Executive summary
Decision briefWhen a new lead arrives, do not begin with an AI product pitch. Confirm the desired business result, map one workflow, quantify the current burden, identify systems and sensitive data, define authority and exceptions, and name the person who owns the result. Offer a bounded evaluation first. If the workflow is suitable, propose a small showcase with approved sample data and measurable acceptance criteria before any live implementation.
- Ask
- Outcome, workflow, volume, systems, current effort, errors, data sensitivity, approvals, exceptions, and owner.
- Offer
- A free fit evaluation and, when practical, one contained five-business-day workflow showcase.
- Avoid
- No guaranteed savings, broad transformation promise, live-system access, or sensitive-data transfer during initial qualification.
A realistic example
An owner submits a request saying the team needs an AI agent for customer follow-up. Instead of recommending a platform immediately, CSLM asks where inquiries arrive, how staff decide who should receive them, which details are missing, what promises require approval, and how many requests are handled each week. The first useful opportunity may be a completeness check and review queue—not an autonomous sales agent.
A controlled workflow
Move routine work forward. Route exceptions to people.
Pause and request human review when the request includes sensitive data, unclear authority, unsupported claims, unsafe autonomy, or a workflow that cannot be measured.
Before and after
Before
- The lead receives a generic AI-services pitch.
- Tool selection happens before the workflow and data are understood.
- The prospect is encouraged to send files, credentials, or customer records too early.
- A free conversation becomes an undefined promise to build everything.
After a controlled improvement
- The lead receives a prompt acknowledgment and a short set of relevant questions.
- One workflow, business outcome, owner, and baseline are documented.
- Sensitive data and credentials stay out of initial intake.
- The evaluation ends with a clear no-fit, prerequisites, showcase, or implementation recommendation.
The first questions to ask a new lead
What outcome matters?
Ask what should become faster, more accurate, less expensive, or easier to control—and why it matters now.
Where does the workflow start and end?
Identify the trigger, final output, responsible owner, and every system or person the work crosses.
How often does it happen?
Record weekly or monthly volume, peaks, backlog, average handling time, and how long customers or staff wait.
Where is effort or quality lost?
Ask about repeated entry, searching, handoffs, corrections, missed follow-up, duplicate work, and common exceptions.
Which systems hold the source of truth?
List the CRM, inbox, forms, spreadsheets, accounting tools, calendars, document stores, and existing integrations—without requesting credentials.
What information is sensitive?
Identify personal, payment, health, employment, legal, credential, or confidential business data and exclude it from the first sample unless a later approved design requires it.
What may the workflow do?
Separate reading, classifying, drafting, recommending, updating, sending, approving, and deleting. Name every action that requires a person.
What would make the evaluation successful?
Agree on a baseline, sample, measurement method, acceptance threshold, stop condition, and decision owner.
Tools and process components
- Lead acknowledgment template
- Ten-minute qualification checklist
- Current-state workflow map
- Baseline and exception sample
- Data and permission boundary
- Showcase acceptance criteria
- Implementation decision summary
What CSLM should offer
Match the next step to the evidence
Not every inquiry should become an implementation project. The first offer should reduce uncertainty and protect both parties.
| Lead condition | Offer | Deliverable | Boundary |
|---|---|---|---|
| Problem is still broad | Free fit evaluation | One workflow hypothesis, missing information, risks, and recommended next step | No implementation promise |
| Prerequisites are missing | Readiness recommendation | Process, data, ownership, or integration work to complete first | May recommend ordinary automation or no AI |
| One workflow is testable | Contained five-business-day showcase | Workflow map, bounded prototype or demonstration, observations, and measurement plan | Uses approved sample or redacted data; no autonomous live commitments |
| Showcase meets acceptance criteria | Paid implementation scope | Responsibilities, integrations, controls, milestones, support, and pricing for approval | Starts only after scope and access are explicitly approved |
Continue from example to implementation
Choose the next page for the workflow you want to improve.
These links continue the specific operating examples above. They do not assume that every workflow needs AI.
From first inquiry to fit review
AI consulting for small businesses
Compare the free review, contained showcase, and separately quoted implementation before sending any sensitive information.Continue with this pathMeasurement plan
Measure the workflow, not the demonstration.
| Metric | How to measure it | Useful signal |
|---|---|---|
| Response time | Time from lead arrival to human acknowledgment | Shows whether genuine inquiries are being handled consistently |
| Qualification completeness | Required discovery fields answered before solution design | Reduces premature proposals and repeated follow-up |
| Baseline burden | Current volume, handling time, corrections, delays, and exceptions | Creates a comparison point for a showcase |
| Showcase decision | No-fit, prerequisites, proceed, revise, or stop with written reasons | Prevents an experiment from becoming an open-ended project |
Safeguards to keep
- Do not ask a new lead to email passwords, API keys, payment data, health data, or complete customer records.
- Do not promise a percentage of savings before observing a representative workflow and documenting the calculation.
- Do not imply that the free evaluation guarantees a showcase or production implementation.
- Do not let a prototype send messages, approve transactions, change customer records, or make commitments without explicit scope and human authorization.
Common questions
AI consulting lead intake FAQ
What should the first reply to an AI-consulting lead say?
Acknowledge the request, state that the first step is understanding one workflow and business outcome, ask only the essential qualification questions, and tell the lead not to send credentials or sensitive records.
What is included in CSLM's free evaluation?
CSLM reviews the stated workflow, current tools, burden, data boundary, authority, and measurement opportunity. The result is a fit decision and recommended next step. A contained five-business-day showcase is offered only when the workflow is practical and its boundaries can be agreed.
When should pricing be discussed?
Ask early whether there is an approved budget or decision process, but provide implementation pricing only after the workflow, responsibilities, controls, integrations, and acceptance criteria are sufficiently defined.
Continue the topic
Related practical guides
The practical takeaway
A good first lead conversation makes the problem smaller and the decision clearer. Ask enough to map one workflow, then offer only the next step the evidence supports.