Dextra AssuranceOps

Scan, test and monitor security from one workspace.

Continuous scanning, AI-powered pentesting, AI agent monitoring and evidence packages — connected to compliance, remediation and human approval.

In development. See current implementation and what comes next.

Built forSecurity, compliance, AI governance and engineering teams.

Illustrative AssuranceOps workspacePRODUCT IN DEVELOPMENT
ASSURANCEOPS · CONTINUOUS ASSURANCEScans, agent monitoring and evidence in one workflow
01Systems02Agents03Scans04Evidence05Decisions
Scan & test
SCOPED
Monitor
OBSERVE
Remediate
REVIEW
Package
APPROVE
CONNECTED RESULTKeep findings, fixes, human decisions and evidence connected as systems change.
LifecycleEnd to endAuthorityHuman retainedChangeTargeted reopen

What AssuranceOps brings together.

The complete AssuranceOps product in development: from recurring checks and authorized testing to remediation, evidence packages and ongoing review.

01
Continuous scanning and monitoringKeep checking approved systems, refresh evidence and surface changes that need attention.
  • Scheduled and event-triggered checks within approved targets, sources and execution windows
  • Scanner findings, control failures, stale evidence and collection gaps linked to accountable owners
  • Declared check frequency, last successful observation and follow-up when collection fails
02
AI-powered pentestingRun authorized Strix assessments, investigate findings and verify fixes through retesting.
  • Explicit target authority, permitted methods, execution limits and stop/recovery controls
  • Original attempts, findings and artifacts retained for human validation
  • Remediation, retest and a reviewed report with scope, results and limitations
03
AI agent monitoring and governanceFollow agent configuration, behavior, tools and outputs, with decisions tied to the exact revision.
  • Inventory of systems, agents, owners, models, prompts, data, tools and permissions
  • Infrastructure, task-quality, behavioral and business-outcome monitoring with source coverage and gaps
  • Input, model and output risk signals; intended-use evaluations; incident investigation and reassessment
  • Attributable human approval tied to the reviewed case state; override and renewed review after relevant changes or evidence expiry
04
Evidence packages and audit preparationTurn the work into a scoped package another security reviewer, customer or assessor can inspect.
  • Source evidence references, versions, covered periods and links to controls, findings, exceptions and remediation
  • Exact contents, exclusions, limitations and accountable approval before release
  • Versioned reports and packages with integrity checks, correction history and separate recipient acknowledgement
05
Compliance operations and healthSee what applies, what is supported, what is failing and who needs to act next.
  • Program and system scope, versioned obligations, reviewed framework mappings and coverage gaps
  • Evidence requests, owners, due dates, rejected submissions, reminders and overdue work
  • Current and historical health with visible failed, unknown and stale states
06
Remediation, exceptions and reassessmentCarry each finding through an owned fix, retest, decision and later change.
  • Finding-to-fix history, evidence-qualified closure and human disposition
  • Exceptions with scope, rationale, compensating controls, authority and expiry
  • Affected work reopens when the source, system, permissions, requirements or risk changes
07
Commercial and federal assuranceConnect buyer and supplier reviews with the controls, reporting and remediation each environment requires.
  • Source-linked questionnaire answers, supplier determinations and reviewed customer responses
  • Federal findings, POA&M, assessment and continuous-monitoring documentation with OSCAL exchange
  • Customer-selected applicability and reporting profiles with external assessment and authorization retained by the responsible authority
08
Referenced AI assistance and deployment controlAsk about the work in plain English and follow the supporting records, under the chosen data policy.
  • System-wide answers and proposed document or control improvements grounded in authorized source records
  • Reviewable proposed changes with human approval before consequential action
  • Shared, dedicated and customer-hosted deployment profiles with tenant-specific model, key and retention policies

Current product view · synthetic workflow

Keep evidence, exceptions, remediation, approval, reporting, and reopening in one security assurance case.

The open fictional buyer-review story demonstrates evidence review, named approval, and a simulated return path.

This public view uses synthetic, rights-cleared data. It illustrates a workflow without exposing customer records or implementation-sensitive detail.

