The $1,000/hour Solo AI business (Full Course)

Greg Isenberg · 1:01:05 · Full free-course episode with Corey Ganim: a prescription-style AI consulting business for small businesses — no audience, no capital, no coding required. You diagnose time drains and prescribe off-the-shelf tools; the assessment is the foot in the door to multi-thousand-dollar engagements.

Why now

• Millions of US businesses need AI; ~5% use anything beyond prompting ChatGPT. Owners are either too busy in the weeds (40–60h weeks) or head-in-sand scared of competitors.
• The “why give away the playbook” answer: so much blue ocean that every viewer implementing it still wouldn’t dent demand.
• You’re not building anything — you prescribe existing tools like a doctor writing prescriptions (“a doctor that needs no qualifications”).

The core offer: the AI Tools Assessment ($999)

• Client sweet spot: small business owners, 2–20 employees, 500K–5M/yr revenue.
• Promise: 45-min structured interview → prescribe 3–7 off-the-shelf AI tools reclaiming 5–10 hours/week, money-back guarantee if fewer than 5 hours found (average client saves ~7). Only cost to client = 45 minutes.
• His 2026 numbers: ~15 assessments sold; ~50% convert to implementation work; LTV per assessment client $3–10K+.

Fulfillment — four phases

1. Discovery call (recorded with an AI note-taker: Fathom/Otter/Fireflies). Pure probing, zero pitching — bite your tongue when you already know the tool. Key questions: walk me through yesterday; which tasks do you dread; where does work pile up; what automation attempts failed; magic-wand question (delete one process forever? → most common answer: email).
2. AI analysis of the transcript — feed transcript to Claude (he built a skill) with research: find off-the-shelf SaaS/AI tools per pain point. QA pass = judgment calls (Claude prescribed Salesforce to a 4-person landscaping business → swap for SMB CRM). Feed past transcripts + finished reports back into the skill so it learns what good looks like: first 2–3 assessments ≈ 60–70% there, by #4–6 often copy-paste. Tool directories he uses: futurepedia.io, theresanaiforthat.com.
3. Report generation — templatized deck (built in Claude Design now, Gamma before); template downloadable at audittemplate.ai. Slides: executive summary (main pain point, outcome, hours reclaimed, primary focus = which ROI lever: money/effectiveness, time/efficiency, quality) → effort-vs-impact matrix (report targets the quick-wins quadrant) → pain-point→tool summaries (“5 hrs/wk email” → SaneBox; handwritten meeting notes → Fathom) → recommended solutions deep-dive (tool, cost, setup time, hours saved each) → four-day quick-start plan (≤10 min/day actions to fight freeze paralysis) → “what comes after quick wins” (tees up major projects = upsell seeds) → financial impact slide: weekly hours × hourly rate − tool cost (~$60/mo avg tools) = always four-figure monthly net ROI, sometimes five. Iterated this report 12× on one principle: a confused mind doesn’t implement, doesn’t get ROI, doesn’t upsell.
4. Review call (~30 min, screen-share the report, ~50% of time on recommendations slide). Then three closing questions: Which recommendation is most urgent? DIY or do you want my help implementing? What’s your timeline — is this killing you or livable for 60 more days? → 50–60% want implementation.

Why audits rule

Clients literally pay you to discover more reasons to pay you. Greg’s scaling note: once proven, 200 of the 999 into Meta ads = self-liquidating funnel.

Upsell menu (the door was only ever $999)

1. Process redesign ($3–3.5K): fix a broken process before automating it (16 steps → 7). Example: e-commerce seller drowning 10 h/wk monitoring Amazon ad spend wanted AI immediately; sell him the future-state blueprint first; automating the new process is a separate engagement.
2. Simple automation builds (~$1.5K): Zapier/Make/n8n for cut-and-dry 1–3 step flows where a Claude skill is overkill (wedding-venue ops manager duplicating Asana templates manually).
3. Knowledge systems: business broker client got 400–500 emails asking the same 5 questions per listing → custom GPT trained on each listing’s marketing package, link sent instead of his inbox. Both buyer and seller said best experience they’d had in a deal. Pulls efficiency + quality levers simultaneously.
4. Custom workflows / Claude skills packs ($3K+): build skills + reference files for their proprietary processes, train the team; attach maintenance retainers since processes change.
5. (Retainers optional around any of these: “2 knowledge systems/mo for 3K", "3 automations/mo for 2K”.)
6. AI concierge — see below, his favorite.

