Within Five Years, an AI Agent Will Be as Standard as a Phone Number.

By Khamare Clarke · August 2025 · 9 min read

In the late 1990s, a significant number of UK small businesses resisted building websites. The reasons were familiar: “My customers don't use the internet”, “I get all my business from word of mouth”, “It's too expensive”, “I'll look into it next year.” A decade later, those same businesses were scrambling to catch up with competitors who had compounded two or three years of online visibility while they waited.

The same pattern played out with Google Business Profile. When Google Maps became the default way people found local businesses, the early adopters who claimed, completed, and optimised their profiles had a head start that translated directly into leads and revenue. The late movers paid a catch-up cost in both time and money.

AI agents are the third iteration of this adoption curve. The main difference is that the curve is compressing. What took websites fifteen years and Google profiles ten years is happening in AI agents in three to five. If you are reading this in 2025 and you have not seriously evaluated what an AI agent would do for your business, you are at the beginning of the gap, not the middle.

What is an AI agent actually doing for a trades business?

Let us be specific, because “AI agent” is a phrase that can mean anything from a basic chatbot to a fully autonomous system. For a trades business (roofer, plumber, electrician, landscaper), an AI agent is a system that handles the front-of-funnel work that currently either falls on the business owner at inconvenient hours or falls through the cracks entirely.

Enquiries at 2am

A customer searches for an emergency roofer at 11pm on a Wednesday after noticing a leak. They find your Google Business Profile and WhatsApp you. If your AI agent is configured correctly, it responds within 30 seconds with a qualifying question: “What is the nature of the damage? Is it an active leak?” The conversation qualifies the lead. The agent captures the address, the urgency level, and the customer's preferred contact time. In the morning, you see a structured lead summary, not a cold message that you have to re-read three times to understand.

Your competitor with no agent responds at 9am the next morning. In the time between, the customer has probably messaged two other roofers. If one of them also has an agent that responded at 11pm and booked a survey time for Thursday morning, you have already lost the lead before your working day started.

Automatic follow-up

A prospect messages, gets a quote, and then goes quiet. Without a follow-up system, that lead is dead in the water. Most trades business owners do not have time to chase every prospect manually, and the ones they do chase are chased inconsistently.

An AI agent with a follow-up sequence sends a message at 48 hours: “Hi, just checking in on the quote we sent through. Any questions I can help with?” If there is no response at 72 hours, it sends a different message. The sequence is designed around how long the purchasing decision cycle is for your specific service. For a roofing repair, the cycle is short and urgency is high. For a full re-roof, the cycle is longer and the follow-up needs to be softer. The agent handles both without your attention.

Review requests

Google reviews are the single highest-leverage free action a local trades business can take for its search visibility. A business with 200 recent, high-quality reviews dominates the map pack in a way that keyword optimisation alone cannot achieve. The problem is that asking customers for reviews is awkward, easy to forget, and gets deprioritised in the daily pressure of running a business.

An AI agent automates this. When a job is marked complete (or when a payment is received, or when a message is sent), the agent sends a review request at the optimal time: 24 hours after the work is done, when the customer is satisfied but the experience is fresh. The message includes a direct link to your Google review page. No friction, no forgetting.

For Upgrade Roofing Solutions, this kind of structured engagement strategy contributed directly to the 538% increase in Google Business Profile interactions over 90 days. The review volume was a component of that outcome. When reviews compound, so does map pack visibility. When map pack visibility grows, so does the enquiry volume that the agent then handles.

Lead qualification before the owner sees it

Not every enquiry is worth the same time. A business owner who reviews every raw message individually is spending executive time on leads that a well-designed qualification sequence can sort automatically. An AI agent pre-qualifies: it asks for location (to check service area), job type, budget indication where appropriate, and timeline.

What arrives in your inbox is a qualified, structured lead: customer name, job description, address, urgency level, preferred contact method. You spend 30 seconds reading it and either book the survey or pass. The triage work is done. Your time is spent on the decision, not the data gathering.

What do early adopters gain concretely?

The compounding advantage of early adoption is response speed. In the local service market, the business that responds first converts at a significantly higher rate. A 2019 study by Lead Connect found that response within five minutes increases lead conversion probability by 400% compared to response within 30 minutes. AI agents make five-minute response available at all hours, not just business hours.

The second advantage is review accumulation. A business that has been automatically requesting reviews for 12 months has significantly more reviews than a business that relies on memory. In local search, review velocity and recency are documented ranking signals. The business with 150 reviews from the past six months consistently outranks the business with 50 reviews from three years ago.

The third advantage is operational. An owner who is not manually fielding evening enquiries has more mental bandwidth for the work itself. That is not measurable in the same way lead conversion is, but it is real. The businesses I have built agent systems for consistently describe the operational relief as significant within the first month.

What do late adopters pay?

Late adopters pay the gap cost. When a competitor has been operating with an AI agent for two years and has compounded reviews, response speed, and follow-up systematically, the catch-up cost is not just the cost of implementing the agent. It is the cost of recovering the review deficit, the lost leads, and the brand perception gap that built up in the meantime.

This is the pattern from the website adoption cycle and the Google profile cycle. The businesses that moved early did not just gain the tool. They gained the compounded output of the tool operating over time. Catching up to that is expensive and slow.

The question is not “do I need an AI agent?”. The question is “when will my competitors have one, and do I want to be ahead of that or behind it?” The adoption curve suggests the answer to the first question is within 18 to 36 months for the trades market in most UK cities.

What does implementation actually involve?

I build AI agents as part of the broader systems I design for clients. The implementation involves: connecting the agent to the channels where enquiries arrive (WhatsApp, web chat, SMS, email), designing the qualification conversation flow specific to the service type, configuring the follow-up sequences, and integrating review request automation with the job completion workflow.

The build takes a few days. The configuration is specific to each business. What you end up with is not a generic chatbot with canned responses; it is a system that understands your service area, your job types, your pricing structure, and your qualification criteria.

My MSc in AI at Keele University (completing 2027) informs how I think about the underlying system design. The commercial application is what I build for clients. The two things are connected: understanding how these systems behave at a technical level produces better outcomes than plugging a generic solution into a business and hoping it fits.