Illustrative synthetic product view
What the example shows and how to evaluate it

The open buyer-review scenario uses fictional records and a simulated return path. Selected workflows, including first-time setup, have separately been implemented and tested in a local test environment. Platform scope describes the intended full direction. Live evaluations, coverage of AI agents at runtime, and delivery to downstream systems remain to be verified.

Compare with your current tools: When evidence changes after approval, can your current tools show what was reviewed, who decided, and which answers need review again?

Illustrative synthetic Dextra AssuranceOps buyer-review evidence
Illustrative synthetic product view — not customer data, deployment, or a production result
Watch the product workflowOptional visual walkthrough · illustrative
DEXTRA / ASSURANCEOPSSCROLL TO FOLLOW THE WORK

Illustrative sequence for AssuranceOps. Reviews, findings, evidence, exceptions, owners, approvals, and follow-up live in different portals, files, tickets, and inboxes. Teams repeatedly rebuild what applied, what was supported, and who accepted the remaining risk. We're building AssuranceOps to connect continuous scanning, AI-powered pentesting, AI agent monitoring, compliance operations and evidence packages in one workspace. Find risks, assign fixes, verify remediation and keep the evidence behind each human-approved decision as systems change. AI may classify, compare, summarize, and draft. Named people retain authority for assertions, exceptions, risk acceptance, attestations, and external responses.

  1. 01Where the work breaks

    Critical decisions are being rebuilt from scattered evidence.

    Reviews, findings, evidence, exceptions, owners, approvals, and follow-up live in different portals, files, tickets, and inboxes. Teams repeatedly rebuild what applied, what was supported, and who accepted the remaining risk.

    CEDARLINE HEALTH · SOC 2 SECURITY REVIEW

    Do you encrypt customer data at rest?

    Unresolved
    Current evidence
    None bound
    Approved answer
    None on file

    Three questions imported. No prior evidence or answers to carry forward.

    SYNTHETIC TENANTA question the review cannot answer yet

  2. 02What changes

    Move from scattered evidence to one defensible security decision record.

    We're building AssuranceOps to connect continuous scanning, AI-powered pentesting, AI agent monitoring, compliance operations and evidence packages in one workspace. Find risks, assign fixes, verify remediation and keep the evidence behind each human-approved decision as systems change.

    CEDARLINE HEALTH · SOC 2 SECURITY REVIEW

    Do you encrypt customer data at rest?

    Partially supported
    Bound evidence
    Configuration review — customer data stores
    Bound by
    Rowan Okonkwo (admin)
    Content hash
    afe58af5f6745ae43710

    The posture changed because a named person bound a versioned snapshot to it.

    SYNTHETIC TENANTThe same question, now resting on something

  3. 03Where authority stops

    Connect evidence and action without blurring assessor, certification, buyer, or risk authority.

    AI may classify, compare, summarize, and draft. Named people retain authority for assertions, exceptions, risk acceptance, attestations, and external responses.

    DEC-2026-0001

    Buyer response package — Cedarline Health

    Gated
    PrerequisitesExact payloadPost-approval driftClaim-to-action
    Authority
    Rowan Okonkwo (admin)
    Approval digest
    8a0f9d1e5581ad67bbac057c1c01e0ad664b35a4
    Evidence digest (current)
    595e2d8f9518c1e9efa374b5636cd22d1c815687

    SYNTHETIC TENANTFour gates, a named authority, and the digests it is bound to

  4. 04What you are left with

    A defensible security assurance decision with targeted follow-up

    The next reviewer can follow the evidence, responsible person, unresolved conditions, and reason for reopening without treating an earlier approval as a permanent answer.

    Start an AssuranceOps fit review

    BUYER PACKAGE · VERSION 1

    Versioned, digest-verifiable

    Not approved
    Digest
    a8d4c9c826a1bffc76957e1b
    Answered
    1 of 3 · 2 limitations stated

    The package says it is not approved rather than implying consent. Building it is not delivery: nothing has been sent, received, or accepted.

    SYNTHETIC TENANTA result that states what it is not

Illustrative sequence · not a customer result Human authority remains explicit

Where AssuranceOps helps

THE FRICTION

Critical decisions are being rebuilt from scattered evidence.

