Field notes / AI
AI can make incoming enquiries easier to understand, route and respond to. The value is not the chatbot. It is the structure behind it.
Igniter Studio · 2 Oct 2026 · 4 min read
A lot of businesses receive enquiries in the same way.
Someone fills in a form.
An email arrives.
A person reads it.
They work out what the customer wants, whether it fits, who should handle it and what should happen next.
Sometimes this takes thirty seconds.
Sometimes it takes ten minutes.
Sometimes important context is missed completely.
This is where an AI-assisted intake system can be useful.
Not because it replaces the person reading the enquiry.
Because it makes the information easier to act on.
Most enquiry forms collect information.
Very few actually organise it well.
A customer might write:
We have an existing website but it is slow and we are not getting many enquiries. We need something new before February because we are opening another location. We are also wondering whether bookings could happen online.
A human immediately extracts several useful facts:
Traditional form fields can capture some of this, but users rarely communicate in perfectly structured data.
They tell stories.
AI is useful because it can turn that messy input into structure.
One response to poor enquiries is to make the form longer.
Industry.
Budget.
Location.
Business size.
Website URL.
Target date.
Project type.
Number of employees.
Marketing spend.
Favourite breakfast cereal.
Eventually the form becomes an application process.
That can work for some services, but it also creates friction.
AI allows another approach.
Let the customer explain the problem naturally, then extract useful details behind the scenes.
For example:
Customer message
We manage four gyms and currently use spreadsheets for memberships and maintenance jobs. Staff keep missing requests from different locations. We want a simple system where each branch can log issues and head office can track them.
The system could derive:
Need: internal operations tool
Locations: 4
Current system: spreadsheets
Primary pain: missed maintenance requests
Users: branch staff + head office
Likely capability: workflow software
Suggested priority: high operational fit
The original message remains available.
The structure simply makes it easier to process.
This distinction matters.
A good intake system might identify:
Budget not provided.
A bad one might infer:
Likely budget: $10,000–$20,000.
Unless there is a strong reason to calculate something, missing information should remain missing.
The system should clearly separate:
That makes the output easier to trust.
People often associate lead qualification with automatically deciding whether somebody is worth speaking to.
That is only one use.
Qualification can also mean understanding what kind of conversation should happen next.
For example:
Website enquiry
→ ask for current URL and target launch date.
Automation enquiry
→ ask which systems are currently involved.
Internal software enquiry
→ schedule a discovery conversation.
Very early idea
→ send a short planning guide first.
This makes the next step more relevant without automatically rejecting anyone.
Once the information is structured, the system can trigger predictable actions.
An intake workflow might:
The important thing is that not every step needs AI.
Use deterministic rules where deterministic rules are enough.
AI is useful for interpreting the messy parts.
Normal software should handle the rest.
If an AI-generated draft is about to be sent to a prospective customer, someone may want to review it.
That should be easy.
The interface could show:
Original enquiry
Extracted details
Suggested response
Approve / Edit / Reject
This is usually more useful than pretending the system is autonomous.
It also creates an opportunity to improve the system over time.
If users repeatedly edit the same type of response, that is valuable feedback.
Saving time is useful.
But structured intake also creates better business data.
Over time, you can see:
That turns a contact form into a source of operational insight.
An AI intake system does not need to become a giant CRM.
A useful first version might only do three things:
summarise the enquiry
extract important facts
suggest the next action
If that saves time and improves consistency, expand from there.
The important thing is not how much AI is involved.
It is whether the person handling the enquiry can understand the situation and act faster.
That is the product.
Your next chapter starts here
Tell me what you’re working on. I’ll help you find a practical way forward.
Let’s talk about it