The fastest way to lose a promising lead is not always a slow reply. It is replying with a polished message that ignores what the person actually asked.
Small businesses often receive enquiries through a mixture of email, website forms, social messages and referrals. The information is uneven, the urgency is unclear and the person responding has to reconstruct the context from scratch. AI can make that first pass much faster. It should not decide whether a person is worth serving or send an unchecked answer under your name.
This workflow turns a raw enquiry into a compact lead brief that a human can review in minutes.
Define what a useful lead brief contains
Before opening an AI tool, agree on the fields that genuinely change the next action. For most service businesses, a useful brief includes:
- the contact and business name, if provided;
- the problem in the prospect’s own words;
- the requested service or outcome;
- timing, budget or location signals;
- known constraints and missing information;
- a recommended next step for a human to approve.
Do not confuse this with a lead score. A number can be useful later, but it is not a reason to discard someone. The first purpose is clarity: what do we know, what do we need to ask, and who should take the conversation forward?
Capture the original enquiry intact
Store the original email, form submission or message beside the AI summary. It is the source of truth when a detail matters. If you use a form, keep the questions short and ask only for information that affects the service: desired outcome, rough timeline and a preferred way to talk.
One enquiry should produce one record in your CRM or a simple spreadsheet. This makes it possible to see where leads came from and which questions recur.
Give AI a bounded classification job
Use an instruction that makes uncertainty visible:
Read this new enquiry and create a lead brief. Extract only details explicitly stated. Include: the requested outcome, service needed, timing, budget signals, constraints, missing questions and a suggested next step. Do not infer urgency, budget, fit or intent. Quote the exact phrase when a claim could affect a proposal or promise.
The output is an internal draft. The person responsible for sales or delivery should check it before it enters the CRM or shapes a response.
Ask the smallest useful follow-up
A good qualification reply does not send a long questionnaire. It acknowledges the request, confirms the next step and asks only for the missing detail that prevents useful advice.
For example:
Thanks for getting in touch about improving your booking flow. To suggest the right next step, could you share the current website and whether you are aiming to launch before a specific date?
AI can draft this, but review the tone, names and every statement about availability. Never let a model invent a timeline, price or capability.
Use a simple review table
| Check | Question |
|---|---|
| Accuracy | Does the brief match the original enquiry? |
| Completeness | What single missing fact changes the next action? |
| Ownership | Who replies, and by when? |
| Fit | Is the request in scope, needs a referral, or needs a discovery call? |
| Learning | Did the enquiry expose a gap in the website, offer or form? |
This takes less than two minutes when the draft is good. If it takes longer, the prompt or source material needs improvement.
Measure whether the workflow earns its place
Track a small set of signals for four weeks:
- time from enquiry to first useful reply;
- percentage of leads with a clear next action;
- calls booked from qualified enquiries;
- repeated unanswered questions;
- cases where the AI brief missed or distorted an important detail.
The goal is not to maximize the number of leads labelled “good.” It is to make the first conversation more relevant and to give the team more time for real judgment.
The rule
Automate preparation, not rejection. A lead workflow should help a person understand the request better — not hide behind a score.
Start small.
Make it repeatable.
- Input: Define the job, audience, constraints and source material.
- Draft: Use AI for structure and options, not the final unchecked answer.
- Review: Check facts, judgment, tone and the consequence of a mistake.
- Deliver: Put the useful output where work actually happens.
- Learn: Record what improved, failed or should change next time.
Make the workflow yours.
Get the H² AI Workflow Starter Pack: five practical systems, prompts and decision frameworks to turn useful AI ideas into repeatable work.
Get the free Starter Pack ↘Ready to make it
operational?
The AI Business Operations Pack turns the core workflow into six reusable systems, prompts and templates for small-business operations.
Explore the €29 Operations Pack ↗