AI Foundations for Small Business · Lesson 11 of 11

Pitching AI to an SMB Owner

Time: ~30 min reading + ~20 min worksheet Goal: Convert a discovery-validated opportunity into a signed engagement. Pricing, objections, decks, and the art of the close.

---

An SMB pitch is not a Fortune-500 pitch. Owners don't want 40 slides. They don't want a business case deck with an executive summary, an appendix, and a steering committee.

They want:

• A clear problem statement in their words
• A specific, scoped solution
• Real numbers (value, cost, payback)
• Clear risks and how you'll handle them
• A specific next step they can say yes to today

That's the whole thing. Ten slides max, often fewer.

This is the structure. Use it or adapt it, but don't skip sections.

Slide 1 — Title + Context

Cover. Company name, your logo, the date, and a one-line subtitle that captures the purpose.

> "AI Opportunity Assessment — [Company Name] — [Date] > Prepared by [Your name], [Your firm]"

Slide 2 — What We Heard

Two or three bullets summarizing the business and the top pains, in the owner's language. This is the "you were listening" slide.

> "CleanCraft is a 30-person residential cleaning business with strong growth but thin office-side capacity. The biggest time sinks Joe identified are scheduling changes and customer communication — 3–4 hours per day across the office staff."

Slide 3 — The Opportunities

The 2–3 opportunities you're recommending, one per line. Not the whole universe — your top picks. Ranked.

> 1. AI-drafted customer communication (appointment confirmations, changes, follow-ups) > 2. Automated scheduling change triage (email in → Jobber update draft → one-click approval) > 3. Monthly review narrative for Joe (turn QBO + Jobber data into a plain-English "what happened this month")

Slide 4 — The Lead Opportunity, Deep

Take #1 from slide 3 and go deep. Problem → Approach → Output.

> Problem: Office staff spends 2+ hours/day writing routine client emails. > Approach: Custom GPT configured with CleanCraft voice, common scenarios, and CRM integration. Staff drafts 10x faster, consistency improves. > Ownership: We build it, your team uses it, we tune it monthly for the first quarter.

Slide 5 — The Numbers

Value, cost, payback. Honest.

| Line item | Year 1 | Year 2 | |---|---|---| | Annual time savings value | $X | $Y | | Annual cost (all-in) | $X | $Y | | Net | $X | $Y |

Payback: X months.

Assumptions listed at the bottom — adoption rate, effective time savings %, hourly rate. Transparent.

Slide 6 — The Plan

What happens in month 1, 2, 3. Plain English.

> Month 1: Interviews, access to CleanCraft voice samples, build v1 of the Custom GPT, enable for 2 pilot users. > Month 2: Full rollout to office team, training sessions, tune based on first 2 weeks of use. > Month 3: Add opportunity #2 (scheduling triage), measure time savings, report to Joe.

Slide 7 — Risks and How We'll Handle Them

Two or three real risks. Name them before the owner does. Credibility bump every time.

> - Adoption risk: Office team is busy. Mitigation: we train in short sessions, build the tool into their daily workflow, check in weekly for 6 weeks. > - Data privacy: Client info used in AI drafts. Mitigation: Claude Team with business data terms, no client info trained into models, AUP. > - Output quality: First drafts will need editing. Mitigation: we expect this and build review into the workflow. Expect 2–3 weeks before it "clicks."

Slide 8 — The Ask

Specific. Dated. Actionable.

> Engagement: 90-day pilot, $14K one-time + $1,250/month subscriptions (20 seats) > Start date: [Date] > Decision needed by: [Date, 5–10 business days from now] > Next step: Sign attached engagement letter, we kick off [day of week].

That's the whole deck. Eight slides. 15 minutes to present. 15 minutes to discuss. Done.

Pricing is where most new consultants sabotage themselves. Two rules:

Rule 1: Price the value, not the hours

"This will save you $50K/year; it costs $18K one-time" is a winning pitch. "I charge $150/hour and I estimate 120 hours" is a losing one, even if it's the same dollar amount.

Rule 2: Have a pricing ladder

For any client, offer three options. Not "here's the price, take it or leave it." Something like:

• Good: $X — core recommendation, the one you'd pick
• Better: $Y — recommendation + ambitious add-ons
• Best: $Z — full program, including things you'd do in month 6+

Most clients pick Good. Some pick Better. A few pick Best. Offering the ladder makes the Good price feel reasonable, not expensive.

