The Borrowed Name Problem: Why Honest Chatbots Sell Better Than Fake Humans

You’ve seen the widget. It pops up with a warm “Hi, I’m Sarah 👋” and a smiling avatar. You ask something mildly complicated — delivery cut-offs, a bulk discount, whether the bracket fits the 40mm rail — and “Sarah” replies in 0.4 seconds with a paragraph that somehow says nothing. Then she asks if you’d like to speak to a human. Sarah was never Sarah. And the buyer knows it.

That tiny piece of theatre is what I’d call a borrowed name problem. It’s not a technology failure. It’s a trust tax. And for UK SMEs selling to other businesses, it’s a tax you don’t need to pay.

The tell isn’t the speed. It’s the performance.

Buyers are not confused by automation in 2026. They use it all day. What irritates them is being asked to play along with a fiction. A human name, a stock headshot and a fake typing delay don’t make a bot feel human. They make the company feel a bit dodgy.

That matters more in B2B than in a quick consumer purchase. A £12 order can survive a slightly clumsy automated exchange. A £12,000 quote enquiry cannot. The person on the other side is often deciding whether your firm is organised, honest and safe to buy from. If the first interaction is a small lie, everything after it gets discounted.

There’s a compliance angle too. UK GDPR expects people to understand when they’re interacting with automated systems, and regulators are increasingly alert to misleading online design. I’m not your lawyer, and this isn’t legal advice. But the practical direction of travel is obvious: don’t disguise automation as a named human and hope nobody notices.

The real cost: conversations that start in the wrong place

A fake-human bot doesn’t just annoy buyers. It corrupts the sales process.

A prospect asks a precise question. The bot deflects, reassures and offers a link. The prospect repeats the question in different words. The bot loops. By the time a human picks up, the buyer is frustrated and the rep is working from a transcript that looks like a bad first date.

Worse, the metrics lie. Containment looks healthy because the bot “handled” the chat. Resolution looks fine because nobody marked it as failed. Meanwhile, your team is spending the afternoon rebuilding trust that a borrowed name helped break.

For SMEs, that’s expensive. You don’t have a huge sales team to absorb wasted conversations. Every avoidable loop is time not spent on qualified enquiries, quote follow-ups or the awkward but valuable work of actually selling.

Disclosure without the conversion cliff

Here’s the fear I hear from sales managers: “If we tell them it’s a bot, they’ll leave.” Usually, that fear is backwards. People leave when they feel tricked. They stay when they know what’s happening and can see a route to a real person.

Good disclosure doesn’t need to be a legal disclaimer in 8pt grey text. It needs to be clear, early and useful.

Try this opener instead of a fake name:

Hi — I’m the digital assistant for [Company]. I can help with delivery times, product specs and starting a quote. If you need a person, type ‘human’ at any point and I’ll bring one in.

That does four jobs. It labels the automation. It sets expectations. It tells the buyer what the bot is good at. And it puts the escape hatch on the table immediately, which paradoxically makes people less likely to bolt.

Use a role label rather than a human name: digital assistant, virtual assistant, automated helper, chat assistant. If you want a name for brand consistency, make it obviously non-human — “Ask Acme” or “Acme Assist” rather than “Sarah from Leeds”. Avoid photorealistic avatars. Use an icon or abstract mark. Don’t fake typing delays for dramatic effect; just answer at a sensible pace.

Where the handover has to happen

A bot should not be judged by how many chats it keeps. It should be judged by how quickly it gets the right chats to the right human.

Set hard handover triggers. The obvious ones: a direct request for a person, a complaint, anything involving contract terms, a quote above a set value, a third repeat of the same question, or a sharp drop in sentiment. Add softer triggers too — mentions of deadlines, competitors, procurement, or phrases like “can you just confirm” that usually mean the buyer wants accountability.

The handover itself matters. Don’t dump the buyer into a fresh queue and ask them to repeat everything. Pass the transcript, the product SKU, the quote reference and a one-line summary. A warm handover says “we were listening”. A cold one says “we were performing”.

Three scripts that keep buyers talking

1. The honest opener

Hi — you’re chatting with our digital assistant. I’m best for delivery questions, specs and quote starts. Want a human? Just say so.

2. The useful dead end

I can’t confirm that discount myself — it depends on volume and lead time. I can pass you to a colleague with your details now, or take your email and have them reply within the hour. Which suits?

3. The clean handover

Got it. I’m bringing in Priya from the sales team. I’ve shared your questions and the product link, so you shouldn’t need to repeat yourself. She’ll be with you shortly.

Notice what these don’t do. They don’t apologise for existing. They don’t pretend the bot can do everything. They don’t use a human name to soften the blow. They just tell the truth and keep moving.

Measure honesty, not containment

If your dashboard rewards containment above all else, your bot will learn to trap people. Instead, track repeat-question rate, post-handover satisfaction, quote completion and the percentage of chats where a human request was honoured within one minute. A bot that says “I don’t know, here’s a human” faster is a better sales asset than one that keeps a buyer circling.

The name tag test is simple. If a colleague asked “is Sarah real?”, would you be comfortable answering? If not, change the name. Honest automation isn’t a conversion killer. Pretending is. And in a market where buyers can smell a scripted “hello” from three screens away, the brands that label the bot and hand over cleanly will win the conversations that actually pay.