The Pitch Meeting Problem

You know AI support would save your team time and money. But your boss doesn't want to hear about cool technology. They want to know: what it costs, what it saves, and what happens if it goes wrong.

Here's how to answer all three in a 10-minute conversation.

The Cost Slide

Keep it simple. One number.

"Per-resolution AI support for our volume costs [$X/month]. No setup fees. No contracts. We can cancel anytime."

For most small teams, this is $50 to $150/month. That's less than a nice team dinner. Present it in those terms: "For the cost of one team dinner per month, we automate 60% of our support."

If your boss needs a comparison: "We currently spend [$Y/month] on support (labor + tools). This would reduce that by [40-60%]. Net savings: [$Y - $X - new tool cost]/month."

The Savings Slide

Run the math with your actual numbers:

Current state:

• Support messages per month: [your number]
• Average time to handle one message: [your number, usually 5-8 minutes]
• Who handles support: [names and their hourly loaded cost]
• Monthly labor cost on support: [calculate]

With automation:

• Messages auto-resolved (60-70%): [calculate]
• Time saved per month: [auto-resolved × avg handle time]
• Value of time saved: [hours × hourly cost of whoever's doing support]
• Automation cost: [messages × $0.20-$0.30]
• Net monthly savings: [time saved value - automation cost]

If the founder is doing support at an opportunity cost of $150/hour, the savings are dramatic. If a $25/hour support rep is doing it, the savings are still meaningful.

Present both numbers: "Direct savings are [$X/month]. But if [founder name] gets back [Y hours/month], that time goes toward [revenue-generating activity]."

The Risk Slide

Your boss is thinking: "What if it gives wrong answers and makes customers angry?"

Answer: "The system has a confidence threshold. It only auto-responds when it's 85%+ confident in its classification. Everything below that gets sent to our team for manual handling. We control every response template — the AI never generates freeform answers. And there's no contract, so if it doesn't work, we turn it off."

This addresses the three fears:

• Wrong answers: Confidence thresholds prevent low-confidence auto-responses
• Loss of control: You write all response templates, not the AI
• Lock-in: No contract, cancel anytime, pay only for what you use

The Trial Proposal

Don't ask for a permanent commitment. Ask for a trial.

"Let's run it for 30 days. $5 free credits cover the first [X] messages. If the automation rate is above 50% and we see no customer complaints, we continue. If it doesn't work, we stop."

This makes the decision easy. The risk is near zero ($0 to $50 for a month of data). The potential upside is clear. And you've given your boss a specific success metric to evaluate.

Common Objections and Responses

"Our customers expect human support."

"They'll still get it for complex issues. The AI handles the routine stuff — password resets, billing questions, order tracking. These don't benefit from a human touch. They benefit from speed."

"What about our brand voice?"

"We write every response template ourselves. The AI decides which template to send, not what to say. Our brand voice stays intact because we control every word."

"Can't we just hire someone?"

"A part-time support person costs $1,500-$2,500/month. Automation costs $50-$150/month and handles the same volume of routine questions. We should automate the routine stuff first, then hire for the complex stuff if we outgrow automation."

"I've heard AI chatbots are terrible."

"Most AI chatbots try to have conversations with customers using LLMs. Those are hit or miss. This is different — it classifies the customer's intent and sends a pre-written response. No hallucination, no weird answers. Just fast routing."

"Let's wait and see what competitors do."

"Our competitors are already doing this. [Name a competitor using AI support if you know one.] The question isn't whether to automate support — it's whether we want to be ahead or behind."

After the Meeting

If you get the green light, set it up the same day. The faster you show results, the more credibility you have for future tool requests.

Send a one-week update: "We auto-resolved [X%] of messages. Average response time dropped from [hours] to [seconds]. Zero customer complaints. Monthly cost: [$X]."

That email is your proof that the investment worked. It also makes the next tool request easier.