We ran the strategy through a simulated audience before a word of it shipped. Here is what they heard — and the half they missed.
Before committing to a Second Voice — a confidence-graded, expert-led channel offering interim guidance before formal guidelines exist — we generated seven stakeholder personas from the strategy brief itself and let them react across six rounds and 125 posts and reshares on stand-in social platforms. Entirely on local hardware; no data left the building. Read this as a dress rehearsal, not field data.
Read the full simulation report →
The Second Voice's real risk isn't speed — it's trust. Across every segment, the recurring objection was never "this is too fast." It was "concerns over accuracy and trust will need to be addressed."
And the strategy's own differentiator — confidence-grading — barely surfaced in the synthetic discourse. The audience re-voiced "timely guidance" far more than "graded certainty." They heard the riskier half of the idea.
The wedge: the Second Voice succeeds or fails on whether the confidence grade itself is the visible hero of the message. Ungraded, it reads as the AHA abandoning its rigor. Graded and front-and-center, it converts "incentive-free expertise" into the modern credential the brief is chasing.
| Overall posture | Cautious optimism, not rejection. The modelled arc moved from initial skepticism toward acceptance — the brief's thesis held under simulation. |
| Loudest framing | "Scientific rigor vs. early guidance that could save lives." Dr. Ashton's cautious-optimism line propagated widest; the launch post led at ~9 reshares. |
| Recurring objection | Accuracy & trust — named independently by the consumer proxy, the AHA's own voice, and others. |
| Under-communicated | The confidence-grading mechanism. Central to the brief, nearly absent in how the audience re-voiced it. |
| Competitive tell | A peer foundation framed it as a race ("will the AHA follow suit?"). First-mover positioning is available and unclaimed. |
"While it's important to maintain scientific rigor, providing early guidance could save lives."
The expert voice carries the strategy if it can hold both halves at once. This is the line that traveled — lead with it.
"Health consumers may welcome the AHA's move… However, concerns over accuracy and trust will need to be addressed."
The clearest statement of the reframe: welcome, conditional on trust being protected. Also surfaced a pediatric angle the brief doesn't name.
"…critical interim guidance between new research findings and the development of formal clinical guidelines… to protect individuals at risk."
On-message — but it sells timeliness, not graded confidence. The institution under-uses its best differentiator.
"The British Heart Foundation has been proactive in engaging with digital platforms. Will the American Heart Association follow suit?"
Peers read this as a category positioning move. First-mover credibility is contested, not given.
"Let's raise awareness about cardiovascular diseases disproportionately affecting women. Stay informed and empower yourselves."
The AI-chat surface — a rival in the brief — re-voiced the AHA's women's-heart message favorably. A co-option path, not only a threat.
Engine. MiroFish-Offline — a local multi-agent simulation: Neo4j knowledge graph + Ollama qwen2.5:14b, no cloud. Document → ontology → persona graph → social simulation → analysis.
Inputs. The AHA "Strategic Reorientation: Closing the Gaps in AHA Consumer Relevance" brief, verbatim.
Run. 7 personas · 8 rounds · Twitter+Reddit stand-in platforms · 120+ logged actions. The simulated discourse was written back into the knowledge graph live, so the full report is grounded in what the synthetic audience actually said — not just the source brief.
Confidence. This is a simulation — a structured prediction of reception for pre-testing and message-shaping, not a substitute for fielded research. Persona realism is bounded by a 14B local model; a production run tiers the model up and widens the persona set.
Why it matters. This is the "simulate" stage of the ShurIQ flywheel — a dress rehearsal between draft and publish, and — because it runs locally — the working proof of "capability without custody."