• Closing tricks: credit the 999 toward implementation ("basically makes the assessment free") while marking the upsell up 1K so margins hold but psychology wins; warm clients who just paid $999 convert far better than cold ad traffic.

Seven client-acquisition methods (no capital, no audience)

1. Local “AI for Business” meetup — partner with a co-working space for free room by promising 30 local owners through the door; host gives instant expert status. Format: 15 min networking + 20-min Claude presentation + collect contacts at door + follow-up within 24h with the assessment offer. SEO-like: results come from months 5–12, not event #1. Combine with method 5 (his January meetup co-hosted with a plugged-in local realtor inviting her book).
2. Door knocking — best sales education that exists; a listener hit 30 local service businesses → 5 meetings → 2 clients. Immediate-results channel opposite of meetups; ~150K people heard the advice, one did it.
3. LinkedIn DMs — no AI-slop spam; target local owners, probe pain, never pitch in message one. Voice notes work (“I probably live down the street from you” hook).
4. Free mini-audits of your network — gym/church/parent contacts; 15 minutes, show one AI win, dangle full assessment as upsell. Framing Greg suggests: “Worst case you learn two tools you’ve never heard of.”
5. Agency/professional partners — insurance agents, accountants, marketers all field client AI questions; be their referral answer (+ referral fee), follow up every 2–3 weeks (“any clients ask about AI lately?”); co-branded workshops keep both parties top-of-mind.
6. AI office hours at co-working spaces — sit in their space 1 hr/week answering anyone’s AI questions; value-add for space, free consulting for tenants, ICP access for you. Community member Dennis: pitched 10 spaces, got 2, door-knocked restaurants for $175 of gift cards as giveaways, drew 11 attendees, booked 2 follow-ups from night one.
7. Post your wins / build in public — a win doesn’t need to be huge; “someone texted me a ChatGPT question and here’s my good answer” counts. Early days: manufacture content value (“90 days of one great prompt/day”) until real client results arrive to layer in.

• Meta-point: 99% won’t do these; those who try quit in 3 days. Persist 7 days → top 0.001%.

The flagship upsell: AI Concierge (done-with-you retainer)

• Offer: two 45-min Zoom calls/month screen-sharing while you help them use Claude Cowork and build Claude skills, plus unlimited Voxer access (12-business-hour response).
• Price ladder he actually used on consecutive yeses: 1,200 → 1,500 → 1,800 → 2,000/mo. Two calls at 1,500 = **~1,000/hour effective rate** — hence the title.
• First 5 pitches of this offer closed 5 → $8K MRR in 10 days.
• First call = foundation only, via a Claude plugin walkthrough: connect tools, context files, global instructions, demo scheduled tasks and skills. Every later call runs AOA per workflow: Audit (show me how you do it today) → Optimize (13 steps → 7) → Automate (turn it into a Claude skill). Rinse-repeat through their whole operation.
• Premium feel, near-zero load: pre-engagement Google Form defining the 90-day win (4 of 5 clients hit theirs within ~60 days — reference it on renewal conversations), plus a Notion hub inventorying everything accomplished, emailed after every call. The scary-looking unlimited Voxer produced 3 messages total across 5 clients in 2.5 months.
• Scarcity mechanics (only if legitimate): visible price-step landing page (4 spots at 800 → 1,200…), honest caps (“one slot left, two other calls scheduled this week”), then drop churned capacity into an even more premium package — services markets always have a higher luxury tier.

Strategic close

Differentiate or drown in the coming crowd of AI consultants: pick an industry vertical (ex-financial-services → “AI assessments for financial services”) or a geography (“the AI guy in Charlotte”), because niching raises prices and lowers pricing pressure — then niche down again.