Corey Ganim's new walkthrough answers a practical question his broader AI-services videos leave open: what happens between the client saying yes and receiving a report they can use? He shows the call, Claude-assisted analysis, impact-versus-effort priorities, report slides, and review conversation behind his $999 AI Tools Assessment.
The product is a paid diagnosis, not a software build. Interview the owner, measure specific bottlenecks, research existing solutions, put the best options in an actionable order, and review the report together. Corey's three-outcome filter is simple: every recommendation should plausibly help the client make more money, save time, or improve the quality of their service. A proposed benefit is still a hypothesis until implementation proves it.
Watch Corey Explain the Delivery
Credit: The offer, three-outcome framework, slide sequence, fee, and retainer figures come from Corey Ganim and the supplied video transcript. See his written assessment guide. The privacy and measurement safeguards below are our editorial additions, not claims that Corey demonstrated them in this recording.
What the Client Actually Buys
At 3:13, Corey describes his target customer as a non-technical owner of a roughly $500,000 to $5 million business. His offer is paid before work begins: a 45- to 60-minute discovery call, a written report with usually three to seven recommendations, and a 30-minute review. One recommendation may be conventional software; another may be a Claude workflow or skill. The assessment identifies the solution. It does not install it.
This boundary matters. A useful report should stand on its own, even if the owner never buys implementation help. Agree up front on the workflow scope, number of recommendations, data the consultant may review, delivery format, revisions, and what is expressly out of scope.
1. Run a Discovery Call That Produces Evidence
The first call is not a tour of AI products. Ask the owner to walk through where work queues, who handles it, what gets re-entered, what breaks, and what happens when it is late. At 5:45, Corey describes using Claude to prepare a plain-English question bank ordered from easy to harder questions. The important move is to request a number or example for each answer: hours per week, transaction volume, rework, missed leads, or an actual quality complaint.
Corey records the call and uses its transcript as source material. Obtain explicit consent first. Confirm where the recording and transcript will be stored, who may access them, whether AI processing is permitted, and when they will be deleted. Redact sensitive customer and employee details before sharing a transcript with an assistant. If the client cannot authorize that processing, work from approved notes instead.
2. Research Tools, Then Score the Fit
At 8:16, Corey feeds the transcript to Claude to extract pain points, search for possible tools, and place recommendations on an impact-versus-effort matrix. His paid community offers a skill for this step, but the method itself is not dependent on that skill. Claude Skills can package repeatable instructions; they do not independently verify a tool's suitability.
For every candidate, trace the problem to a sentence from the call, then check the vendor's current pricing, integrations, access requirements, limits, and data handling. Score impact against the owner's stated baseline and effort against setup, migration, training, and ongoing maintenance. A high-impact, low-effort option belongs in the quick-win quadrant. High-impact, high-effort work is a later project, not a four-day promise. Remove a recommendation if it cannot honestly serve one of the three outcomes.
3. Make the Report a Decision Tool
Corey walks through his branded report from 11:03. The valuable part is the sequence, not the slide styling:
- Executive summary: the owner's top problems, the first action, and the main uncertainty.
- Impact-versus-effort matrix: quick wins separated from larger projects.
- Recommendations: one tool or workflow per priority pain point, with cost, setup effort, owner, and a reason it fits.
- Four-day quick-start plan: small, client-approved steps to trial the first solution rather than a promise to complete every project in four days.
- Financial impact and next steps: show the client's input numbers, expected tool cost, assumptions, and what will be measured after launch.
A sample calculation: if a task currently takes three hours weekly and the client's agreed capacity value is $40 per hour, the gross potential is about $480 over four weeks. A $40 monthly tool would leave $440 of potential capacity value before training and oversight. That is not $440 of realized cash savings unless those hours are actually removed from cost or redeployed productively. Use a range when the baseline is uncertain, and do not double-count the same time in several recommendations.
Corey offers his report template through an email signup. Our separate Responsible AI Tools Assessment companion is directly accessible and adds evidence, permission, and review fields; it is not Corey's template.
4. Review the Findings Without Forcing an Upsell
On the 30-minute review call, explain why each option made the list, where it might fail, and which quick win the owner is willing to test. Corey then introduces his optional AI Concierge offer, which he says commonly costs $1,500-$2,000 per month and covers implementation, Claude training, and skills. That is a separate scope with separate accountability. Record who will implement each item, how success will be checked, and whether the client prefers to do the work themselves.
His lead-in offer is a free 15-minute mini assessment that identifies one bottleneck and one possible solution. It can be a good test of the interview process, provided it gives a useful answer even when the business declines the full report. For the fuller offer ladder, see our earlier guide.
The Two-Hour Claim Needs a Full Clock
Corey says the $999 engagement takes about two hours, yielding roughly $500 per delivery hour. The two scheduled calls alone take 75-90 minutes, leaving 30-45 minutes for transcript review, vendor research, pricing checks, ROI math, report assembly, and quality control. That pace may reflect a practiced operator with templates and a familiar tool stack. It is not a beginner guarantee. Add sales time, no-shows, revisions, software, admin, and non-billable learning when deciding your price. Corey reports selling about 26 assessments this year; this is a creator-reported count, not an independently audited result.
The defensible promise is narrower: a concise, evidence-backed starting plan. The client gets value only when someone tests the recommendation and measures the result. Our full assessment guide covers deeper discovery, privacy, and unit economics.
Turn the Framework Into a Business Idea
Use this in any AI assistant to shape a small assessment around an industry you know. It asks for a real buyer and a validation step before assuming Corey's pricing or results will transfer.
Sell a useful diagnosis
Three offer ideas and a seven-day validation plan.
Video Chapters
0:00 The assessment model · 1:43 Engagement details · 4:58 Discovery call · 8:16 Claude analysis · 9:40 The report · 14:24 Review and upsell · 15:37 Offer ladder.
Source Map
- Primary creator source: Corey Ganim's video, published 5 October 2026, and the supplied transcript. The price, audience, time estimate, assessment count, and retainer range are his reported figures.
- Creator resources: Corey's written guide, email-gated report template, and AI Operator Hub. Course membership and the free resource page are separate from this article.
- Product context: Anthropic's Claude Cowork page and Skills documentation. Access, capabilities, and prices should be rechecked before recommending them to a client.