Raava / Qwalo

02 · Case study

Lead qualification, run by agents.

Qwalo reads the replies sellers send to a land buyer's text campaigns and decides which leads are worth following up.

Live · first clientProperty check in shadow
Qwalo: a seller's reply goes to a fast model, then a stronger model when the first is unsure, then a person. Each settles what it can. Human calls go back to the models as examples. Seller replies Fast modelsettles the clear ones unsure Stronger modela second look still unsure A personmakes the call Settledfollow upor drop each human call returns to the models as an example
Each tier settles what it can and passes the rest down.

A cascade, not one model

  1. 1A fast model settles the clear replies.
  2. 2When it is not sure, a stronger model takes a second look.
  3. 3When that one is not sure either, the lead goes to a person.

Each human decision goes back to the models as a worked example, so the share they can settle grows over time.

Guard rails first

Sending is off by default. The model may not reply on its own until a backtest on held-out history shows it is right at least 90% of the time, and every change to a prompt or model must pass an evaluation before it reaches production.

Next: the property itself

A second step, now in shadow, checks the land behind the lead: who owns it, whether a road reaches it, how much of it is wetland.

Where it is

Qwalo went live with its first client in October 2026.

Building something agents should run?

Raava Solutions takes on Palantir Foundry and AI agent work for clients. Tell us about the operation.

contact@raavasolutions.com