AI Consulting Business: How to Start One in 2026 (Step-by-Step)
Most people who want an AI consulting business start by trying to build something — an automation, a chatbot, a custom tool. The model in this guide does the opposite. It sells a $999 AI Tools Assessment: a roughly 45-minute structured interview with a small business owner, followed by a prescription of three to seven off-the-shelf AI or SaaS tools. This guide breaks down the offer, the four-phase fulfilment system, the upsell ladder, the seven client-acquisition methods and a realistic first-30-days plan, based on Greg Isenberg's Startup Ideas Podcast episode with guest Corey Ganim. Every pricing, conversion and income figure below is Corey's own claim, not a verified result.
What You Will Find in This Guide
Jump to any section.
What Is the $999 AI Tools Assessment Business?
The AI Tools Assessment is a productized consulting offer for small business owners. Rather than implementing anything, the consultant runs a structured discovery interview, records it with an AI note-taker, feeds the transcript into Claude for analysis, produces a short templatized report, and walks the client through it on a review call. The deliverable is a prescription of three to seven off-the-shelf AI or SaaS tools the owner can sign up for and start using immediately. Corey describes the target client as a small business owner with roughly two to 20 employees and $500K to $5M in annual revenue. The offer carries a full-refund guarantee tied to finding at least five reclaimed hours per week, which is what makes a $999 price feel low-risk to a busy owner. What makes it different from a done-for-you automation agency is that nothing gets built on the first engagement — the consultant only prescribes tools that already exist.
Who Should Start This Business?
- Solo AI consultants: They use the assessment as a paid foot-in-the-door offer, then upsell implementation work to roughly half of buyers. It lets one person run a full sales and delivery cycle without hiring.
- Freelancers adding an AI line: Designers, marketers and ops freelancers bolt the assessment onto an existing service list to reach business owners they already know. The report template means delivery does not require deep technical skill.
- Agency owners: Marketing, accounting and insurance agencies use the assessment as a paid diagnostic before pitching larger retainers. It converts an existing client relationship into a structured, billable first step.
- Career changers with AI fluency: People who already understand AI tools well use the model to convert that knowledge into revenue without coding. The work is interviewing, research and reporting rather than engineering.
- Consultants who want to sell only large done-for-you builds — the $999 assessment is deliberately a small first step, and Corey says lifetime value comes from what follows, not the assessment itself
- Anyone without genuine working knowledge of AI and SaaS tools — the whole offer rests on prescribing the right tool for a specific pain point, and Corey is explicit that you need to understand AI to do this well
The 4-Phase System and Upsell Menu
A $999 assessment with a refund guarantee that removes the client's risk
The core promise is a roughly 45-minute structured interview followed by a prescription of three to seven off-the-shelf tools aimed at reclaiming five to ten hours per week. Corey backs it with a full refund if he cannot find at least five hours per week of AI-addressable opportunity. That guarantee is what makes the price easy to say yes to — worst case the client loses 45 minutes. According to Corey, the average client ends up reclaiming about seven hours per week.
Workflow outcome: a paid diagnostic that a busy owner can approve without a procurement process.
A structured interview that probes pain instead of pitching
Phase one is a recorded Zoom or Google Meet call with an AI note-taker such as Fathom capturing the transcript. Corey asks questions like what the owner did yesterday, which tasks they dread, where work piles up, what they tried to automate and failed, and — his favourite — which process they would delete with a magic wand. He stresses that the first call is only for probing, and that he has to force himself not to prescribe on it. Saving every prescription for the report is what makes the paid deliverable feel substantial.
Workflow outcome: a transcript rich enough that the analysis phase becomes mostly research, not guesswork.
Claude turns the transcript into a researched tool shortlist
Phase two feeds the transcript to Claude with a prompt asking it to research off-the-shelf SaaS or AI tools that fix the pain points raised. Corey has built this into a reusable skill, and says Claude catches patterns and solvable problems he might have missed. He also runs a quality-assurance pass — if Claude suggests Salesforce for a four-person landscaping business, he substitutes something appropriate. He points to directories like Futurepedia and There's An AI For That as research aids, the latter filterable by industry such as real estate.
Workflow outcome: a first draft shortlist in minutes that a human then edits for fit.
