The pack is nine production-grade prompts that map to the five most common moments inside a paid AI discovery: scoping the engagement, running stakeholder interviews, analyzing source documents, synthesizing findings into a roadmap, and packaging the deliverable. Each prompt is field-tested across real consulting engagements and ships with the input shape, the model behavior to expect, and the failure modes to avoid.
Who the prompts are for
The pack is built for solo consultants, AI agency owners, and boutique advisory teams who already sell discovery work and want to stop reinventing the prompt for every engagement. If you are running your first paid AI audit, start with the First Paid Audit Kit instead. If you are already running engagements and want to standardize how your team prompts inside them, the prompt pack is the right next step.
How each prompt is structured
Every prompt follows the same four-part structure so it is durable across model changes and predictable for new team members:
Role and scope — what the model is being asked to be, and what it is explicitly not.
Context block — the structured input the model needs (transcript, document excerpt, scoring rubric).
Quality gate — the explicit criteria the output has to meet before you ship it.
The quality gate is the part most consultants skip and the reason most prompt outputs feel "ChatGPT-ish." The gate forces the model to self-check the output against the criteria before returning, which collapses the rework loop from three rounds to one.
What changes when you switch to structured prompts
Two things shift. First, the model output stops sounding like ChatGPT and starts sounding like consulting work. The structure is the difference. Second, junior team members can run discovery work that previously required the senior consultant. The prompt encodes the senior consultant's judgment, so a junior associate can run a stakeholder interview and produce findings the senior would have written.
This is the same mechanism Audity uses internally. The product is the workflow layer wrapped around the same family of prompts, plus citation tracking, source-paragraph linking, and deliverable assembly. If you want the prompts on their own to use inside your existing stack, the pack is yours. If you want the workflow, the team handoffs, and the white-label deliverables on top, that is what Audity adds.
Where the prompts came from
The pack is distilled from roughly 80 paid discovery engagements run by Ed Krystosik and the consulting teams using Audity. Each prompt was iterated through at least 20 client engagements before it was added to the pack. The prompts that did not survive contact with real client work were dropped. What is in the file is what we still use ourselves on live engagements, not aspirational examples.
