Discovery

The fraction of a fixed roster of eight high-signal discovery questions that have an answer. It is a progress metric and never a gate – nothing blocks on it, and an assessment can be completed with it well short of full. It deliberately ignores the long tail of optional detail fields so it tracks the questions that matter.

Only where AI autofill has been switched on, which is a second flag beneath the module flag and defaults to off. With it off – which is how it ships – nothing here can trigger a paid call. With it on, the rule is narrow and worth knowing exactly: it drafts into empty fields only, and a field somebody answered is never overwritten. What it can pre-fill includes the issue fields that become What We Heard rows, so if your objection is to a machine drafting anything in the customer voice, that is the flag to leave alone.

Discovery does not. It is twenty-seven questions somebody answers, with nothing to install and no access to ask for, which is why it works in a first meeting. Scanning is a separate add-on, the Network Scanner: endpoint agents, a perimeter check, an email-security check and a cloud posture read, each run only after a signed authorisation for that company.

Because the deck built on top of it speaks in the customer’s voice. A default written into an empty answer becomes a sentence on a slide attributed to them, and there is no way for the reader to tell it apart from something they actually said. The deck grader exists partly for this: it checks every What We Heard row against what the deal record actually holds.

It locks with the rest of the content. Reopening the assessment is the only way to change it, which means a completed audit is a fixed account of what was asked and answered rather than something that drifts afterwards.