We updated the line on our website recently. It used to say Real Humans. No Chatbots. Now it says Real Humans, for Customers.
That is not a softening. It is a sharpening.
And the reason for the change is sitting on our main inbound number, answering calls. He is called AP, and we built him in-house. Here is why that is actually a stronger version of the promise we have always made, not a weaker one.
Real Humans.
No Chatbots.
An absolute claim about who we are
Real Humans,
for Customers.
A precise claim about what customers get
The promise that mattered, and the bit we never quite articulated
When we launched AssurePath, Real Humans. No Chatbots. was meant to land a specific point. When something is broken, a person fixes it. No AI deflection. No phone tree. No "I'm sorry, I don't understand that, please rephrase." A human, who knows your setup, picking up and getting on with it.
That promise was always about customers. About the people paying us to keep their business running. The line never had anything to do with whether the receptionist was human. Nobody has ever signed an MSP contract because they liked the way the front desk answered. They signed because the engineers were good.
But "no chatbots" full stop was painting with a broader brush than the actual promise. So we narrowed it to what we always meant. Real humans, for customers. Customer support is, was and always will be staffed by engineers. Humans who pick up the dedicated support line, know your environment and fix the problem. That has not changed and will not.
What we can do, now we have drawn the line cleanly, is be honest about everything that is not customer support. And put a smarter front door on it.
The problem we were quietly ignoring
AssurePath has two phone numbers. The customer support line goes straight to engineers. Always has, always will. The other is the main inbound number on the website. Prospects, suppliers, recruitment agencies trying to sell us SEO, the occasional "is this Pizza Hut?" all land there.
Here is the bit we do not usually talk about. Every time that main line rang during the day, an engineer had to stop what they were doing, almost certainly customer work, and pick it up. Sometimes it was a great prospect call worth the interruption. More often it was someone trying to sell us LinkedIn lead generation, or a candidate looking for a recruiter at a different company entirely, or a five-minute "do you do this" question we could have answered by email.
Every one of those calls cost a customer somewhere a few minutes of their engineer's attention. That is not what we promised when we said real engineers, real problems. It was the opposite.
What we built
AP. He is our in-house AI receptionist, sitting on the main inbound number only. He answers, asks what the call is about, takes the details if it is worth passing on, politely sees off the cold sales calls and emails the team a clean summary within seconds. New leads land in our CRM automatically. Hot prospects get flagged for callback within a couple of hours.
He does not touch the customer support line. Customers still get humans. That is the promise, real humans for customers, and AP is precisely how we keep it.
We built him in-house using the same stack we use for client AI projects through our AI Solutions practice. Same tooling, same prompt discipline, same iteration loop. We eat our own cooking.
Why the website update matters
The new line, real humans, for customers, is not a retreat. It is a more honest version of the same promise.
The old line was an absolute claim about who we are. The new one is a precise claim about what customers get. The first was easier to print on a homepage. The second is easier to defend on a phone call.
If anything, having AP on the main line protects the human-engineer promise. Every call he handles is one our engineers do not have to break customer focus for. The ratio of engineer time spent on actual engineering work has gone up. That is the only number that matters.
The old line was a claim about who we are. The new one is a claim about what customers get. The second is easier to defend.
What surprised us
A few things we did not expect when we went live.
Cold sales calls dropped to zero overhead
AP is unfailingly polite about declining. He does not get tired, does not get drawn into "just five minutes" conversations and does not take it personally when someone ignores his "we don't take sales on this line" and pitches anyway. He thanks them, points them at the contact form and ends the call. Our SDR-baiting time has gone from "occasional 10-minute distractions" to nothing.
Prospects actually like talking to him
We expected complaints. We got curiosity. Several callers asked, mid-conversation, who built it. That led to conversations about our AI Solutions practice that we would not otherwise have had. Turns out the receptionist is also an inadvertent product demo.
The data we capture is better than what humans were getting
When an engineer answered the phone between tickets, they took whatever notes they had time for. AP captures full name, company, phone, email, what the caller wanted, rough size, role, timeline and lead score. Every time, in the same format. The team's first action on a new lead used to be "ring them back to find out who they are." Now it is "ring them back to talk about the actual thing."
The first version was bad
This is the most important part. The first opener sounded fake. The first prompt let AP hallucinate reasons the team were not picking up ("everyone's gone home for the weekend", at 11am on a Tuesday). The first email-readback got stuck in a loop and made one of us shout "you are broken" at our own AI. We rebuilt it three or four times before it was good enough to keep. If you are thinking of doing the same, budget for the iteration, not just the build.
The bit that genuinely matters
We have been telling clients for months that the right way to use AI in business is to take the boring work off humans so they can do the work that actually needs them. That is the line on every AI Solutions proposal we send.
AP is us, taking our own advice. The engineers spend more time on engineering. The customers get more attention. The receptionist function, which was always a distraction from the real work, gets handled cleanly.
The website now reads real humans, for customers because that is the precise promise we can keep, every time, without exception. Customer support is human, full stop. The main inbound is AP, by design. Two channels, two jobs, one consistent value.
If you are thinking about doing this yourself
A few things we would tell ourselves if we could go back.
The opener matters more than anything else
Callers decide in the first six seconds whether they are being respected or fobbed off. Test ten versions before picking one.
Do not let the AI hallucinate
The first time AP told a caller "everyone's off for the weekend" on a Tuesday, we realised the prompt needed explicit instructions about what the AI does not know. Time of day, day of the week, where specific people are. None of that should ever be invented.
Decide upfront what the AI is allowed to do
Pricing, technical advice, commitments on timelines. Ours does none of those. That decision saved us from several uncomfortable callbacks.
The CRM integration is what turns a novelty into a business asset
Without it, you have an AI taking messages that nobody reads. With it, every call becomes a structured record someone can act on.
And finally, be precise about what you are promising. We narrowed our website language not because we had done something wrong, but because the more accurate version was the better version. The honest story is more interesting than the marketing one anyway.
If you would like to talk to AP yourself, ring our main line. He will take it from there. If you would like to talk to one of our engineers about building something similar for your own business, drop us a note at hello@assurepath.co.uk. That one is still humans only.
Where to start with practical AI
AP is one example of what we mean by practical AI. Take a boring, repetitive job that is pulling humans away from work that matters, hand it to an AI built and supervised by people who understand both the tech and your business. If that sounds useful, here is where to look next.
- AI Solutions: how we approach AI projects for UK businesses
- Workflow Automation: the boring tasks worth taking off humans first
- AI Project Rescue: for projects that have stalled at proof of concept
- MVP Creation: build a working version before committing to the full thing