AIMentor — AI Expertise Learning Platform

13 questions · pass mark 80% · 20 minute limit. The timer auto-submits at zero.

Time remaining: 20:00

Question 1easy
What is the dominant cause of failed AI projects?
Slow model inferenceBuilding the wrong thing — solving a problem the client didn't really haveUsing the wrong programming languageInsufficient GPU memory
Question 2easy
'We want a chatbot / an AI agent / to use GPT' is an example of what?
A well-defined problem statementA solution looking for a problem — you must uncover the real business problem underneathA feasibility assessmentA success metric
Question 3medium
In a discovery conversation, when should you first discuss AI?
Immediately, to set expectationsOnly after the business problem, its cost, and its owner are clear — then ask if AI is the right toolNeverBefore understanding the problem, to sound expert
Question 4medium
Which questions belong in a discovery session? (Select all that apply)
What's the painful/repetitive/slow thing, and what does it cost?Who does it, how often, and what happens when it goes wrong?If solved, what changes and how would you measure success?Which LLM should we fine-tune?
Select all that apply.
Question 5medium
A client asks for 'a chatbot on our website.' What's the reframing move?
Immediately quote a chatbot buildUncover the real problem (support overwhelmed? users can't find answers?) and how to measure successTell them chatbots don't workFine-tune a model first
Question 6hard
Which feasibility axes should you check before scoping an AI build? (Select all that apply)
Is it even an AI problem, or would rules/a query/a workflow solve it better?Is the data/context there (knowledge source, examples, reliable tools)?Is the error tolerance realistic for a probabilistic system?Is the model the newest available?
Select all that apply.
Question 7hard
Why is being willing to say 'AI isn't the right tool here' valuable for a consultant?
It avoids all workThe honest feasibility call builds trust and differentiates you from vendors who just sell the buildIt is legally requiredIt raises your rate automatically
Question 8hard
After confirming a real, feasible, valuable problem, how should you scope the first project?
Boil the ocean — automate everything at onceA narrow, complete, de-risked vertical slice with measurable success criteriaShip nothing until it's perfectBuild the easiest part regardless of value
Question 9medium
In scoping, why build the part most likely to fail first?
To make the demo look harderTo de-risk the unknowns early and learn cheaply before investing in the restBecause it's the most funTo increase token usage
Question 10hard
How can discovery be a business model, not just free pre-sales?
By giving it away to win the buildAs a paid operational diagnostic that maps processes, prioritizes opportunities, and hands over a roadmap — valuable on its own and converting to the buildBy skipping it entirelyBy charging only after the build ships
Question 11medium
Which are discovery/scoping pitfalls? (Select all that apply)
Accepting the client's solution as the brief without reframingSkipping feasibility (data, error tolerance, is-it-AI)Boiling the ocean with a broad first scopeDefining measurable success criteria up front
Select all that apply.
Question 12medium
A request phrased as 'we want a chatbot' is a ____ statement, not a problem statement (one word).
Question 13hard
What paid engagement (two words) maps a client's processes and hands them a prioritized AI roadmap, converting into the build?

0 / 13 answered

Discovery & Scoping — Final Exam