The client who asked “Can’t a general-purpose AI do this?”: how to answer without discounting
It arrives politely: “Can’t we just do this with a general-purpose AI?” Many sellers hear it as an attack on the product and defend, or hear it as a price signal and discount. It is neither. It is a buyer telling you what they think the job is — usually a smaller job than the one you solve. The answer is a question that makes the whole job visible, and then a plain account of the part a general assistant does not do.
Say: “For part of it, yes. Which part are you picturing it doing — and what happens to the rest?” Then name the rest. The data going in, doing it every time, accuracy on their real examples, accountability when it is wrong, the record it leaves. Do not discount; a lower price agrees the comparison is fair.
What the question means
“Can’t a general-purpose AI do this?” is a statement about scope. The buyer has pictured one step of the job — writing the summary, drafting the email, answering the question — and a general assistant does that step well. They have not pictured the steps around it: getting the right data in, doing it on every case, knowing when it is wrong, and leaving a record. Those steps are invisible in a demo of one example.
The question is also a fair one. If the answer for their case is “yes, it can”, the honest response is to say so and leave. Sellers who cannot imagine that answer sound defensive on every version of the question.
The line to say
“For part of it, yes. Which part are you picturing it doing — and what happens to the rest?”
The first clause agrees, which removes the argument. The second asks them to describe the job, which is where the scope becomes visible — usually to them, while they are describing it. The third asks about the rest. Then listen; the buyer will name the step they think is the whole job, and you can walk through what surrounds it in their own workflow.
The five things a general assistant does not do for their case
| The rest of the job | How to say it for their case |
|---|---|
| Getting the data in | “It would need the call, the client record and the renewal date every time — where does someone paste those in, and who does it on Tuesday when they’re busy?” |
| Doing it every time, the same way | “Once is a demo. Every deal, every quarter, is the job.” |
| Being right on their examples | “Send me your messiest one and let’s see what each does with it.” |
| Accountability when it is wrong | “When it gets a number wrong on a client call, who notices, and what happens next?” |
| The record it leaves | “Next quarter, can the person covering for you see what was said and promised?” |
Say two or three of these, not all five, and only the ones true of their case. A list recited in full sounds rehearsed; two specific ones sound like you understood the job.
NextSay’s Roleplay rehearses the call against an AI counterpart built from your own deal, then debriefs the moment worth replaying. On the live call, Copilot surfaces the next move and the Key Signals.
Where your state or theirs requires it, tell the other person the call is being transcribed and get their consent.
Why discounting confirms the comparison
A discount in response to “can’t a general-purpose AI do this?” says: yes, we are the same category, and here is a price closer to theirs. It concedes the scope argument you have not yet made. In our view, it also risks anchoring later negotiations to a general assistant’s subscription price. If price comes up later on its own terms, tie it to the outcome the pilot measures. Do not tie it to a tool that does a different job.
If, after the walk-through, the buyer genuinely only needs the one step, say so: “Then a general assistant is the right tool for that. If the rest becomes the problem, that’s what we do.” You lose a deal you were unlikely to win, and keep the person.
Where NextSay fits
In a sales conversation, a general assistant does not listen to the live call and cue the next move while it is still useful. NextSay’s Debrief writes the record after the call: every commitment in the words actually said, and a follow-up drafted from real promises only.
Frequently asked questions
How do you respond when a client says “can’t a general-purpose AI do this”?
Agree that it can do part of it, then ask which part they are picturing and what happens to the rest. The buyer names the step they think is the whole job; you walk through what surrounds it in their workflow — the data, doing it every time, accuracy on their cases, accountability, the record.
Should you discount when a prospect compares you to a general-purpose AI?
No. A discount concedes that the comparison is fair and risks anchoring later negotiations to a general assistant’s subscription price. If price comes up on its own terms later, tie it to the outcome a pilot measures.
What does a general AI assistant not do for a business workflow?
Get the right data in every time, run the same way on every case, hold up on the customer’s messy examples, be accountable when it is wrong, and leave a record the next person can use. Those steps are invisible in a one-example demo and are usually where the work is.
What if a general-purpose AI really is enough for the client?
Say so and leave the door open: “Then a general assistant is the right tool for that step. If the rest becomes the problem, that is what we do.” You lose a deal you were unlikely to win and keep the relationship.
Make the whole job visible, then answer for the rest.
NextSay Sales is the sales desk of NextSay AI: it prepares, researches and rehearses each sales call, guides the live call with the next move and Key Signals, and writes the record and follow-up.
Where your state or theirs requires it, tell the other person the call is being transcribed and get their consent.