A deliberately simple client deliverable built from a reusable template
Phase three produces the client-facing report, which Corey says his team iterated on twelve times to keep it clear. It opens with an executive summary listing the main pain points, the outcome, the hours reclaimed per week and whether the focus is effectiveness, efficiency or quality. Then comes an effort-versus-impact matrix spotlighting high-impact, low-effort quick wins, the tool recommendations with cost and setup time, a four-day quick-start plan, a slide teeing up larger projects, and a financial-impact slide showing monthly net ROI. It is built in Claude Design from a template Corey gives away free.
Workflow outcome: a report a non-technical owner can act on without a second meeting to interpret it.
A 30-minute walkthrough with three closing questions baked in
Phase four screenshares the report and walks the client through each recommendation — this pain point, this tool, this cost, this setup time, this time saved. Corey then asks three questions: which recommendation is most urgent, whether they want to do it themselves or want help implementing, and what their timeline is. He says that roughly 50%, sometimes 60%, of clients then want implementation. The review call is where the assessment stops being a report and becomes a pipeline.
Workflow outcome: a natural, non-pushy transition from a $999 diagnostic into implementation revenue.
Five implementation offers plus a recurring retainer
Corey lists process redesign, automation builds, knowledge systems, custom workflows and full implementation packages as the upsell menu. He has sold process redesigns for $3,000 and $3,500, a simple Zapier build for around $1,500, and describes full implementation packages as combinations that might reach $8,000. His favourite close is crediting the $999 assessment toward the implementation while marking the implementation up, so the client feels a discount and the math stays profitable. He puts assessment-client lifetime value at $3K to $10K or more.
Workflow outcome: a small paid engagement that funds and feeds much larger work.
7 Ways to Find Your First Clients
Host a local AI-for-business meetup
Corey's first client-acquisition method is hosting a local AI meetup. He suggests bartering with a co-working space for a free conference room in exchange for bringing business owners into their space, then running 10-15 minutes of networking, a roughly 20-minute presentation on using Claude, and collecting contact details at the door. Follow-up happens within 24 hours with an offer of the assessment. He likens it to SEO: clients tend to come from the third to sixth event rather than the first, and he plans to build it into a monthly local event brand.
Knock on doors for immediate results
Door knocking is the method Corey says produces the fastest results — if you need a client today, you knock today. The approach is to ask for the owner or manager at local service businesses. He relays a listener who told him he visited 30 local service businesses, set five meetings, and converted two into clients. Corey is candid that almost nobody follows through on this method, which is exactly why the competition for it is so low, and he calls it the single best way to learn sales.
Send local LinkedIn DMs that ask, not pitch
Corey is clear these are not the spammy AI DMs everyone gets. You target business owners in your own town, ask how they currently use AI and where it hurts, and never pitch in the first message. His example opener is a short note along the lines of “I probably live down the street” offering some free feedback on where AI could help, and he suggests a voice message can work too. He argues the competition for these methods is low because most people quit after a few days.
Offer free mini-audits to your own network
Text the business owners you already know, from the gym, church or your kids’ school, and offer 15 free minutes to show one way AI could help their business. You prescribe one tool, and the full paid assessment becomes the upsell. Corey’s framing tip is to present it as “worst case you learn a tool or two you had not heard of”, which makes it easy to say yes to.
Build agency referral partners
Insurance agents, accountants, marketing agency owners, coaches and consultants all have clients who ask them AI questions they cannot answer well. Corey's approach is to offer to act as the AI resource for those clients and pay a referral fee for anyone who becomes a client. The critical part is follow-up: he says you should check in every two to three weeks asking whether any client has asked about AI lately, because a referral partner rarely materialises from a single outreach. He also suggests co-branded workshops as a way to stay top of mind while producing value for their audience.
Run AI office hours at a co-working space
The office-hours play is to approach a co-working space and offer to sit there for an hour a week as a free AI resource for their tenants. The space gets a value-add to offer members, tenants get free consulting, and the consultant plugs directly into their ideal client profile. Corey cites Dennis from his community, who contacted about ten co-working spaces, got two to agree, collected $175 in gift cards from local businesses by door knocking to use as a draw, and got 11 attendees at his first session with two follow-up calls booked.
Post your wins in public
The last method is documenting what you do as you go, through short videos, posts on X or images. Corey stresses a win does not have to be big: a friend texting a ChatGPT question and you giving a good answer counts. Greg suggests manufacturing consistency early on, for example sharing one useful prompt a day for 90 days, and layering in real client results as they arrive.
