Two hundred customers sounds like a useful thing to have. Fifteen sounds like a problem. In the story Shifu Ventures tells about Photon Legal, however, the smaller number belongs to the bigger business. The intellectual-property firm shifted toward enterprise accounts, and Shifu reports that its average contract value rose from $25,000 to $250,000. Sometimes the crowded address book is the thing getting in the way.
- Shifu supplies operators to build alongside founders.
- Photon Legal is its first published co-building case.
- The lesson worth borrowing: choose the buyer before scaling the machinery.
That is a more interesting starting point than another promise to help startups grow. It suggests that growth can require subtraction: fewer buyers to understand, fewer competing demands, a clearer reason to buy. It also poses an awkward question for founders collecting small wins. Are those wins teaching the company how to become the business it wants to be?
The curious case of fewer customers
Prateek Sethi, Shifu’s co-founder, has described the starting condition bluntly in a public post: Photon was a traditional law firm with years of plateaued growth, selling too broadly. In that earlier account, he put revenue at roughly $700,000 and the later run-rate at about $2.5 million. His concern was whether intensive operating help could produce profitable growth before Shifu expanded the model.
His account describes work across sales strategy, product, hiring, culture and systems. The published numbers are snapshots from different stages, rather than a tidy, independently measured experiment. What makes the account useful is the sequence: first discover an operating problem, then work inside the company, then decide whether the result justifies doing it again. The studio’s confidence followed an engagement.
The current Shifu case puts Photon at $3.5 million in annual recurring revenue, up from $600,000. Treat those as the studio’s reported results. The revenue label alone does not tell a reader how much comes from subscriptions, legal work or a blend of both. That distinction matters when judging margins, staffing needs and how easily the next dollar can be earned.
About 5.8× the reported starting ARR. Shifu’s $10M ambition remains a separate target.
There is a catch in the arithmetic. A larger contract can fund more specialized work, but it can also purchase more complexity. The test is whether the company can deliver that work repeatedly without rebuilding its service for every buyer. A bigger invoice is a useful clue; it is not the whole diagnosis.
Before the studio, a detour
Sri Maneru’s route to Shifu included a plan that he abandoned. After co-founding alumni-engagement software company Almabase, he considered electoral politics. In his personal account, the work’s emotional demands and its effect on his family forced a reconsideration. He concluded that the fit was wrong. That admission gives the later company-building mission a useful human scale: choosing work includes choosing what you can bear doing.
He returned to entrepreneurship after an exploratory period involving pilot projects, travel and conversations. Sethi was already a familiar collaborator from Almabase. The pair’s Bharat SaaS Yatra took the inquiry onto the road; an eChai account describes the plan to visit 20 cities and tell founders’ stories beyond the usual metropolitan circuit. The studio grew out of listening as well as operating experience.
A product inside the practice
Photon Legal gives this operating philosophy something concrete to examine. Its founders, Amit Koshal and Prayank Khandelwal, established the firm in 2019. Its own timeline places the launch of Photon Pulse in 2025. That makes the case more specific than a software studio simply adding a chatbot: an existing professional practice was developing a product around work it already understood.
Pulse’s product page starts with a wonderfully ordinary question: where do employees keep ideas that might deserve patents? It offers idea submission, patentability assessment and a way to manage the process. The promised speed of an initial assessment is a product claim, not a patent grant. A useful interface can help surface an invention; the legal outcome remains another matter.
Here is the practical idea a reader can borrow. Start with the recurring work customers already value. Find the handoffs where ideas disappear or decisions stall. Build software around those moments, then examine whether a different buyer will pay for the resulting service. Installing AI and finding a stronger commercial proposition are separate jobs, and both need doing.
Five seats change the work
Shifu’s current model caps active co-building at five companies, with a three-month working pilot during selection. It looks for full-time founders building AI-native B2B businesses in India for global customers, typically with about $100,000 ARR. Past $10 million, the stated plan is to step back into advisory and release a slot. That is a staffing commitment with a growth target attached.
“We don’t invest in the companies we co-build.”Shifu Ventures’ current FAQ
The wording matters. Founders seeking funding alone have a different problem to solve. Those considering an operating partner should negotiate decision rights, commercial terms, milestones and the practical division of work. They should also ask whether their sales cycle and delivery economics suit enterprise accounts. Fifteen large customers can bring focus; losing one can hurt considerably. The same subtraction has two edges.
Capacity is the real wager. A studio can promise deep involvement only while its people have time to provide it. Shifu’s stated preference for a team of 10 to 15, and its unusually long 15-year partner vesting schedule, make that bet explicit. A founder hiring this kind of help is choosing colleagues for consequential decisions, so a trial together has obvious value.
For the customer, the distinction should be visible in what gets finished. A positioning decision should change the sales conversation. A product decision should remove a recurring obstacle. An operating partner earns its place by connecting those decisions to daily work. Otherwise, the extra people become another meeting on the calendar.
The district beyond the deal
The related Shifu Foundation serves a different starting line. Its fellowship targets people building full-time, especially outside major metros, and describes its support as equity-free. Hometown Talks, run with NITI Aayog and Atal Innovation Mission, brings founders back to their hometowns to speak with local students and entrepreneurs. Inspiration and operating help have different audiences, even when they share an origin.
Shifu’s most portable idea is therefore smaller than its ambition to build 100 companies. Spend time close enough to see what customers actually buy. Test the working relationship. Put a limit on commitments. For a founder surrounded by invitations to do more, choosing what deserves sustained attention might be the useful first move.