DecisionTrail design-partner pilot brief

A buyer-facing brief covering pilot scope, deployment boundary, setup requirements, weekly success metrics, and the low-overhead operating model for regulated Teams workflows.

Who this pilot is for

DecisionTrail is best suited to compliance, legal, risk, investigations, and operations teams that need to reconstruct high-stakes decisions across Microsoft Teams without stitching screenshots and exported threads by hand.

Standard pilot promise

The standard design-partner promise is intentionally narrow:

• one high-cost workflow
• one Microsoft tenant
• one privacy-safe deployment model
• one set of measurable weekly success metrics
• one narrative export path teams can review and challenge

What is in scope

• Tenant-wide Teams coverage for channels and group chats
• Private-chat coverage as an explicit client admin choice
• Cross-conversation timeline assembly for named business identifiers
• Automatic publishing for clear evidence
• Background QA for sampled or uncertain cases
• Optional stricter holdbacks if the client wants them
• Narrative export with chronology, actor context, and source references

What is not in scope

• Generic AI assistant or broad productivity claims
• Self-serve multi-tenant onboarding
• A promise of zero human oversight in every environment
• A claim that the public website by itself proves regulatory compliance

Minimal operational overhead means

DecisionTrail should not require a bank to stand up a new analyst team just to use the product.

In the default operating model:

• clear events flow straight into timelines and exports
• background QA samples the edge cases
• manual corrections are auditable and low-volume
• stricter holdbacks are optional rather than mandatory

Required client setup

• Approved Microsoft Entra access for operators
• Client decision on coverage policy and private-chat scope
• Azure deployment boundary aligned to the client’s data-custody requirements
• One named workflow the pilot is expected to improve

Weekly success metrics

• time to first usable timeline
• time to narrative export
• QA correction rate
• blocking holdback count
• cross-conversation timeline existence
• export success with source references intact

Current live proof snapshot

The March 30, 2026 production synthetic walkthroughs are the current reusable proof baseline:

• Meridian Bank Test: 33 messages, 91 events, one 7-conversation timeline, 24.05s to first usable timeline, 2.92s to export, 0 blocking holdbacks
• Zephyr Fund Test: 33 messages, 135 events, one 9-conversation timeline, 36.02s to first usable timeline, 8.65s to export, 0 blocking holdbacks

Those runs were reset into the default background-QA posture first, so the walkthrough reflects the current low-overhead operating model rather than an older stricter review policy.

ROI language to keep consistent

DecisionTrail should be framed as:

faster reconstruction, lower manual stitching, more defensible chronology

Recommended next steps

Review the proof pack, inspect the sample report, then book a demo using one real Teams workflow you want to pressure-test.