The Offer Ladder and Price Points
Pricing here is not a software subscription — it is the consultant's own offer ladder, and every figure below is Corey Ganim's stated number from the episode, not a verified market rate. The entry point is a $999 AI Tools Assessment, backed by a full refund if he cannot find at least five hours per week of AI-addressable opportunity. Implementation upsells are quoted per project: Corey says he has sold process redesigns for $3,000 and $3,500, a simple Zapier build for around $1,500, and describes full implementation packages that might reach $8,000. He puts assessment-client lifetime value at $3K to $10K or more. The recurring piece is the AI Concierge retainer, which he says he priced from $1,200 up to $2,000 per month as clients kept saying yes. Because these are one consultant's figures, treat them as a starting reference point for your own pricing rather than a benchmark.
| Plan | Price | What You Get |
|---|---|---|
| AI Tools Assessment Best Value | $999 one-time | Roughly 45-minute discovery call, AI-assisted analysis, templatized report and a review call. Full refund if at least five reclaimed hours per week cannot be identified. Corey's stated figure. |
| Implementation upsells | $1,500-$8,000 per project | Process redesign, automation builds, knowledge systems, custom workflows or full implementation packages. Corey's stated figures: $3,000-$3,500 for process redesigns, around $1,500 for a simple Zapier build. |
| AI Concierge retainer | $1,200-$2,000/month | Two 45-minute working sessions per month building Claude skills and teaching Claude Cowork, plus unlimited Voxer access with a 12-business-hour response window. Corey's stated figures. |
Honest Pros and Cons
- Low delivery risk on the first engagementYou are prescribing tools that already exist rather than building anything, so the assessment can be fulfilled without engineering work.
- The guarantee makes the sale easyA full refund tied to finding five hours per week means the client's worst case is losing 45 minutes, which removes most of the friction from a $999 decision.
- It is a paid diagnostic, not a free pitchThe client pays you to uncover opportunities to pay you more, which is a far better position than pitching cold.
- The report is templatized and reusableBecause the deliverable is built from a template, delivery time drops sharply after the first few assessments.
- It feeds a real upsell ladderCorey says roughly half of clients buy implementation, which is where he puts lifetime value at $3K to $10K or more.
- You genuinely need to understand AI toolsThe entire offer rests on prescribing the right tool for a specific pain point, and Corey is explicit that this is the one thing you must know.
- The guarantee creates real delivery riskIf you cannot find five hours per week of opportunity, you refund the fee — after having done the work.
- Early assessments need heavy human editingCorey says the first two or three AI analyses get you only 60-70% of the way there before the skill is dialled in.
- The acquisition methods reward persistence, not speedMeetups, agency partnerships and office hours compound over months; Corey compares the meetup to SEO, with clients arriving from the third to sixth event.
- The DealbreakerIf you are not willing to do uncomfortable, unglamorous outreach — knocking on doors, sending DMs, hosting events — the model stalls. The offer is easy to fulfil but hard to sell without consistent prospecting, and Corey is blunt that almost nobody sticks with these methods long enough to see them work.
How to Launch This Business Step by Step
Get genuinely fluent in the tools before selling anything. Corey's whole model depends on knowing which off-the-shelf AI or SaaS tool fixes a specific business problem, so spend time in Claude, in AI note-takers and in tool directories like Futurepedia and There's An AI For That before you take a paying client.
Download and adapt the report template. Corey gives away the exact template he uses, built in Claude Design, so you can plug in a client name, date, business type and primary focus rather than designing a deliverable from scratch. Read it end to end so you understand what each slide is doing before you customise it.
Run your first two or three assessments for free with people in your network. Corey's fourth acquisition method is exactly this — a short message offering 15 free minutes to show one way AI could help, framed as 'worst case you learn a tool or two'. Expect the first few AI analyses to need heavy editing; that is normal and it is how the skill gets dialled in.
Build the analysis into a reusable Claude skill. Once you have a few finished transcripts and reports, feed them back so the analysis knows what good looks like. Corey says that by the fourth, fifth or sixth assessment the output is often close to copy-paste, which is what makes the economics work.
Start one acquisition channel and stay with it for at least seven days. Pick the method that fits your temperament — door knocking for immediate results, LinkedIn DMs to local owners, a meetup if you want to be seen as the local expert, or office hours at a co-working space. Corey's point is that most people quit after three days, and simply persisting puts you ahead of nearly everyone.
