The Honest Case For Not Automating Yet

Review notes: Solution-aware — the only one in the set that isn't problem-aware. Strongest authority play of the seven: a vendor arguing against its own category. Evidence: GP clinic, school group, property agent, independent travel agent. Contains "MC" (medical certificate) in a verbatim quote — regional shorthand, left exact.

We sell automation software. This post argues against buying it in four specific situations. Both of those things are deliberate.

The reason is that the objections we hear on sales calls are usually better-reasoned than the answers our industry gives them. So here are four, verbatim, with honest responses.

1. "I'm scared it will break the thing that works"

A GP, who had been considering automation for years:

"I thought about this automation a long time ago already, but I'm scared of implementing it. What if I miss? I know it's going to be good at generating online consultations, generating online appointments, but I'm scared if it impairs the walk-in treatments."

And then the specific mechanism of his fear:

"Walk-in patients are the ones requiring immediate or urgent treatment. They're sick, they need an MC. So I cannot tell them, okay, you come tomorrow, because I now have an appointment."

This is not technophobia. This is a correct analysis of his business. His walk-in patients are urgent by definition. If online booking fills his day, he cannot serve the person who arrives ill at 11am — and that person is both his existing business and his ethical obligation.

Our honest answer: he's right, and most booking automation would hurt him. The version that wouldn't touches his calendar not at all — it answers questions, gives accurate information, and captures the enquiry, leaving capacity decisions entirely with him. If a vendor's answer to his objection is "our AI optimises your calendar," he should end the call. The correct response to a well-reasoned fear is a narrower scope, not more reassurance.

2. "What if it works and we can't cope?"

Same doctor, immediately after:

"I'm scared, I'm worried if we cannot meet up to the increased demand."

Rarely said out loud, and completely legitimate. A business at capacity gets nothing from more enquiries except a worse reputation. Automation that generates demand you can't serve converts a growth problem into a service-failure problem.

Our honest answer: if you are genuinely at capacity, demand generation is the wrong project. What may still help is the efficiency side — the same tools that answer enquiries faster also remove admin, and removing admin is how you create capacity. Do that first, then open the tap. Sequence matters.

3. "Everything is AI. Can I talk to a real human?"

A head of edtech at a five-campus school group, in the middle of evaluating an AI product:

"I was hoping that it's a real human, because everything is AI. We will get frustrated. Like, can I get to talk to a real human?"

She then raised something sharper — that AI has made her own inbound worse:

"Many of the leads are not genuine... AI is too common already and it's too easily done, that the leads are not genuine."

Her team burns a one-hour response SLA on submissions that were never people. AI created that problem for her before it solved anything.

Our honest answer: both points stand. Any system that traps a customer in an AI loop with no route to a person is a worse experience than a slow human — and for a school talking to a family about a visa or a safeguarding concern, it's unacceptable. The test isn't whether AI is involved. It's whether the handover to a person is fast, obvious and reliable. On fake leads: filtering helps, but she was right that the arms race is real, and we'd rather say so than pretend detection is solved.

4. "It looks overwhelming, and last time I gave up"

A property agent, after a short product walkthrough:

"As for me, I'm not quite familiar with all the systems — everything looks quite overwhelming."

An independent travel agent, on a previous tool:

"They will teach us how to do it and ask us to post on our own, which I do not know how to do, and that's a lot of restriction. That's why I just forget about it."

An international school, on why they deliberately automate less than they could:

"With this automation tool that we use, it can be set up, but once it's broken then you have to manually come in and fix it. And that takes a lot of time. So we try to minimise the amount of automation that we do... I've experienced that way too much time."

That last one deserves attention. A well-run school made a considered decision to automate less, because their previous automation was fragile and fixing it cost more than the automation saved. That's not resistance to change. That's arithmetic.

Our honest answer: the failure rate in this category is high, and the reason is usually implementation rather than product. Which is why the vendor questions in "They Sold You a Builder, Not a Solution" matter more than any feature list — who builds it, what happens in month four, and whether there's a person to call. If you've been burned, the correct posture is scepticism, and the burden is on the vendor.

When you actually should

Having argued the other side properly, the honest positive case. Automate when:

• A specific, named task is eating hours. Not "we're inefficient." An international school computed live on a call that manual applicant data entry cost them roughly 40 hours a year at ten minutes an applicant. That's a target you can hit.
• You're losing customers to timing, not to price. If enquiries go cold overnight and competitors reply first, that's mechanical.
• Quality depends on who's working. Different staff, different experience — this is a process gap, and processes can be encoded.
• You can't answer a basic question about your own funnel. A Head of Admissions who cannot state her conversion rate isn't short of effort; she's short of a system.
• The constraint is admin, not demand. Automation multiplies capacity. It's much worse at creating demand.

And when you shouldn't

• Your problem is too few enquiries. Fix demand first.
• You're at capacity and can't expand. Fix throughput first.
• Nobody owns it internally. Something with no owner becomes shelfware — one school was paying full price while using a third of the features.
• You're doing it because of a board deadline rather than a diagnosed problem.
• Your data is so poor that automating on top of it would industrialise the error.

Honest assessment

We'd rather lose a sale than sell into the doctor's situation and break his walk-in practice. Partly because it's the right call, and mostly because it's the durable one — a customer who churns in month four costs more than the one we didn't sign.

If you're in one of the four situations above, the useful next step isn't a demo. It's writing down the specific task you'd want removed, with a number attached. If you can't fill in the number, it's too early.

LinkedIn native post (link in first comment)

We sell automation software. Here are four reasons not to buy it, all from real customers.

1. "I'm scared it will break the thing that works."

A doctor: "I'm scared if it impairs the walk-in treatments. Walk-in patients require immediate treatment. They're sick, they need an MC. I cannot tell them come tomorrow because I now have an appointment."

He's right. Most booking automation would hurt his practice. If a vendor's answer is "our AI optimises your calendar," he should end the call.

1. "What if it works and we can't cope?"

Same doctor. Rarely said out loud, completely legitimate. If you're at capacity, more enquiries is just a worse reputation.

1. "Everything is AI. Can I get to talk to a real human?"

A head of edtech — while evaluating an AI product. She also pointed out AI made her inbound worse: "Many of the leads are not genuine. AI is too common already." Her team burns its 1-hour SLA on submissions that were never people.

1. "Last time, I gave up."

A school that deliberately automates LESS: "Once it's broken you have to manually come in and fix it. So we try to minimise the amount of automation we do."

That's not resistance to change. That's arithmetic.

The honest test for whether it's time 👇

LinkedIn article angle

Strongest authority play of the seven — a vendor arguing against its own category. Keep the four objections but expand the "when you actually should" section so it doesn't read as false modesty. Title: Four Good Reasons Not To Buy What We Sell.

CTA

The full case, both directions → website.

Yalla is the chat-native CRM where the AI does the work, not just the talking.

See how much of your business it can reach.