Meet Edward.He checked our own platform first.
You hold applicants' Social Security numbers on a platform that's probably 20 years old. We pointed frontier AI at ours first. Now we work with your team to analyze, remediate, and defend yours, in weeks, not months.
AI just found flaws
that hid for decades.
Attackers get the same AI.
Go first.
Examples from Anthropic's public reports, 2026.
Analyze. Remediate.
Defend.
Analyze.
With written permission, read-only first, we use frontier AI to review your code, cloud, access, and data handling.
Prioritize.
We reproduce each finding and rank it by real risk, so your team fixes what matters first.
Remediate.
AI drafts each fix with an undo step. We work alongside your team, and your people approve every change.
Prove and defend.
We replay the attack to prove it's closed, then set up alerts and scheduled checks your team runs after we leave.
Exposure reportSample Co.Scorecard DAY 4
Reproduced and ranked TOP 5
Fix 1 of 6 · Require MFA on the admin portal DRAFT
Verified DAY 21
No fix ships on
one AI's word.
A first week,
in plain English.
- Reviewed 214 access keys · 3 unused for over a yearflagged
- Checked sign-in rules on 6 admin portals · 1 without MFAcritical
- Reproduced a public file listing on a storage bucketconfirmed
- Wrote fix: require MFA on the admin portalready
- Second AI replayed the attack after the fixblocked
- Approved by IT directorshipped
- Compared vendor access to contracts · 2 over-permissionedflagged
- Confirmed backups can't be deleted by automationpassed
- Reviewed email domain authenticationfix drafted
- Patch level on internet-facing servers · 1 behindflagged
- Reviewed 214 access keys · 3 unused for over a yearflagged
- Checked sign-in rules on 6 admin portals · 1 without MFAcritical
- Reproduced a public file listing on a storage bucketconfirmed
- Wrote fix: require MFA on the admin portalready
- Second AI replayed the attack after the fixblocked
- Approved by IT directorshipped
- Compared vendor access to contracts · 2 over-permissionedflagged
- Confirmed backups can't be deleted by automationpassed
- Reviewed email domain authenticationfix drafted
- Patch level on internet-facing servers · 1 behindflagged
Start with a conversation.
Leave with a team that can defend.
Exposure reviewFree · 30 minutes
Your platforms, your biggest gap, and what an assessment would cover. You leave with a short written summary.
Start hereExposure assessment1–2 weeks · fixed scope
We use frontier AI on your code, cloud, access, and data handling. You get every finding ranked by risk, a fix plan, and the start of the evidence file your clients and counsel will ask for.
Most common next stepRemediation sprint2–4 weeks
We fix with your team. AI drafts each change, your people approve it, and we retest to prove it's closed.
After the assessmentDefense setupIn the sprint, or on its own
We set up "keep watching" in your environment: alerts on unusual exports, access reviews, scheduled AI checks, a runbook, and training. Your team runs it.
Your team keeps itAdvisoryOptional · monthly
A monthly review call, a briefing when a major vulnerability lands, and a retest each quarter. We advise. You decide.
OptionalBreach readinessA working session
Practice the night of a bulk export: who calls counsel, clients, and regulators, and when. You leave with a playbook your team has rehearsed.
Add-onYour systems. Your team.
Your call.
Edward started at deverus. We pointed frontier AI at our own 20-year-old screening platform first, found what years of conventional testing had missed, and fixed it, with proof. Now we show your team how.
Any platform: ours, another vendor's, homegrown, or several after acquisitions.
Is Edward a product we buy?
No. Edward is our mascot and the name for how our consultants use AI. You hire our team. We work with yours, in your environment, and your team keeps everything when we're done.
Who does the work?
Our consultants, led by Shawn Rucks, who founded deverus and has spent 28 years in background screening, with frontier AI doing the heavy lifting. That's how we finish in weeks, not months. People decide what's in scope, and your team approves every change.
What AI do you use?
The most powerful cybersecurity AI model available to businesses, used through enterprise AI services that don't train on business data by default. You decide what it can see.
Does our data leave our environment?
We work inside your environment and accounts. We don't take a copy of your applicants' data, and passwords and keys never appear in our reports.
Could this break something?
We start read-only. Nothing changes in your systems without your team's approval, and every fix ships with an undo step.
Can you guarantee we won't be breached?
No one honestly can. What we can do is find and close the gaps an attacker would use first, prove the fixes, and leave your team able to keep watching.
Does this help with client security questionnaires and audits?
Yes. Your evidence file shows what was found, fixed, and verified, which is what clients and auditors ask for. We support your compliance work; we don't certify it.
We're not on deverus. Can you still help?
Yes. We work on any screening platform: another vendor's, homegrown, or several after acquisitions. Your vendor relationship stays yours, and what we find stays between us and your team.
Why a screening company, and not a big security firm?
We know how screening platforms are built, where applicant data flows, and what your clients' security questionnaires ask, because we've run one for 28 years. And we went first, on our own platform.