Theory time is over. Today you're going to actually do it β take a real lead description, run it through AI, customize the response, and have a ready-to-send message by the end of this lesson.
This is the workflow you'll use every day from now on. Once you've done it once, it becomes second nature. Let's walk through it step by step.
Here's the exact process you'll follow every time a lead comes through on RatedPeople, Thumbtack, hipages, or any other platform:
Step 1: Copy the lead description. When you get a notification, open the lead and copy the full job description. On your phone, long-press to select the text and copy it. Takes 5 seconds.
Step 2: Open your AI tool. Switch to the ChatGPT or Claude app on your phone. If you've already got a conversation going for lead responses, use that same thread β the AI will remember your details.
Step 3: Paste and prompt. Use one of the prompt templates from yesterday's lesson, or the simple version: "I'm a [trade] with [X years] experience, [qualifications], fully insured. A homeowner posted this job: [paste description]. Write a professional, personalized response under 120 words. Reference their specific job, ask one smart question, mention my credentials, and suggest a next step."
Step 4: Review and tweak. Read the AI's response. Adjust the tone. Add any local knowledge. Make sure everything is accurate. This should take 30 seconds.
Step 5: Send. Copy the response, go back to the platform, paste it, and hit send.
Total time: 2 minutes. And you've just sent a response that's more personalized and professional than 90% of your competition.
Let's practice with a realistic lead. Imagine this job just came through on your platform:
Lead: "Hi, I need an electrician to install downlights in my kitchen and living room. It's a 1930s semi, currently has one ceiling pendant in each room. Thinking 6 downlights in the kitchen and 4 in the living room. Also want a couple of double sockets added in the kitchen for new appliances. Based in Solihull. Hoping to get this done in the next 2-3 weeks."
Now let's see what happens when we run this through AI with the right prompt.
Let's break down why that AI-generated response would beat most competitors:
It references the specific job β "Downlights and extra sockets in a 1930s semi." The homeowner knows you've read their description, not just blasted a template.
It shows local expertise β "I've done plenty of similar work in the Solihull area." Even if the AI added this, it's the kind of detail that builds instant trust.
It asks a smart question β "Do you have a loft space above both rooms?" This is genuinely relevant to the installation and shows the electrician knows their stuff. It also gets the homeowner to reply, starting a conversation.
It mentions credentials naturally β Part P, NICEIC, insured. Not shoved in awkwardly, just mentioned as part of the flow.
It suggests a concrete next step β "Would an evening or weekend suit you better?" This gives the homeowner something to respond to immediately.
It's the right length β Short enough to read in 30 seconds. Long enough to demonstrate competence.
AI is powerful, but you can still trip up. Watch out for these:
Sounding robotic β If the AI output sounds too polished or corporate, loosen it up. You're a tradesperson, not a management consultant. "I would be delighted to assist you with your plumbing requirements" should become "I can definitely help with this."
Being too generic β If you use the same prompt without pasting the actual job description, the AI will give you a generic response. Always include the specific job details in your prompt.
Claiming things that aren't true β AI might mention qualifications you don't have or experience you haven't got. Always check for accuracy. If the AI says "I've completed over 500 similar installations" and you haven't, change it.
Forgetting to add your personality β The best responses feel human. If you normally crack a light joke, add one. If you're known for being straight-talking, make sure the response reflects that. AI gives you the structure; you add the character.
Sending without reading β Even a 10-second scan catches errors. Don't let a typo or an AI hallucination undermine an otherwise excellent response.