Anchoring tips

• Never lead with price. Value first, scope second, price third.
• State it once, confidently. Don't apologize for the number. Don't say "I know this might seem like a lot..."
• Wait for the reaction. Silence after a price quote is not your job to fill.
• If they push back, ask "what part feels off?" — it's rarely the total price; it's usually scope misalignment.

The five objections you'll hear most, and the honest answer to each:

"It's too expensive."

Probe: "Help me understand — is it the total, or the timing, or the uncertainty about ROI?"

Often it's the third. You can address ROI with data and pilots. You can address timing with phased engagements. If it's really the total and there's no room, ask if a smaller scope is the fit. Don't discount your main offering — that destroys credibility.

"We're not ready for AI."

Response: "Totally fair. What would 'ready' look like? Is it a data issue, a people issue, a priorities issue?"

Often "not ready" is code for "I don't trust the technology." Shift to showing them how limited-scope, human-in-the-loop work lets them learn without betting the farm.

"Can we just use ChatGPT?"

Response: "Yes, and you should. A lot of what I'd recommend wraps ChatGPT or similar — the value I add is configuring it for your voice and your workflows, and building the guardrails. Let me show you an example."

Then show a custom GPT vs. vanilla GPT on one of their actual workflows. The difference sells itself.

"Our IT guy says he can do this."

Response: "That's great — partnership with internal IT is important. Here's the split I'd suggest: they own the infrastructure, we own the workflow design and prompt engineering. Want to bring them into the next conversation?"

Never fight with internal IT. Bring them into the tent.

"Let me think about it."

Response: "Absolutely. What specifically do you want to think through? I want to make sure I've given you everything you need."

If they don't have a specific thing to think through, there's usually an objection they haven't named. Keep asking.

For anything under $25K, skip the 15-page proposal. Use a short engagement letter:

• Scope (what you'll deliver, bulleted)
• Out of scope (what you won't, also bulleted — prevents misunderstanding)
• Timeline
• Price and payment terms (50/50 or 25/50/25 common)
• Assumptions
• Change management clause
• Signature block

Two pages. A client signs a 2-page letter; they drag their feet on a 15-page proposal.

For larger engagements ($50K+), a real Statement of Work is appropriate, but even then — tight, specific, readable.

For most SMBs, the right first engagement is:

• $10K–$25K total, over 60–90 days
• Clear deliverables (three Custom GPTs + training, for example)
• One measurable outcome (X hours saved, Y% faster)
• An option to extend into phase 2

Why this range? Under $10K, owners don't take it seriously. Over $25K, they'll want to involve a committee. The $10K–$25K sweet spot is "the owner can say yes on their own authority, it feels substantial enough to matter, and it's small enough to not feel risky."

Land the first engagement. Deliver it well. Earn the right to the second — which is where the real money lives.

Three things happen right after you present:

• They say yes. Send the engagement letter within 2 hours while they're still warm.
• They ask follow-up questions. Great — it means they're interested. Answer clearly and short.
• They say "let me think about it." Set the follow-up before you leave. "I'll circle back Thursday at 2pm — does that work?"

Never leave the meeting without a next step on the calendar.

Every discovery → brief → pitch cycle takes 3–6 weeks. You'll land maybe 30–50% of the ones you pitch, with a good motion. The others aren't losses — they're future clients you stayed in touch with.

Stay in touch. Share a relevant article every few months. When AI changes (and it will), check in and say "this new thing affects [your pain] — want to grab 20 minutes?" Many of your biggest clients will close six months or a year after the first discovery.

---

• 8-slide deck — Title, What We Heard, Opportunities, Deep Dive, Numbers, Plan, Risks, Ask.
• Price the value, not the hours. Ladder your pricing.
• Address objections honestly. Five common ones — have crisp responses ready.
• Engagement letter (2 pages) beats proposal (15 pages) for first-time deals under $25K.
• Sweet spot for first engagement: $10K–$25K, 60–90 days, clear deliverables.
• Always end with a next step. Dated. Specific.
• Long game: stay in touch with the 50% who didn't say yet.

---

Next up: Your capstone. Time to actually do the work.

