For firms that want to implement their own AI capability
This can be your firm’s in‑house AI capability
Rivas Cuervo is an AI-driven legal team. We built a method around the way partners work with associates. We want to keep building it with you and for you.
“An opportunity to bill clients for specialized AI capabilities.”
The issue
A partner’s work with an associate is a loop. The engagement in the partner’s words, “look at these,” notes in the margin with the exact passage, “write the memo with this,” comments on the draft, a new version in tracked changes, and nothing the partner said gets lost.
Chat-style legal AI is built around the answer, and the loop stays with the lawyer. That loop is what we built.
The aim is to reduce the number of junior associates and paralegals that work needs.
The demo, one part at a time
One matter, from the engagement to the drafts, in six short clips. What the partner says, and what the system does. Fictitious names and sandbox documents; the system as it ran in October 2026.
01 / 06 · The partner says
“Here is what the client asked for.”
The system does
The engagement, in the client’s words
A matter starts with the client and the engagement, pasted as the client sent it. The data room comes in as it is: 49 documents and the buyer’s first draft of the share purchase agreement, each with a short card that says what it is.
Another matter, start to finish
An unpaid promissory note, read and drafted on its own
Four documents and one task marked automatic. The system built the list, read the documents in two batches, and drafted the executive collection complaint in Spanish, with its evidence list, for the lawyer’s review.
What we measured
Cost in model use. Benchmark, October 2026.
| What | Cost |
|---|---|
| Reading a document | 1 to 3cents per page |
| A first draft | About USD1.20 |
| Each round of refining with the partner’s comments | USD0.78 |
This can be your in‑house capability
Some large firms are building AI units of their own. This is one, already working on real matters. We offer to bring it inside one firm and keep building it there, with the people who made it, on the firm’s terms.
It does not depend on any AI vendor. The method lives in its own instructions and code, and the models are called from a single point, so the provider can be changed.
Owning it also gives the firm control over what it costs.
And it opens a new line of work: specialized AI capabilities the firm can bill its clients for.
Who we are
A team of law students in Bogotá. Working inside large firms, we saw what a partner hands to a junior associate and realized it could be done differently, so we built it.