6 min read
A lead fills out the contact form on your site at 9 p.m. You see it the next morning between two other fires, glance at it, decide it looks like a tire-kicker, and leave it for later. Three days pass. By the time you reply, they have already hired someone who answered that same night. You never find out that this one was actually your best-fit client of the month, because you judged the form in four seconds while making coffee.
That is the quiet leak in most solo businesses. It is not that you lack leads. It is that every lead lands in the same undifferentiated pile, and sorting that pile by hand is exactly the kind of work that slides when you are busy. The fix is an AI-powered intake flow that reads each new inquiry, scores how well it fits, drafts a first reply, and tells you which ones deserve your attention today. You can build it in an afternoon, and it runs quietly forever after.
What This Solves and Who It Is For
This walkthrough is for any solo owner who takes inquiries through a form: freelancers, coaches, consultants, service providers, agencies of one. If leads reach you by phone instead, our guide to an AI receptionist is the better starting point, and the two pair well. The goal here is simple: never again lose a good client to a slow or absent reply, without turning yourself into a full-time inbox sorter.
What You Will Need
Three pieces, all with usable free or low-cost tiers:
- A form tool such as Tally or Typeform. Both let you build a smart intake form for free.
- An automation connector such as Make or Zapier, which is the wiring that passes your form answers to the AI and routes the result. If you are unsure which to learn, we compared them directly in Zapier vs Make for solo owners.
- An AI model with an API, meaning a way for other apps to send it text and get text back. ChatGPT, Claude, or Gemini all work.
Budget about two hours for the first build, then near-zero maintenance. Total monthly cost usually lands under the price of one lost client.
Step by Step: Build the Flow
- Build an intake form that asks the right questions. Go beyond name and email. Ask for budget range, timeline, and one open text field like “what are you trying to solve.” Those three answers are what let the AI judge fit later. Keep it to five fields or fewer so you do not scare people off.
- Connect the form to your automation tool. In Make or Zapier, create a new scenario that triggers on “new form submission.” This is the trigger that fires the whole chain every time someone hits send.
- Add an AI step that scores the lead. Send the form answers to your AI model with a clear instruction. Something like: “You are screening leads for a [your business]. Based on the answers below, rate fit from 1 to 5, give one sentence of reasoning, and flag any red flags. Answers: [insert form fields].” Ask it to return the score and reasoning in a simple, consistent format so the next steps can read it.
- Add a second AI step that drafts a reply. Feed the same answers back in with a new instruction: “Draft a warm, specific first reply to this inquiry in my voice. Reference their actual situation. Do not quote a price. End by proposing a short call.” Keep the draft as something you approve, not something that auto-sends, at least at first.
- Route by score. Add a filter or router: a 4 or 5 goes to the top of your inbox and sends you a text alert. A 1 or 2 gets a polite, pre-written “not a fit right now” reply that you approve. The middle stays in a “review later” folder. This is the step that ends the undifferentiated pile.
- Log everything to one place. Send every scored lead to a single spreadsheet or your existing tracker, so you can see over a month which sources bring the best-fit inquiries. That data alone often pays for the whole setup.
The Mistakes That Trip People Up
The most common one is letting the AI send replies automatically from day one. Do not. Run it in draft-and-approve mode for at least two weeks so you can catch the answers that miss your tone or misread a lead. Once you trust it, you can let the low-score “not a fit” replies auto-send while you keep a hand on the high-value ones.
The second mistake is a vague scoring prompt. “Rate this lead” gives you noise. “Rate fit from 1 to 5 based on budget over X, timeline within Y, and a clear problem I actually solve” gives you a signal you can route on. Spend your setup time getting that instruction right, because everything downstream depends on it. The same principle drove our AI proposal workflow guide: the quality of the output is set by the clarity of the instruction.
How You Know It Is Working
Within the first week, you should notice two things: your best leads get a reply within minutes instead of days, and you spend less time deciding who to answer first because the score already told you. Within a month, check your log. If your reply time on high-fit leads dropped and your close rate on them held or rose, the flow is doing its job. If everything is scoring a 3, your prompt needs sharpening.
The deeper win is the one a real estate agent found when she fixed her slow follow-up: speed is a competitive advantage a solo owner can actually win on. You will never out-staff a big competitor, but you can out-respond them, and now you can do it without staying chained to the form.
Your Next Move
Do not build the whole thing tonight. Build just the form and the scoring step first, and let it run for a few days scoring leads while you still reply by hand. Once you see the scores line up with your own gut, add the draft-reply and routing steps. Ship the smallest version that works, then grow it.
Related reading
- How to Set Up an AI Receptionist That Answers Your Phone When You Cannot
- How to Build a No-Code AI Chatbot That Answers Your Website’s Repeat Questions
- Zapier vs Make: Which AI Automation Platform a Solo Owner Should Actually Learn First
What is the first question you would want the AI to ask every new lead? Share it in the comments and we will help you sharpen the scoring prompt around it.
