ADRI Adoption Journey

Take AI agents from local protection to governed enterprise data supply. This guide connects the open-source toolkit to Verodat MCP so you know exactly when to switch on managed services.

Overview

Step	Outcome	ADRI OSS	Verodat MCP / Enterprise	What to show	Revenue Model
1	Define data need	✅	–	Run adri setup --guide and capture "good" data	OSS – Free
2	Block bad data	✅	–	@adri_protected(..., on_failure="raise") blocks failures	OSS – Free
3	Handle error data	✅	–	Review JSON reports in ADRI/dev/assessments	OSS – Free
4	Log for dev (local)	✅	–	Enable local audit in config, tail CSV logs	OSS – Free
5	Log for prod	✅	✅	Send audit events to Verodat workspace	Token burn-down (single app)
6	Analyse performance & compliance	✅	✅*	Use Claude project connected to Verodat ADRI	Token burn-down
7	Publish to datasets	✅	✅*	Publish standards to Verodat datasets	Token burn-down + Dataset licence
8	Prove enterprise scope	✅	✅	Configure agent tools in tools.verodat.io	Token burn-down + Dataset licence
9	Meet compliance requirements	✅	✅	Replay logs + controlled data supply	Token burn-down + Dataset licence
10	Ship AI solutions	✅	✅	Share ADRI standards in marketplace	Token burn-down + Dataset licence

✅* indicates the open-source edition makes it possible, but Verodat MCP reduces the effort to near-zero.

Data flow and logging

Stage 1 – Local Wins (Steps 1–4)

1. Create standards quickly – adri generate-standard gives you a YAML contract in minutes. Store it in git with your agent code.
2. Block bad inputs immediately – Use @adri_protected(standard="...", data_param="...", on_failure="raise") in production pathways.
3. Understand failures – Inspect the JSON report or call adri show-standard to see which rule fired.
4. Capture local evidence – Flip the adri.protection.verbose_protection flag or enable local audit logs when you need traceability during development.

👉 Deliverable: reliable agents, developer trust, no platform commitment yet.

Stage 2 – Shared Visibility (Steps 5–7)

5. Stream logs to Verodat – Configure enterprise logging so assessments land in a governed workspace (adri.logging.enterprise with Verodat credentials). Product teams and compliance view the same evidence.
6. Analyse trends with Claude projects – Use Verodat’s MCP integration to benchmark failure patterns, drill into compliance questions, and auto-generate remediation tickets.
7. Publish standards as datasets – Promote validated ADRI standards into Verodat datasets so downstream teams query the same definitions via managed endpoints.

👉 Outcome: single-source-of-truth for standards + logs, lightweight governance, usage-based billing.

Stage 3 – Enterprise Delivery (Steps 8–10)

8. Expose agent configuration – Manage tool access, rate limits, and data supply governance in tools.verodat.io so every agent run is compliant by default.
9. Prove compliance – Use replayable logs, deterministic data supply, and Verodat audit views to satisfy ISO/SOC/HIPAA reviews.
10. Ship packaged AI solutions – Publish ADRI-backed standards to the marketplace so partners consume the same high-quality data contracts.

👉 Outcome: production-grade data supply for AI agents with revenue alignment to usage.

Implementation Checklist

• Follow the Getting Started guide to complete Steps 1–4.
• Enable local audit logging when stakeholders ask for evidence.
• Schedule a Verodat MCP onboarding session before promoting to production (Steps 5–7).
• Connect enterprise logging and Claude analysis before compliance reviews (Steps 8–9).
• Package your standards for marketplace distribution (Step 10).

How to Talk About the Journey

Persona	Why it matters	Message
Agent engineer	Needs confidence that agents won’t crash on bad data	“Guard your function with one decorator, see the pass/fail score instantly.”
Data platform team	Must provide governed data supply	“Publish ADRI standards to Verodat once, let every agent reuse them.”
Compliance lead	Requires replayable evidence	“Every assessment is logged locally and can be streamed into your controlled workspace.”
Business owner	Wants measurable reliability	“Scorecard every dataset before the agent runs; share dashboards via Verodat MCP.”

Ready to Scale?

• Stay in OSS while you iterate quickly and build proof with local auditing.
• When you need shared compliance logging or managed data supply, reach out at thomas@verodat.com to activate Steps 5–10.

Your first win should happen locally. The adoption journey shows stakeholders how to expand that win into an enterprise deployment without rewriting your agent workflows.