The test for whether a field belongs
Ask: does this field change who the lead goes to, how urgently, or what the first message says? If not, it's not worth requiring at intake — see Lead Routing Automation for how routing logic actually consumes these fields. An accounting firm deciding what to collect is a good test case — engagement type (a one-off question versus an ongoing engagement) changes routing; a caller's exact revenue figure usually doesn't.
Fields most service businesses actually need
- Contact detail and preferred channel (email/WhatsApp/phone)
- What they're looking to get done (free text is fine)
- A qualifying signal specific to your business (budget band, timeline, business type)
- Source page / campaign, so you know which channel is producing real inquiries
- Qualification status (a field the AI or your team sets, not the visitor)
Fields that sound useful but usually aren't
Detailed company size, exact revenue figures, or multi-select checklists of every possible service — these add friction without changing the next action in most setups. Add them only if a specific process genuinely depends on the answer. A recruitment agency intake is a good example of this trap — asking for a full job spec upfront usually adds friction without changing which consultant the inquiry should reach.
How this connects to intake and CRM integration
See What Data Should an AI Intake Collect for the visitor-facing side of this, and Website CRM Integration for how the data actually lands in the CRM automatically.
Questions
Should lead source always be tracked?+
Yes — it's one of the highest-value fields for understanding which channels actually produce qualified leads.
Is qualification status set by the AI or a person?+
Either, depending on setup — the AI can pre-score it, with a person able to override.
What if my CRM already has 40 fields defined?+
Not all of them need to be required at intake — this page is about the minimum that changes what happens next, not an audit of your whole CRM schema.