“It’s not only building the skill, it’s building the knowledge base behind the skill — it’s taking what I’ve given, and it’s asking me all sorts of information about, as I built this index, where did this come from? Where did that come from? Do you have the background on this? Bring them in. Meanwhile, it’s analyzing all this in the back… it’s saying, oh, there’s additional information here from the frameworks that you’ve given me.” — Don Back
Session context: 2026-04-09_Mastermind — Don Back shared his in-progress work onboarding 20 participants in a professional development program, describing how he combined multiple assessment instruments into an AI-powered coaching workflow that surfaced things no single instrument revealed.
Core Idea
Most coaches who bring AI into their practice use it as a smart document processor: they feed in one piece of client data and ask Claude to summarize or interpret it. This is useful but limited — you get AI-quality analysis of one source at the same quality you’d get from a smart assistant reading that source.
Don Back’s Multi-Instrument Client Profile approach is fundamentally different: it gives Claude four sources simultaneously and asks it to find what they reveal together that none reveals alone.
His specific combination for 20 professional development participants:
- Myers-Briggs Type Indicator — communication style and cognitive preferences
- OCEAN (Big Five) — personality structure and how they perceive the world
- Career Claimers Index — attachment to academic environment vs. readiness for industry; current agency level
- 15-minute structured interview transcript — open-ended narration on 6 standard questions
The key insight emerged when Claude processed all four for two participants who happened to be in a relationship. They referenced each other in their interviews. Claude caught this cross-reference, performed meta-analysis between their profiles, and used each to confirm and extend the other’s data — surfacing relationship dynamics that would have been invisible from any single instrument.
This is AI operating as a triangulation engine, not a summarizer.
The knowledge base Claude built in the process: As Don worked with Claude to develop the skill, Claude kept asking: where did this index come from? What’s the background on this law you reference? What’s the source for this framework? This wasn’t friction — it was Claude building the knowledge context that would make every future run of the skill more accurate. The skill, when mature, would carry the full reasoning behind Don’s Career Claimers Index, not just the output format.
The human-in-the-loop design: Don built in a critical pause point: the client-facing report never gets generated until Don has reviewed and approved the coaching insights. AI produces the analysis; the coach approves the delivery. This is the right boundary for high-stakes relational work where the AI might be accurate but the framing needs human calibration.
What “blown away” actually meant: Don described his first two test cases as results that “blew him away.” This wasn’t about Claude being impressive — it was about the meta-analytic output being qualitatively different from what he’d get doing this manually. The AI found patterns across instruments that would require significant deliberate effort to surface by hand, and it surfaced them as a natural output of the profile generation.
Practical Application
The Multi-Instrument Profile Template (adaptable for any coach’s existing assessment stack):
- Choose 3–4 instruments that triangulate on different dimensions of the client (personality, behavior, readiness, communication, history — pick dimensions that genuinely don’t overlap)
- Build the knowledge base first — before automating, feed Claude all your frameworks, the rationale behind your assessment choices, and any canonical background. Let it ask questions. Answer them. This builds the reasoning layer the skill will use.
- Run 2 manual test cases with real client data before automating. Identify where outputs surprise you (good signal) or miss (refinement signal).
- Build pause points at every client-facing output. AI generates; coach approves; client receives.
- Look for cross-references — if any clients know each other, reference each other, or share a context, flag that explicitly. Claude will use those connections productively.
What to look for in the output: The most valuable thing the multi-instrument profile produces isn’t the summary of each instrument — it’s the synthesis that names what the instruments disagree on and why, and what patterns emerge across all of them that none reveals alone.
Related Insights
- Insight - Use AI to Simulate Behavioral Interviews Before They Happen — related use of AI in the client assessment context
- Insight - Trust Before Automation in High-Value Relationships — the principle that governs the human-in-the-loop pause point design
- Insight - The Grounded Query Principle — Context-Locked Answers Reduce Hallucination and Increase Trust — the knowledge base building process ensures grounded analysis
- Insight - Use Claude to Evaluate Whether a Project Should Become a Skill — Don’s process of manual testing before automation follows this insight’s logic
Related Briefs
Evolution Across Sessions
This establishes the baseline for Multi-Instrument Client Profiling as a coaching application. Prior sessions have covered AI in client assessment contexts (Insight - Use AI to Simulate Behavioral Interviews Before They Happen) and the trust-first principle for automation in high-value relationships (Insight - Trust Before Automation in High-Value Relationships). This insight extends both: it shows what happens when you give AI not one data source but four, and demonstrates that the triangulation capability is where the real leverage lives. Future sessions should track Don’s progress as the skill matures and whether the pause-point design holds up with real clients.