The scenario: diagnosing without a script
A client rolls out new accounting software and it stops pulling data from an existing inventory system that worked fine the week before. The client wants it fixed today and has already floated a theory, that the new software is simply broken, without much evidence behind it.
A technical interview built for this role hands over that same situation and asks the candidate to reason through it out loud: what would you check first, what would rule the client's theory in or out, what's the next question you'd ask before touching anything. There's no code to run and no ticket to close, just the reasoning a consultant would actually walk a client through on the phone.
What this session covers, and what it hands off
Expect diagnostic reasoning questions built around integration failures, vendor mismatches, and a system that behaves differently than documented, each one asking for the thinking process rather than a memorized fix. The AI interviewer follows up the way a real technical interviewer would, pushing on a thin explanation or asking what evidence would change your hypothesis.
What this session does not cover is any actual coding, scripting, or query writing. For a posting that explicitly tests hands-on skills like a basic script or a SQL query against a sample database, that lives in technical challenges, a separate feature supporting coding, debugging, SQL, and written formats with real code execution, and it's worth treating as this session's companion rather than a substitute for it.
Scoring
How the five dimensions apply here
Technical is the dimension doing the most work in this session, and its written reasoning explains whether the diagnostic path made sense, not whether it matched one predetermined answer. Communication matters almost as much, since a consultant who reasons well internally but can't narrate that reasoning clearly loses the part of the job that actually gets billed.
Calibration floors and caps apply the same way here as elsewhere, so a fluent-sounding guess that skips the diagnostic steps won't outscore a slower, more methodical answer that actually gets to the right hypothesis.
Frequently asked questions
Will I be asked to write or run any code in this session?
No. The technical interview type is a spoken conversation about diagnostic reasoning. Writing or running actual code, including SQL, happens in the separate technical challenges feature.
How is this different from a job-description-based session for the same role?
The job-description type reads a specific posting and questions you on its named tools and requirements. The technical type tests general diagnostic reasoning without needing a posting to draw from.
What if the role I'm targeting really does require writing SQL?
That's a real, honest gap this session doesn't cover on its own. Pair it with technical challenges, which supports SQL alongside coding, debugging, and written formats with real code execution.
Does a confident-sounding wrong answer score well here?
No. Calibration floors and caps apply to this session the same as every other Intervieux interview, so an answer that sounds fluent but skips the actual diagnostic reasoning won't outscore a slower, more accurate one.
Related pages
Practice explaining a diagnosis out loud
Run a technical AI interview to rehearse the spoken half of IT consulting, then pair it with technical challenges if the role also tests hands-on coding or SQL.