---

Time: ~20 min Goal: Rehearse the pitch. Find your weak spots.

---

For each objection, write your crisp 2–3 sentence response. Use your own voice, not the lesson's wording.

• "This sounds expensive. My nephew uses ChatGPT for $20 a month."

______________________________________________________ ______________________________________________________

• "How do I know this actually works? I don't want to be a guinea pig."

______________________________________________________ ______________________________________________________

• "What happens to my data? I've got client info I can't risk getting leaked."

______________________________________________________ ______________________________________________________

• "I'm not sure my team will actually use it. We tried a new CRM two years ago and half of it got ignored."

______________________________________________________ ______________________________________________________

• "Can you guarantee the ROI?"

______________________________________________________ ______________________________________________________

• "What if you go out of business or I can't reach you?"

______________________________________________________ ______________________________________________________

You're pitching a 40-person professional services firm. You've identified three opportunities from discovery:

• A. Custom GPTs for proposal and engagement letter writing (biggest pain)
• B. AI-enhanced automation for intake/document routing
• C. Internal RAG knowledge base on past engagements

Design the ladder:

Good ($X, the "easy yes"): Scope: ______________ Price: $______ Duration: ____ days

Better ($Y, your recommendation): Scope: ______________ Price: $______ Duration: ____ days

Best ($Z, the "full vision"): Scope: ______________ Price: $______ Duration: ____ days

Write what you'd say in the first 60 seconds of the pitch meeting. Start when you sit down, end when you click to slide 3 ("The Opportunities"). Practice saying it out loud.

Your opening: ```

```

Read it out loud. Is it longer than 60 seconds? Cut. Shorter than 30 seconds? Add.

---

Exercise 1 — a few sample responses:

On price: "Totally fair to start there. ChatGPT at $20 is the raw tool. What you're paying me for is the workflow design — the prompts, the templates, the integrations, the training. Think of it like buying a truck versus hiring the contractor who uses the truck to build your house. The tool is cheap. The work around the tool is where the value shows up."

On proof: "Fair. Here's how we handle that: we structure this as a 90-day pilot with a specific measurable outcome. If we don't hit the outcome, we sit down together and figure out why. I'd rather lose a small engagement than oversell you on a big one that doesn't deliver."

On data: "Good question and the most important one. Here's what I recommend [walk through business-tier tool, data handling terms, AUP, nothing used for training]. I'll put the specific data handling agreement in writing before we start."

On adoption: "That's the #1 risk, and you're right to raise it. Here's how we handle it: small pilot group first, weekly check-ins for the first 6 weeks, and we build the tool into one existing workflow rather than asking people to learn a new app. If adoption is stuck after 30 days, we course-correct or you don't pay for month two."

On ROI guarantee: "I can't guarantee a specific dollar figure — anyone who does is fibbing. What I can guarantee is the scope, the quality, the measurement framework, and the hours. And I'll show you my assumptions so you can judge the math yourself."

On continuity: "Everything we build, you own — the prompts, the configurations, the documentation. If I get hit by a bus, your staff has the playbook and any capable consultant can pick it up. No vendor lock-in."

Exercise 2 — sample ladder:

• Good ($12K, 60 days): Opportunity A only. Two Custom GPTs (proposal + engagement letter), brand voice training, 3 training sessions, one tune-up round. Claude Team licenses extra.
• Better ($22K, 90 days): A + B. Custom GPTs + one Zapier/Make automation for document routing. Monthly tune-ups for a quarter.
• Best ($48K, 5 months): A + B + C. Custom GPTs + automation + small RAG knowledge base on past engagements with 6 months of tuning.

Exercise 3: A good opening doesn't waste time on pleasantries. Try something like:

> "Thanks for the time today. I'll be quick with this — I know you've got a lot going on. I've put together a short deck with three opportunities I think are worth your attention, based on what we talked through last time. I'm going to walk through them, give you the numbers I'm seeing, and then the question for you is which one you want to start with. Ready?"

That's 20 seconds. Respectful, confident, structured. The owner immediately knows this isn't going to be a pitch-fest — it's a working meeting.

---

Check yourself: Can you handle six different objections without fumbling, build a three-tier ladder in 10 minutes, and open a pitch meeting in under 60 seconds? You're ready for the capstone.

Previous: How to Run a Discovery Conversation · Back to the course