Composite walkthrough
Meridian Legal Partners — Document Intelligence Pilot
We’re a new firm. Rather than borrow metrics from prior employers, we ran a fictional 45-attorney litigation firm through our real platform end-to-end and published every artifact a paying client would receive — discovery call to qualified pilot to evaluation gate.
Names and transcript are synthetic. Every number on this page is a real platform output from a live run, including the readiness scorecard you can re-run yourself at the bottom of this page.
ICP fit
0.85
Readiness
95 / 100
Opportunities
5
Investment
$105,000
Brief confidence
0.78
Engagement
From a 30-minute call to a gated pilot
Five stages, each with a written artifact. No stage skipped, no metric fabricated.
- 1
Discovery
30-minute call, written record
A managing partner at Meridian books a discovery call. We capture the transcript, structured notes, and an ICP-fit score against our published rubric. The output is a qualification memo — not a sales pitch.
Artifact
ICP-fit score: 0.85 / 1.00
- 2
Qualification
Rubric-scored, not vibes
Meridian scores high on document volume (~8,000 active matter files), executive sponsorship, and budget readiness. They flag HIPAA-adjacent health-matter work, which routes them through the healthcare data-handling profile from the start.
Artifact
Profile: legal-services + healthcare-adjacent
- 3
Paid Readiness Assessment
10 days, $5,000, scorecard you own
We run the readiness rubric across data, workflows, risk, talent, and governance. Categories are weighted and scored. Findings and recommendations are written, not generated as bullets — every category has reasoning the partner can defend internally.
Artifact
Readiness score: 95 / 100
- 4
Opportunity register
Five candidates, ranked by impact and effort
The platform proposes five pilot candidates ranked by priority score and time-to-value. Each is mapped to an offer in our catalog and an estimated investment. The partner picks one and signs on a fixed-scope pilot — no open-ended retainer to start.
Artifact
5 opportunities · $105,000 investment schedule
- 5
Pilot — Multi-agent legal brief
Cited, evaluated, and gated
Associates feed an active matter and a research question. A supervisor agent routes retrieval, drafting, and citation-checking sub-agents. Every paragraph cites a source. A golden-set eval runs on every change before the model is allowed near a real client matter.
Artifact
Legal-brief confidence: 0.78 (eval gate ≥ 0.75)
What this walkthrough exercises
The same platform every paying client uses
Six platform capabilities are hit end-to-end in this walkthrough. The full service catalog is on /services.
ICP-fit + qualification
Discovery call transcript → structured rubric score → qualification memo with recommended next step.
Readiness scorecard
Seven weighted categories: data, workflows, risk, talent, governance, infrastructure, sponsorship.
Opportunity register
Ranked pilot candidates with offer mapping, time-to-value, and an investment schedule the buyer can present internally.
Multi-agent pilot
Supervisor-routed retrieval + drafting + citation checks. Eval gate must pass before any production deployment.
Citations on every answer
RAG outputs link back to the source document and span. Faithfulness threshold ≥ 0.90 enforced in CI.
Human approval on side effects
Outbound emails, ticket creation, and CRM writes pause for review. Audit trail is enforced, not optional.
Live platform run
Run this scenario against the real readiness endpoint
The narrative above is a written walkthrough. This button posts Meridian’s inputs to POST /v1/assessments/readiness/preview - the same scoring service that powers a paid engagement. No CRM records are created; the scorecard is computed and returned live.
Honest positioning
Meridian is fictional. The platform is not.
We’d rather show you a composite walkthrough on a real platform than dressed-up metrics from prior consulting roles. If you’d like to see the same flow on your own data, the Readiness Assessment is the next step — fixed scope, fixed price, and yours to keep whether or not we continue.