Our story
Started with one frustration: AI that answers with confidence, not facts.
This is the honest version: why we started, what we are building and where we are. We do not have a long corporate history to tell, so we do not pretend to.
Step 1
The problem
Most small and mid-sized businesses run on a patchwork: a spreadsheet for stock, an inbox for customers, a chat app for quotes, a notebook for what sold last week. Answers exist somewhere, but finding them takes a person, and every hand-off is a chance to get something wrong.
Step 2
Why AI changes what is possible
Language models can read a question in plain words and reach across all of that information. But a model that answers from memory will guess, and a guessed price or stock level is worse than no answer. The interesting work is not the model; it is making the model answer only from what the business actually knows.
Step 3
The premise
Pinevon started from one frustration: AI that answers with confidence rather than facts. The alternative is software that treats a business's own data as the only source of truth and says "I don't know" when the data does not support an answer.
Step 4
Anchor
Anchor is the first product built on that idea: a catalog, inventory, quotation and order system with two grounded assistants on top, one that answers customers and one that answers the owner. It is available today as a web app.
Step 5
Where we are
Pinevon is early and small on purpose. We are building Anchor with the businesses that use it, publishing only what is true, and adding capabilities in the open through the changelog.
Step 6
What comes next
Deeper workflows, more channels, better analytics and a developer surface, in that spirit: one careful step at a time. See what is being worked on and what is only being explored on the What’s next page.