How to package AI services so a local business understands them

A joiner does not want AI. He wants to stop losing quotes to whoever answered their phone first. A dental practice does not want conversational voice technology. It wants fewer no-shows on a Tuesday afternoon. If your packages are built around what the technology is rather than what it does for the person paying, you will spend every sales meeting explaining and very few of them closing.

Packaging is the work of translating capability into something a busy person can understand in about ninety seconds and make a decision about. Get it right and the sales conversation becomes short. Get it wrong and no amount of enthusiasm rescues it.

Start from the complaint, not the feature list

Before you build a single package, write down the complaints you actually hear from local business owners. Not the ones you imagine - the ones people say out loud, unprompted, when you ask how the business is going.

• "I miss calls when I'm on a job."
• "By the time I ring back they've booked someone else."
• "I spend my evenings doing quotes and chasing people."
• "People book and don't turn up."
• "I know I should ask for reviews and I never do."
• "My wife answers the phone and it isn't really her job."

Every one of these is a package waiting to be named. None of them mentions artificial intelligence, automation, integrations or workflows. That is the point. Your packages should sound like the answer to a sentence the client has already said.

This is also why generic packaging fails. "AI Starter", "AI Pro", "AI Enterprise" tells the buyer nothing except that there are three prices and one of them is bigger. Names that describe the outcome - call cover, enquiry handling, full front desk - do at least some of the selling before you open your mouth.

Three tiers, and why the middle one matters most

Three is the right number of options. One option is a yes or no decision, which is harder to win. Two options invite a binary comparison where the cheaper usually wins. Five options cause paralysis and long emails asking you to explain the differences.

With three, the question in the prospect's mind changes from whether to buy to which to buy. That is a considerably better question to be answering.

Design the three deliberately:

The entry tier

This exists to be real, but it is not where you want most clients. It should genuinely solve one problem properly - usually call answering and message capture - for a sole trader or a very small firm. Around £495 setup and £149 a month is a common starting point for this level, though what your market supports is something only your market can tell you.

Keep it narrow and keep it honest. Do not cripple it artificially so that it fails; a client on the entry tier who is happy will move up later, and one who feels short-changed will tell people.

The middle tier

This is the one you want to sell. It should be the package that fits the ordinary case: a small firm with a few staff or vans, taking enquiries all day, wanting them answered and booked and followed up rather than merely noted down. Around £895 setup and £299 a month sits in the range these services commonly go for.

The middle tier should feel like the sensible adult choice. Not the cheap option, not the extravagant one. When you present the three, spend most of your time here and let the other two do their job as context.

The full tier

The top tier does two things. It serves the genuinely larger client - the firm with several staff, real call volume, a proper diary to protect - and it makes the middle tier look reasonable. Around £1,495 setup and £549 a month is a common figure for broad automation across enquiries, quoting, reminders and reviews.

It must be a real package that you can deliver. A top tier that exists purely as a decoy is obvious to anyone who reads it carefully, and you will eventually sell one to someone.

Make the differences legible in one line each

The most common packaging failure is a comparison table with fourteen rows, most of which mean nothing to the buyer. The prospect's eyes glaze, he asks "so what's the difference really?", and you end up explaining verbally what the table should have done for you.

Each tier should have one sentence describing who it is for, and then no more than about five or six lines of what it includes. If a line item would need explaining, either cut it or rewrite it as an outcome.

Compare these two ways of saying the same thing:

Multi-channel conversational routing with CRM webhook integration and configurable escalation logic.

Calls and web enquiries are answered, and anything urgent gets put straight through to you.

The first sounds impressive to nobody who matters. The second describes something the buyer can picture happening on a Tuesday morning.

Package by trade where you can

A generic package has to appeal to everyone, which means it appeals strongly to no one. If you have decided to focus on particular trades - and focusing is usually the right move early on - build packages that name them.

A plumbing and heating package can talk about emergency calls out of hours, because that is a real feature of that trade's day. A dental or aesthetics package can talk about appointment reminders and rebooking, because that is where their money leaks. A lettings agency package can talk about viewing enquiries arriving in the evening when nobody is in the office.

The underlying system is largely the same. The packaging is not. A prospect reading a page that mentions his own working week believes you understand his business, and that belief is worth more than any feature you could add. There is more on this in who this suits.

What goes in the setup fee

Be specific about what the setup fee buys, because vagueness invites suspicion. It typically covers configuration of the call handling, writing and testing the conversation flows, setting up numbers and messaging, connecting whatever booking or enquiry system the client uses, and a period of tuning once it goes live.

Say that plainly. "The setup covers building it around how your business actually works, testing it properly, and adjusting it in the first couple of weeks once we see real calls coming through." That is a description of labour, and business owners understand paying for labour.

Never present the setup fee as an administrative charge. Administrative charges feel like something to negotiate away. Work does not.

Price the usage in, visibly

Packaging and usage costs interact, and the interaction is where margins go to die. Voice minutes, SMS and email volume all cost money per use, and the client who loves the service most is generating the most of it.

Build the allowance into the package definition rather than leaving it as a footnote. Something like: the monthly fee includes a stated number of call minutes and messages, with a clear rate beyond that. Business owners are entirely familiar with this shape from their mobile contracts and their card machines. Nobody is offended by it.

What does offend people is a surprise. If you sell a flat fee and then come back three months later asking for more money because they were busier than expected, you have damaged the relationship and you will probably lose the account. Establish volume during discovery and put the numbers in the package.

A high-call-volume client on an all-you-can-eat flat fee is not a good customer. It is a slow leak. Tier them properly at the point of sale, when you still have the leverage of them wanting to start.

Optional extras, and keeping them optional

Extras are useful for two reasons. They let you concede something in a negotiation without touching the headline price, and they give existing clients an obvious way to spend more with you later.

Good extras are discrete and easy to describe: an additional phone number, out-of-hours cover, a review request flow, a second location, a lead follow-up sequence. Each should have its own small monthly figure so that adding one is a minor decision rather than a renegotiation.

Bad extras are the ones that should have been in the package. If every client asks for a thing, it belongs in the tier, not on a list of add-ons. Charging separately for the obviously necessary makes the base package look deliberately hollowed out.

Say what it does not do

This feels counterintuitive and it closes more deals than it costs. Every package should carry a short, plain statement of its limits.

It does not replace you on complex technical conversations. It does not price a job. It does not handle a complaint from an angry customer, and it should not try. It will pass those to a human, which is the correct behaviour.

Stating limits does three things. It makes everything else you say more credible, because you are visibly not overselling. It prevents the disappointment that causes month-three churn. And it pre-empts the objection the prospect was already forming, which is usually some version of "but it won't understand my business".

The same discipline applies to claims. Do not build a package around a promised increase in leads or bookings. Build it around what the system verifiably does - answers, captures, books, reminds, follows up. What that produces commercially depends on the client's own pricing, capacity and close rate, none of which you control.

Test the packaging on someone who is not a prospect

Before you take a package to market, describe it to a family member who runs no business at all and has no interest in technology. If they can tell you back what it does, it is ready. If they say "so it's like a robot thing?", it is not.

The best test of all is the demo. If you can show what the package does in under a minute on a phone - see the live demo for the shape of it - then your packaging is doing its job, because the package is simply the name for something the prospect has just watched happen. Details of what sits behind each tier are on the included features page, and the pricing page shows how the levels relate to one another.