Reviews, findings, evidence, exceptions, owners, approvals, and follow-up live in different portals, files, tickets, and inboxes. Teams repeatedly rebuild what applied, what was supported, and who accepted the remaining risk.

THE COMPOUNDING VALUE

History becomes operating context

The next reviewer can follow the evidence, responsible person, unresolved conditions, and reason for reopening without treating an earlier approval as a permanent answer.

Carry the exact review or decision from scope and evidence through approval, action, and reassessment.

  1. 01

    Scope

    Name the service, use case, affected assets, decision owner, deadline, and governing criteria.

  2. 02

    Qualify

    Collect permitted observations and run authorized tests. Tie each answer or finding to current, applicable evidence.

  3. 03

    Resolve

    Validate findings, assign fixes, retest and route the exact exception or missing fact to its owner.

  4. 04

    Approve

    Present the evidence, limitations, and proposed result to the accountable human authority.

  5. 05

    Return

    Prepare the approved response, supplier decision, report or evidence package for its intended reviewer.

  6. 06

    Reopen

    When evidence, access, scope, incidents, or requirements change, reopen only the affected work.

Choose the path by the decision at risk. Keep one governed operating pattern underneath it.

01

Close a buyer security review

A buyer-format response with cited evidence, explicit limitations, owned remediation, accountable approval, and change-aware reuse.
For
Software, AI, and service suppliers with a deal, onboarding, or renewal at risk
What keeps breaking
The hard questions outrun the approved evidence library, while scanner, pentest, remediation, and engineering records remain disconnected from the answer.
02

Decide supplier and AI-vendor risk

One use-specific decision with evidence, conditions, accountable authority, downstream actions, expiry, and reassessment history.
For
Procurement, TPRM, security, AI governance, legal, and accountable business owners
What keeps breaking
The supplier, intended use, data and agent access, evidence, test results, conditions, contract, and renewal decision drift apart across systems.
03

Operate federal vulnerability and authorization evidence

A provider-owned, evidence-linked response and package path that preserves assessor, certification, agency, and authorizing-official authority.
For
Cloud-provider and program security, engineering, compliance, control-owner, and authorization-support teams
What keeps breaking
Findings, provider evaluation, remediation, retest, authorization artifacts, reporting, and the governing record diverge across operational and compliance systems.
04

Map a consequential AI-assisted decision

One reconstructable decision record connecting the source, configuration, evaluation, named human authority, action, result, and event that requires reassessment.
For
Government programs, regulated operators, AI governance, security, mission, and accountable business owners
What keeps breaking
Source evidence, model and prompt state, tool permissions, required checks, reviewer judgment, and downstream action drift across systems while a human-in-the-loop label says little about what was actually reviewed.

Questions about AssuranceOps in your workflow

What does AssuranceOps do, and what is still being built?

We're building a connected product for continuous scanning and monitoring, authorized AI-powered pentesting, AI agent assessment and governance, compliance operations, remediation and reviewable evidence packages. It will carry work from scoped checks through findings, fixes, retesting, human decisions and ongoing reassessment. The build-status section below separates current local implementation from remaining work; the public example illustrates one buyer-review workflow.

When can we reuse a security questionnaire answer?

You can reuse a security questionnaire answer when its meaning, scope and supporting evidence still fit the new question. Similar wording alone is not enough. Check the service, deployment, exceptions, source version and approval before returning the answer.

Does AssuranceOps replace our evidence store or buyer portal?

Your existing systems remain authoritative for their records. AssuranceOps is designed to connect the question, evidence, owner and human decision around one review. The public story uses a simulated return path; a connection to your actual portal or evidence store needs separate implementation and verification.

What should we test before choosing security questionnaire software?

Use the same synthetic questions and approved sample documents in each tool. Include a supported answer, missing evidence, a service-specific exception and conflicting sources. Check review, changed evidence and the exact returned wording as well as the first draft. These are evaluation checks, not measured AssuranceOps outcomes.

Use the security questionnaire checklist · Explore security review services

Explore the build behind the product.

AssuranceOps is in development. The complete platform scope above is the product we're building; selected workflows can already be inspected in local implementation or the public example.

