What does an AI-services business look like after the first promising demo? In Corey Ganim's interview with Denis Laflamme, the answer is less about a single agent and more about the sequence around it: listen to a business, charge for a useful diagnosis, implement a few bounded changes, and keep meeting until the owner can operate the system.
Denis describes a free 15-minute conversation, a roughly $1,000 assessment, and a separately priced implementation. Corey introduces him as having earned more than $100,000 in first-year AI-services revenue. That is his reported revenue, not an audited result or a beginner earnings benchmark. Denis brings around 30 years of IT consulting and courseware experience to the work; the interview is not a story of starting from zero.
Watch the Interview
Credit: The client stories, fees, and results below come from Corey Ganim's interview with Denis Laflamme and the supplied transcript. Verification and privacy recommendations are our editorial additions.
A Clinic Assistant Built Around the Existing Work
At 2:43, Denis shows a speech-language pathology assistant called Natalie, built for a clinic in Gatineau, Quebec. His account is that clinicians used to spend about an hour after a session writing notes and could draft them in about five minutes with the new workflow. He also demonstrates SOAP-note drafting, analysis, goals, and follow-up tasks. These are reported workflow results; the video does not independently validate the time reduction or clinical accuracy.
Because therapy notes can contain sensitive information about children, a real deployment needs more than a good draft. A clinician must review and approve every note. The clinic should decide which records may be processed, where they are stored, who can access them, and how errors are corrected before any live-data rollout. A demonstration does not establish compliance or clinical safety.
The Sales Ladder: Diagnose Before You Build
Denis starts with a brief, free conversation about bottlenecks. If a business wants deeper analysis, he describes a roughly $1,000 paid assessment based on an approximately hour-long call, a transcript, and a report. The deliverable is a prioritized plan, not an implied commitment to build everything in it. Corey has shown a more detailed version of the assessment workflow in another video.
The useful habit is to ask for a real example before recommending software: What was delayed last week? How often does it happen? Which system holds the information? Who has to approve the output? A paid report should be useful even if the business declines the implementation.
A Four-Month Project for an Equipment Shop
Denis then describes an outdoor power-equipment business run by two owners. They handle inventory, customer calls, email, quotes, and marketing with little spare time. The proposed quick wins include dictation, drafts of social posts from product photos, email triage, inventory and invoice matching, a bilingual phone assistant, and a quote builder. The interview moves between ideas, early setup, and the longer project; it does not show that every proposed automation was fully deployed or that each produced a measured saving.
His first adoption signal is more modest and more useful: one owner tried Wispr Flow and, according to Denis, dictated about 2,000 words in 48 hours. From there he planned to connect existing tools and build a knowledge base while meeting the client weekly. That sequence lets the business discover whether the tool fits daily work before adding more autonomy.
For client-facing actions, keep a review gate. A photographed product can become a draft Facebook post; a person checks the details and publishes it. A phone assistant should have an escalation path. Inventory and invoice matches need exception handling, not silent changes to the system of record.
The $12,000 Quote Is Not the Same as $120,000 Saved
In the conversation, Denis estimates the equipment-shop project could return about 15 owner-hours per week. Corey illustrates the potential value using an assumed hourly rate, arriving at roughly $10,000 a month, or $120,000 a year. Denis says he quoted $12,000 for a four-month concierge engagement, with half due up front and weekly meetings. Those are sales assumptions and a quoted fee, not evidence of $120,000 in realized savings.
To check the business case, measure the current time spent per task, run the new workflow for several weeks, and count the time still spent correcting AI output. Multiply only the net hours returned by a value the client agrees is meaningful. If the owner uses freed time to win more sales, measure that separately; capacity is not cash until it is turned into an outcome. Include subscriptions, setup, support, and handover in the comparison.
The Smaller Referral Deal
A separate referral asked Denis for a website and backend to sell a course. The interview briefly muddles the price, then clarifies it: $2,000 for the build, one time, plus $100 per month for maintenance. Denis says the first build took under an hour after a longer discovery discussion and that he used the client's book as context for the site's voice. That is a reported production anecdote, not proof that a production-ready site, payment flow, accessibility, and security can be delivered in an hour.
The client should approve any use of her manuscript with an AI provider, inspect the final copy, and agree on what the monthly maintenance fee covers. If the work includes payments, customer accounts, or course access, test those flows before calling the site launched.
A Better First Pilot
- Choose one buyer you can reach. Ask about one repeated task and request a concrete recent example.
- Record the baseline. Time five instances or count errors and rework before recommending anything.
- Deliver a paid diagnosis. Show two or three options, costs, access requirements, and a small test the client can approve.
- Build only the first workflow. Start with approved sample data and human-reviewed drafts; add integrations after permissions are clear.
- Compare and decide. Re-measure, document failures, and let the client choose whether a larger implementation is worthwhile.
The episode promotes Corey's training community. That may be useful to some viewers, but the reported revenue and close rates are individual experiences. They do not establish typical earnings or prove that a course caused the results.
Turn the Case Study Into a Testable Offer
This prompt keeps the commercial insight from the interview while asking for consent, baseline evidence, and a stop rule before a bigger proposal.
Assess one real bottleneck
Define the buyer, pilot, evidence, and handover.
Video Chapters
| Time | What Denis covers |
|---|---|
| 0:00 | First-year revenue and background |
| 2:43 | Speech-therapy assistant |
| 4:55 | Free call and paid assessment |
| 6:00 | Equipment-shop workflow |
| 10:10 | First adoption signals |
| 14:30 | Pricing the engagement |
| 20:30 | Referral website project |
Sources and Limits
- Primary source: Corey Ganim's full interview and the transcript supplied with this post. Revenue, project prices, saved-time estimates, and build speed are the speakers' reports; we did not audit client records.
- People and resources: Denis Laflamme's Claude Academy, Corey's assessment template (email signup), and his free AI Operator Hub. These are creator resources, not independent verification of the case-study numbers.
- Our editorial additions: baseline measurement, client approvals, clinical-data caution, draft review, and realized-ROI checks are safeguards recommended for readers; the video does not demonstrate that every client implemented them.