piedras.ai

Offers

Diagnosis first. Then one tool in production.

We do not sell a company-wide platform. We sit with whoever owns the call — office or field — pick one expensive miss, and ship a tool they own. The person running the job still decides.

Call

Thirty-minute triage

No charge

The last expensive technical miss, who owns the answer, and whether the source documents can be reviewed under NDA.

  • What problem ate the most hours — office or location
  • Who gets blamed if the answer is wrong
  • Where the truth lives — drawing, procedure, last job, or someone’s head

10 business days

Problem Sprint

$29,500

Credits 100% into a subsequent build. We walk if there is no owner or no source of truth.

  • Map of how the decision is made today
  • The source documents and the dangerous file
  • Recommended tool and a written statement of work
  • A working slice on your sample

8–12 weeks

One tool in production

$95k–$145k

A knowledge system or a rebuilt tool. Not a platform. A named person still releases anything that goes to the field.

  • Working tool they own
  • Citations, runbook, and a named owner
  • Read-only against real cases before any write
  • Narrow writes only after that holds

After production

Tool-ops retainer

$7,500–$9,500 / month

After go-live we keep watching the tool: failures, model and document changes, and a monthly scorecard. Plant tools drift. This is how it stays useful.

Engagement sequence

  1. 01

    Triage

    Thirty minutes on the last job that stalled between office and field.

  2. 02

    Sprint

    Ten days. Map, sources, written next step. Credit into a build.

  3. 03

    Shadow

    Read-only against historical and live cases.

  4. 04

    Production

    Named person decides. Narrow writes only later.

Bring the last technical miss. Not the AI roadmap.

Book the triage