AI agents that propose. Your team decides.
Every agent is scoped to one job, shows its reasoning, and stops at a human review queue — never at your audit record.
Asset Agent
Reads asset data from imports and existing tools, and proposes structured, attribute-complete entries for your inventory.
What it does
- Parses spreadsheets and existing exports into structured assets
- Suggests attributes and categorization based on patterns it recognizes
- Flags low-confidence matches for manual review instead of guessing
What it doesn't do
- Write directly to your live asset inventory
- Delete or overwrite existing asset records
- Make ownership or criticality decisions on its own
Risk Agent
Works through your assets, risks and company profile step by step, then applies your configured formula and thresholds to score risk and propose treatment options.
What it does
- Reads assets, risks and vulnerabilities with read-only tools, one step at a time, and shows that reasoning live — not just a final answer
- Scores risks against the formula and thresholds you've configured, and proposes treatment options (mitigate, transfer, accept, avoid)
- Rejects its own proposal if it references an asset or risk it never actually looked up
What it doesn't do
- Change your risk formula or thresholds itself
- Approve or close a risk without human sign-off
- Override a risk score a reviewer has already set
Gap Analysis Agent
Compares your current controls against a target framework and drafts a prioritized list of what's missing.
What it does
- Compares implemented controls against a chosen framework
- Drafts a prioritized remediation list with rationale
- Points to existing evidence that may already satisfy a gap
What it doesn't do
- Mark a control as compliant on your behalf
- Submit findings to an auditor or regulator
- Assume a framework requirement is met without evidence
Audit Planning Agent
Reads your candidate risks, risk context, open findings and past audits step by step to propose a scoped audit plan.
What it does
- Works through assets, risks, treatment plans and prior audit history using read-only tools, one step at a time
- Shows its actual steps and reasoning live while it works, not just a final answer
- Discards any proposal that references a risk it never actually looked up
What it doesn't do
- Decide the audit scope on its own — every plan is a proposal awaiting review
- Skip steps or guess at data it hasn't read
- Start an audit or notify anyone without a human approving the plan
Audit Interview Agent
Conducts structured audit interviews, collects responses and evidence, and checks them against what's already in your platform.
What it does
- Asks scoped questions based on the audit's framework and control set, and collects written answers and file evidence
- Cross-checks interview answers against existing control, evidence and continuous-monitoring records, and asks about contradictions
- Drafts a summary and finding proposals for the lead auditor once the interview ends
What it doesn't do
- Create a finding without auditor approval
- Share platform data with the interviewee beyond what the interview calls for
- Run without notice — interviewees see a privacy notice first whenever device evidence is in scope
See Govexa on your own data
A working session with your assets, your frameworks, your risk model — not a generic slide deck.