled dashboards could present client-specific or anonymized aggregate information under the firm’s own branding. Walker framed the need in terms of clients expecting guidance grounded in actual market experience.
For a founder who had worked in-house, that audience was familiar. The client wants to understand the available choices. A record of comparable transactions can give counsel something more concrete to discuss, provided the data and the permissions are handled carefully. Sharing the insight becomes another product problem to solve.
The next case
In October 2026, Centari announced Agentic Review in preview. The feature was designed to recognize patterns in approved outputs and automatically verify high-confidence answers, while retaining human oversight and an audit log. The same release brought updates for comparing live drafts with deal history and asking questions through email.
The direction is telling. Once information has been extracted, somebody still has to check it and decide how to use it. Walker’s company is spending effort on that less theatrical part of the workflow, where a promising answer has to become something a professional can rely on.
His ambition reaches beyond recovering an old agreement. He wants each completed transaction to leave behind knowledge that makes subsequent work more informed. The quarter-a-case detective business has long since given way to enterprise software. The durable interest is in locating the answer, putting it within reach, and giving someone a reason to use it.