memorandum · the thesisBuilt to replace the stack.
A memorandum from the two people who founded Niska, on the firms the AI wave passed over and Taju, the system we are building to replace their stack rather than extend it.
beverly hills & frederick · july 2026
memo no. 001
from · the founders of niska
re · the firms the ai wave passed over
Every wave of software so far has gone wide. It picked the easy markets, the ones with clean APIs and predictable workflows. The firms whose work product is a document — a contract, a workpaper, a deposition, a clinical note, a SOW — were left to the last decade's enterprise stack and the basic chatbots that cannot read what actually matters.
We started Niska for those firms. We are rebuilding the operational stack they already run — replacing it rather than extending it — with Taju, the firm's own compiled institutional brain, at the core. Taju is regenerated from the firm's own documents, conventions, and corrections; it is not a tool added beside the systems a firm already runs. It becomes the system.
We are convinced the gap is the same in every industry where documents are the product. It does not matter whether the firm is a financial services partnership, a defense contractor, a medical practice, a legal team, an insurance carrier, an architecture studio, or any other firm where the answer to the next question lives inside the file room. The same three things are always true. The basic chatbots cannot read all of the documents and context that matter. The enterprise software does not understand how the firm works. And the AI demos fall apart the first time a partner asks how this actually helps their bottom line.
Niska is built for that gap, and the gap is vertical agnostic. Taju is the same compiled engine in every deployment; what it generates — the answers, the automated workflows, the operational stack itself — is generated entirely from your firm's own documents, your people, and your process. Nothing generic ever ships to a customer.
We founded Niska because we agreed on one thing. The industries that run on documents and institutional knowledge are some of the most valuable in the economy, they had been left behind by the current wave of AI, and someone had to actually build the system that replaces their stack from the inside. So we are. That is the whole of the agreement, and this memorandum is the position we hold.
signed,
Tucker Peters
co-founder
Leo Roberto
co-founder
- 01
Tucker Peters
co-founder
Tucker is a sophomore at William & Mary pursuing a double major in Finance and Government, and concurrently interns at WGS Systems, a defense technology contractor. This summer he joins Keystone Strategy in New York as a Summer Associate.
He founded the William & Mary AI Club, a 130+ member organization, and previously built AI-enhanced learning products through GenEdu LLC. Originally from Westport, Connecticut.
- 02
Leo Roberto
co-founder
Leo is a student at William & Mary with a track record of building revenue-generating ventures. He founded RenderRise Digital, a websites and AI systems agency that scaled to $20K MRR serving SMBs, and has held growth and publisher operations roles at Falcon Labs in the San Francisco Bay Area.
He co-founded The Economic Review at William & Mary, the university's student journal for economics. His earlier experience spans internships at PBS NewsHour, Kontrolmatik Technologies, and The George Washington University School of Business.
Partner on a vertical
For operators who have spent a decade or more inside their industry. We bring the engineering, Taju, and the capital; you bring the vertical.
- Equity in the operating company — the vertical operator is a co-founder, not a customer
- One partner per vertical
- Multi-year relationship, not an engagement
- First call to signed term sheet, usually six to eight weeks
Buy the system
For firms that need the system to show up, run inside their environment, and start taking work off their partners' desks.
- Deployed inside your environment — on-prem, hybrid, or cloud
- Stood up in four to six weeks
- Annual pricing, scoped to firm size and document volume
- No per-seat pricing, no per-query pricing, no selling the data
Write to tucker@niska.ai, leo@niska.ai for partnerships and press. Or visit the contact page.