Use the review call to open implementation, then layer a retainer. Ask the three closing questions — what is most urgent, do you want help, what is your timeline — and credit the $999 toward the implementation if you need to close it. Once a client is working with you, Corey's AI Concierge retainer is the natural next step: two working sessions a month plus async access, which he says he priced from $1,200 to $2,000 per month.
Key Ideas From the Episode
Paraphrased points made by the guest and host in the episode — attributed to them, not anonymous quotes.
This is the single most important framing in the episode. The $999 is not the business — it is the paid entry point to the business. If you evaluate this model on the assessment fee alone, it looks like modest freelance work; if you evaluate it on what follows, it looks like a consulting practice. Corey attributes roughly half of clients converting to implementation as the mechanism.
Corey admits he has to bite his tongue because he often knows the tool within minutes of hearing a problem. Saving every prescription for the paid report is what makes the deliverable feel worth $999 rather than a free consultation. Consultants who prescribe on the discovery call undermine the entire offer structure.
Note that these were people he already knew and told directly, not cold traffic, and it is his own account of a small sample. The lesson is about offer design rather than a conversion benchmark: a done-with-you retainer with a clear deliverable and capped availability is easier to sell than an open-ended consulting arrangement. Treat the numbers as his claim, not a forecast.
How This Compares to Other AI Service Business Models
| Decision Area | The $999 AI Assessment | When Another Option Wins |
|---|---|---|
| Best suited for | Consultants who want a paid, low-risk entry offer that opens larger implementation work with small business owners | A done-for-you AI automation agency wins when the buyer already knows what they want built and is ready to fund a build immediately |
| Pricing position | $999 one-time entry point, with implementation upsells Corey puts at $1,500 to $8,000 and a retainer at $1,200-$2,000 per month | Freelance AI content services win on speed to first invoice because they can be sold as small, repeatable deliverables with no diagnostic phase |
| Primary differentiator | Nothing is built on the first engagement — you prescribe off-the-shelf tools, which keeps delivery risk low | An automation agency differentiates on owning the build, which commands higher project fees when the client values a turnkey outcome |
| Ease of onboarding | The templatized report means a new client can be onboarded in a single discovery call | AI training workshops onboard faster in one sense — there is no per-client research phase — but they scale through audience, which most beginners do not have |
| Team collaboration | Designed for a solo operator; Corey caps his concierge roster at six clients | A done-for-you agency scales by adding delivery staff, which the assessment model does not naturally support |
| API and integrations | Not applicable — the deliverable is a report and a set of tool recommendations, not software | An automation agency's value is precisely the integrations it builds, which is a different and more technical sale |
| Long-term scaling | Scales by raising price and adding retainers rather than headcount, which suits a solo consultant | Workshops scale through audience and repeat events, which can outgrow a one-to-one consulting model over time |
The $999 AI Assessment vs Done-for-you AI automation agency
A done-for-you agency sells the build itself — Zapier flows, custom workflows, knowledge systems — rather than a diagnostic that precedes it. Corey's model actually feeds this kind of work as an upsell, and he describes automation builds, knowledge systems and custom workflows as part of his own menu. The difference is sequencing and risk: the agency leads with a build and needs the technical skill to deliver it, while the assessment leads with a report and defers the build. Zapier and Make are the tools Corey names for the simple end of that work.
Choose The $999 AI Assessment if: You want a paid first step that you can deliver without engineering work and that naturally surfaces build projects Choose Done-for-you AI automation agency if: You already have strong automation skills and want to sell the build directly at a higher project fee
The $999 AI Assessment vs Freelance AI content services
Freelance AI content work — writing, editing, repurposing — is the fastest of these models to first invoice because the deliverable is familiar and the buyer already understands what they are paying for. It is also the most exposed to price pressure, since the output is comparable across providers. The assessment model is slower to sell because it requires the buyer to accept a diagnostic premise, but it positions the consultant as an advisor rather than a pair of hands. Content services also rarely produce the implementation upsell that Corey says drives lifetime value.
Choose The $999 AI Assessment if: You want to be positioned as an advisor and build toward implementation and retainer revenue Choose Freelance AI content services if: You need revenue quickly and already have a portfolio of content work to show
Frequently Asked Questions
How does the $999 AI Tools Assessment work exactly?
