From their evidence to a defensible score.
No blank questionnaire. No trust-me percentages. Four steps that turn a pile of real evidence into an assessment your team — and your client's board — can stand behind.
- 1Step 1Ingest(feed the beast)
Bring in the client's existing policies, procedures, tickets, exports, screenshots — whatever they already have. Di Svella reads across formats and builds a working memory of what's actually in place, tagged by source.
- 2Step 2Pre-fill
Every control in the framework comes back with a drafted answer, a citation to the exact document and section it came from, and a confidence rating. The blank questionnaire fills itself — and shows its work.
- 3Step 3Review & interview
Accept high-confidence answers in a click. Interviews narrow to the small set of controls the evidence couldn't answer, or where the answer felt thin. Surgical, not exhausting — and every question is on the page for a reason.
- 4Step 4Score & reveal
Two scores, side by side, each 0–100: how built the control is (Implementation) and how well it actually runs (Operational Maturity). Findings, instability flags, and citations sit next to the numbers so the report defends itself.
Why the output holds up.
Assessment and audit aren't separate products — they're a depth dial you set per control. Assess by default; audit the high-risk controls against real evidence, all in one engagement.
A Control Implementation Score and an Operational Maturity Score, side by side, each 0–100. Because '82% compliant' has never stopped a breach.
Every answer carries a visible trail back to its source. When the evidence contradicts what someone attested, the evidence wins — and the score reflects it.
When a key number changes three times across interviews, Di Svella flags the instability itself as a finding — automatically.
Di Svella weighs every finding by who said it. Your report survives the 'well, that's just our PM' objection before it's ever raised.
Two numbers. One honest picture.
Because "82% compliant" has never stopped a breach.