AI Client Intake and Qualification
Acknowledge and qualify an inbound inquiry within minutes at any hour, produce a structured record, and hand to a person before any commitment is made.
Runtime state: Running in the business. This is an operating procedure with tooling around it rather than a deployed service, and the human takeover before any commitment is a process rule, not a code gate.
- Business objective
- Win the inquiry on response speed without ever letting a model make a promise on the business's behalf.
- Trigger
- An inbound inquiry arrives through any channel, at any hour.
- Inputs
- The inquiry itself
- Qualification criteria and the standing question set
- Service capability and current capacity
- Outputs
- An immediate qualified acknowledgment to the inquirer
- A structured qualification record
- A routing and priority decision
- A person assigned before any commitment is discussed
The insight this rests on
Speed and commitment are separable. The thing that wins an inquiry is being first and being relevant. The thing that carries risk is promising something. Only the second needs a person, and treating them as one problem is what makes businesses either slow or reckless.
The workflow
Inquiry to assigned owner
Under two minutes to first response. Zero commitments made without a person.
- External service
Inquiry arrives
Any channel, any hour, including evenings and weekends.
- AI / LLM
Intent extraction
What is being asked for, in what context, with what apparent urgency.
- AI / LLM
Qualified acknowledgment sent
Acknowledges the inquiry and asks the qualifying questions a good account manager would ask. Contains no timeline, price or scope.
- AI / LLM
Structured record produced
Requirement, context, apparent fit and open questions, in a consistent shape regardless of who is available.
- Automated software
Routing and priority scoring
Decision point: Rules applied to the structured record decide owner and priority. Not a model decision.
- Approval gate
Human takeover before commitment
Decision point: A person reviews the qualification and owns everything from here. Nothing involving timeline, price or scope has been said yet.
- Human action
Discovery conversation
The relationship step. Not automated, and not a candidate for automation.
- AI / LLM
Brief synthesis
Conversation and requirements structured into a brief that production will inherit.
- Approval gate
Brief approved
Decision point: A person confirms the brief before production starts, because every downstream step inherits it.
- Human action
- Automated software
- AI / LLM
- External service
- Approval gate
Decision points
- Is this within our service capability? Flagged by the qualification, decided by a person.
- Priority and owner — deterministic rules over the structured record.
- Proceed to discovery, or decline? Human, always.
- Is the brief right? Approved by a person before production begins.
Responsibilities
AI / LLM
- Extracting intent and requirements from the inquiry.
- Composing the qualified acknowledgment.
- Asking the standing qualification questions.
- Producing the structured record.
- Synthesizing the discovery conversation into a brief.
Human and deterministic
- Anything involving timeline, price or scope. Human, without exception.
- The discovery conversation itself.
- Routing and priority scoring — rules, not inference.
- Approving the brief.
- Declining an inquiry.
Human approval points
- Before any commitment. The takeover gate. Non-negotiable.
- Brief approval, before production starts.
- Declining an inquiry, which is a relationship decision.
Failure handling
| Failure | Response |
|---|---|
| Intent extraction is wrong | The human takeover corrects it; the acknowledgment asked questions rather than asserting facts, so the cost is a clarification. |
| Inquiry is ambiguous | Qualification asks rather than assumes. Ambiguity is flagged for the reviewer. |
| Model or channel unavailable | Inquiry queues for human handling. Slower, never dropped. |
| Inquirer replies before takeover | Escalated immediately; a person joins rather than the assistant continuing. |
| Out-of-scope inquiry | Flagged in the record; a person declines with a referral where appropriate. |
Monitoring
Time to first response, qualification completeness, takeover latency, correction rate at takeover, and conversion from inquiry through to engagement. A rising correction rate is the signal that the standing question set needs revision.
Business value
Speed to first response is what decides who has the conversation. This captures that advantage while keeping every commitment — timeline, price, scope — with a human.
Related projects
Related workflows
- AI Content Production
Produce written content from an approved structured brief rather than from a blank page — with editing, fact-checking and a quality gate owned by a person.
- Lead Discovery and Enrichment Pipeline
Find publicly published business contacts through reproducible searches, validate and classify them, then require a human to review before anything can become a sending list.