The Blind Read Protocol.
A practical way to compare AI meeting platforms without relying on vendor demos, feature matrices, or curated recordings.
1. Select three real meetings
Use ordinary client meetings from the current week: one straightforward review, one planning-heavy conversation, and one meeting with ambiguity, emotion, or a material life event.
2. Run the same source through each platform
Keep speaker attribution, recording quality, and source material as consistent as possible. Do not let either vendor choose the sample.
3. Apply the send test
Would the advisor send the follow-up under their own name after one read? Measure rewrite time. A 20-minute rewrite is not automation savings.
4. Hunt for the soft signals
Check whether the system caught the beneficiary comment, hesitation, changed retirement timing, question behind the question, and any decision pattern that should alter the advisor response.
5. Ask the CCO to review the record
Evaluate source fidelity, speaker attribution, disclosure context, supervisory workflow, audit history, and whether the artifact supports—not substitutes for—professional judgment.
6. Score the full chain
| Criterion | What to measure |
|---|---|
| Material capture | Important facts, commitments, planning issues, and soft signals caught or missed. |
| Follow-up trust | Advisor rewrite time, factual corrections, and willingness to send. |
| Behavioral depth | Decision patterns and action strategies—not generic positive/negative sentiment. |
| Compliance context | Speaker, excerpt, context, severity, reviewer path, resolution, and preserved source. |
| Action quality | Specific owner, due date, source, priority, and CRM execution path. |
| Downstream fidelity | Whether health scores, tasks, and dashboards inherit reliable source intelligence. |
Run the protocol on your own meetings.
RIAX will walk through the full output chain and let your advisors and CCO judge it directly.