It runs in four phases. First, a roughly 45-minute recorded discovery call where you probe for bottlenecks without prescribing anything. Second, you feed the transcript to Claude to research off-the-shelf tools that fix the pain points, then quality-check the output yourself. Third, you generate a templatized report with an executive summary, an effort-versus-impact matrix, recommendations, a four-day quick-start plan and an ROI slide. Fourth, a review call where you walk through it live and ask three closing questions.
Is this an AI consulting business model free to start?
The acquisition methods Corey describes require no capital and no audience — meetups, door knocking, LinkedIn DMs, free mini-audits for your network, agency partnerships, co-working office hours and posting your wins. What it does require is genuine understanding of AI and SaaS tools, plus the willingness to do consistent outreach. Corey gives away the report template free, so the main investment is your time and existing AI fluency.
What are the best alternatives to this model in 2026?
A done-for-you AI automation agency leads with the build rather than a diagnostic, which commands higher project fees but requires real technical delivery skill. Freelance AI content services reach first invoice fastest but face the most price pressure. AI training workshops scale through audience but need an audience to work. The assessment model sits between them: slower to sell than content work, lower delivery risk than a build agency.
Who is the AI Tools Assessment best suited for?
Corey's target client is a small business owner with roughly two to 20 employees and $500K to $5M in annual revenue — busy enough to have real bottlenecks, small enough to make a decision without a committee. On the consultant side, it suits a solo operator who wants a paid entry offer and is willing to prospect consistently. It suits people who already understand AI tools and does not suit anyone hoping to skip that step.
What tools does the assessment actually recommend?
Corey says many recommendations are ordinary SaaS tools with no AI involved at all — things the owner simply did not know existed. He names SaneBox as his most-prescribed tool for getting out of email and Fathom for meeting notes. Where a workflow is a clear bottleneck, the recommendation is often a Claude skill, which is what makes the AI Concierge retainer a natural follow-on.
Key Takeaways
- The core offer is a $999 AI Tools Assessment — a roughly 45-minute structured interview that ends in a prescription of three to seven off-the-shelf AI or SaaS tools, backed by a full-refund guarantee tied to finding at least five reclaimed hours per week
- Fulfilment runs in four phases: a recorded discovery call, an AI-assisted analysis in Claude, a deliberately simple templatized report, and a review call that Corey says converts roughly half of clients into implementation work
- The report stays short on purpose — executive summary, effort-versus-impact matrix, quick wins, a four-day quick-start plan and a financial-impact slide — because a confused client does not implement, and a client who does not implement does not upsell
- The upsell ladder spans process redesign, automation builds, knowledge systems, custom workflows and full implementation, with Corey putting assessment-client lifetime value at $3K to $10K or more
- The recurring piece is the AI Concierge retainer — two 45-minute sessions a month plus async access — which Corey says he priced from $1,200 to $2,000 per month and capped at six clients
Related Reading
- Claude — The analysis engine at the centre of phase two, and the tool Corey teaches clients to build skills in during the concierge retainer. Choose it if you want a single model handling transcript analysis, report drafting and skill building.
- Fathom — The AI note-taker Corey uses to record discovery calls and produce the transcript that feeds the analysis. Choose it if your priority is reliable meeting capture with minimal setup.
- Claude Design — Where Corey builds the client-facing report from his reusable template. Choose it if you want a templatized deliverable you can regenerate for each new client quickly.
- Zapier — The tool Corey names for simple automation builds in the upsell menu, such as a repeatable multi-step task with a clear input and output. Choose it if you want to sell lightweight automation work without writing code.
- AI Business Ideas in 2026 — the broader guide to making money with AI, with more business models to compare.
- Best AI Tools for Agency Owners — tools worth knowing if you start selling AI services to clients.
Bottom Line: Is This AI Consulting Model Worth Trying?
Bottom Line: This is one of the more accessible AI consulting business models for a solo operator, because the first engagement prescribes tools rather than building them — which keeps delivery risk low and makes the $999 price easy to justify with a refund guarantee. It is genuinely well suited to someone who already understands AI tools and is willing to prospect consistently through meetups, door knocking or office hours. It is not suited to anyone hoping to avoid outreach, or to anyone without real AI fluency, because the entire value of the assessment rests on prescribing the right tool for a specific problem. Treat every figure here as Corey Ganim's stated claim from the episode, not a guaranteed outcome.>
Last Updated: September 2026 | Written by theaitoolsbox.com editorial team
