Module 15 · Lesson 02

Client Intake — Asking the Right Questions First

Reading time: 16 minutes Track: Yungsten Tech Employee Curriculum · Required for all staff

The most expensive mistake in consulting

The most expensive mistake in AI consulting is starting to build before you understand the real problem. The client says "we need an AI agent for customer support." You build it. Three months in, they reveal that their customer support team refused to use it because nobody asked them, their data is in a system the agent can't access, and what they actually needed was something much simpler that addressed a specific bottleneck.

Good intake prevents this. It takes 90 minutes. It's worth weeks.

The four intake dimensions

Dimension 1: The actual current state

Not what the client says they have — what you can verify.

• Show us the prototype/MVP. What does it actually do today?
• Who uses it now, how often, and for what?
• What's the biggest pain point in its current form?
• What have you already tried that didn't work?

The gap between described state and actual state is where most project scope errors live. Clients describe what they wish they had; you need to see what they do have.

Dimension 2: The team that will operate it

• Who is the day-to-day operator of this system after handoff?
• What's their technical comfort level? (Be honest — this affects every architectural decision)
• Who makes decisions when something breaks?
• Is there a dedicated person or is this a side responsibility?

Dimension 3: The data situation

• What data does this system need to work?
• Where does that data live today?
• What's their data classification policy? What can go in an AI tool?
• Any regulatory requirements? (HIPAA, SOC2, GDPR, industry-specific)

This conversation often reveals the biggest hidden constraints. Many clients haven't thought carefully about data governance for AI systems. Surface it early, not after you've built something that can't use their actual data.

Dimension 4: Success definition

• What does success look like at 30 days? 90 days? 12 months?
• How will you know it's working?
• What's the minimum viable success — the thing that would make this worthwhile even if nothing else goes right?
• Who needs to sign off that it's working?

The rungs: matching service to need

After intake, you can place the client on the right rung:

Starter rung: They have a vague AI interest and a specific workflow pain. They need to experience what's possible before committing. Deliver: one named agent + a starter wiki. Let them determine if the rhythm is right.

Full engagement: They have a clear problem, a team with some capacity to engage, and organizational buy-in. Deliver: biweekly C-suite sessions, wiki development, four agents over the first quarter, monthly team seminar.

Custom: Enterprise scale, regulated industry, or complex existing infrastructure. Scoped specifically to their situation.

The right rung matters more than the right solution. A client on the wrong rung — either too much or too little service — will have a bad experience regardless of output quality.

Documenting intake

Every intake conversation gets a written summary within 24 hours covering the four dimensions plus the recommended rung and first-90-days plan. This becomes the reference document for the engagement. Any scope changes get validated against it.