Ask a service business where AI could help and most people describe content. Write the posts, write the emails, make the images. Reasonable, and rarely the thing that changes the month.
The bigger number is usually sitting in the gap between when someone enquires and when a human being responds.
What the research actually says
The most cited work here is a lead response study led by Professor James Oldroyd, then at MIT Sloan, carried out with InsideSales.com and published in 2007. It analysed a large volume of leads and call attempts across several companies and found that the odds of making contact were roughly 100 times greater when the first call went out within five minutes rather than thirty, and the odds of qualifying the lead around 21 times greater.
Two honest caveats. It is nearly two decades old, and it was conducted with a company selling sales-acceleration software. Treat the exact multipliers as indicative rather than precise.
The mechanism, though, has not aged. Someone enquiring about a service is usually enquiring with more than one provider, in a window when the problem is on their mind. Being first is not a trick. It is being there while the decision is still open.
Why most businesses are slow
Almost never because they do not care. The enquiry arrives while the owner is on a roof, under a sink, or with a customer. It lands in an inbox nobody is watching. By the evening there are eleven other things, and by Thursday the person has hired someone else.
This is a coverage problem, not an effort problem. Which is exactly the kind of problem automation is good at.
What to automate, and what not to
Automate the acknowledgement, the triage and the routing. A message within a minute or two that confirms the enquiry arrived, sets expectations for when a person will respond, and asks the two or three questions your team always has to ask anyway.
Do not automate the quote, the diagnosis, or anything that reads as a considered human judgement when it is not. Customers forgive an obvious automatic acknowledgement. They do not forgive discovering that a message they took seriously was generated.
The line is straightforward: automate the part that says “we have you and here is what happens next.” Keep the part that requires knowing something about their situation.
Where AI adds something beyond a canned reply
An autoresponder has existed for thirty years. What is newer is useful triage: reading the enquiry, working out which service it concerns, whether it is inside your service area, how urgent it sounds, and what the person actually asked. Then routing it and drafting a reply your team edits rather than composes.
That turns a fifteen-minute task into a two-minute one, and it works at eleven at night. The person still presses send.
Measure the gap, not the tool
Before changing anything, get three numbers: how long your median first response takes, how many enquiries never received a second contact, and how many became customers. Most businesses have never looked, and the answers are usually worse than the estimate.
Those three numbers also tell you whether an automation worked, which a dashboard of message volumes will not.
A reasonable version of this
Every enquiry gets an acknowledgement within two minutes, whatever the hour. Every enquiry is routed to a named person. Nothing sits unanswered past the next working morning. Every enquiry gets at least three follow-up attempts before it is closed. Someone can see, at a glance, what is still open.
None of that requires sophisticated technology. It requires that the process exists and that something reliable runs it.
This is the work our YALID Flow system was built around, and it is usually the fastest return available to a service business. If you want to know what your current response gap is costing, schedule a call with us and we will work it out with you.