Current implementation and remaining work
  1. 01 / BRING

    Requirements, evidence and findings

    Name the system or service, the owner, the applicable criteria and the period being reviewed.

  2. 02 / WORK

    Find gaps and move work to owners

    Connect controls, tests and evidence; track remediation and present the exact decision for human review.

  3. 03 / KEEP

    A record another reviewer can follow

    Preserve sources, versions, unresolved conditions, named decisions and the changes that require renewed review.

Current local implementation and remaining work

These workflows are locally implemented; the complete platform is not generally available as production software. The local application is separate from the fictional public example. A customer deployment needs its own approved hosting, identity, source connections and acceptance checks.

Production integrations, customer-specific compliance, operational authorization, and accepted results require implementation and evidence in the selected environment.

Controls and evidence

IMPLEMENTED LOCALLY

Track control records, evidence versions, owners and validity windows. Retain the evidence behind a recorded test and flag missing or stale support.

NEXT / STILL TO VERIFY

Complete scope-specific obligation and framework mapping beyond the current representative catalogs, show uncovered requirements, and extend source-linked control advice with proposed policy and document changes for human review.

Monitoring and control health

IMPLEMENTED LOCALLY

Calculate control health, evidence gaps and overdue work from recorded controls and evidence. Record scheduled reassessment attempts. A Tenable read adapter imports finding windows; it does not run scans.

NEXT / STILL TO VERIFY

Validate ongoing operation with approved live sources, source-specific freshness and complete historical health views. Comprehensive hosted monitoring is not verified.

Buyer and supplier reviews

IMPLEMENTED LOCALLY

Link questionnaire drafts to recorded facts, review supplier evidence and retain a versioned determination. Keep changed evidence and due reviews visible.

NEXT / STILL TO VERIFY

Validate customer portal connections and the full document-assisted advisory path. Preparing a response does not establish delivery or buyer acceptance.

Vulnerabilities and security testing

IMPLEMENTED LOCALLY

Record authorized testing scope, findings, remediation and retest history. Governed execution uses a synthetic provider; live Strix execution is not enabled.

NEXT / STILL TO VERIFY

Complete and verify isolation and execution lifecycle controls before authorized live Strix testing, then evaluate results independently.

AI governance and assisted work

IMPLEMENTED LOCALLY

Record agent revisions, declared tools and permissions, evidence and named approvals. Preserve the reviewed revision after change. Case-bound assistance requires approval before applying a proposed change.

NEXT / STILL TO VERIFY

Complete AI inventory, risk and obligation mapping, intended-use quality, fairness, privacy and robustness evaluations, live runtime observation, and incident response with reassessment. System-wide plain-English chat with supporting record links is planned; case assistance does not prove full coverage.

Federal remediation and reporting

IMPLEMENTED LOCALLY

Track evidence-qualified POA&M closure and reopening, federal reporting records and bounded OSCAL exchange with diagnostics.

NEXT / STILL TO VERIFY

Validate the customer's framework, system boundary, reporting profile and receiving tool. Bounded schema checks do not establish full program conformance or agency authorization.

One workflow, within the whole product. The open buyer-review example lets you inspect synthetic evidence, approval and change handling. The scoped buyer-review service is a starting point. Neither represents the complete AssuranceOps platform.

People retain risk acceptance, external release and assessment authority. Recorded permissions are not proof of runtime enforcement; an export is not recipient acceptance.

Connect evidence and action without blurring assessor, certification, buyer, or risk authority.

Integration design, safeguards and compliance scope
AUTHORITATIVE INPUTS

Buyer portals, evidence stores, scanners, trust centers, contracts, identity, CRM, GRC, and ticketing remain authoritative for the records they own.

ASSURANCEOPS

AssuranceOps keeps the scoped question, applicable evidence, exception, owner, decision, returned result, and later change in one security assurance case.

CONTROLLED OUTPUTS

Responses, decisions, reports, and tasks can be prepared for selected destinations. The public story simulates the return path; actual transport, delivery, and recipient acceptance require separate evidence.

Connect only the customer-approved read or write path needed for the workflow, with explicit identity, permission, receipt, failure, and recovery behavior.

Evidence before assertion

Unsupported, stale, contradictory, or out-of-scope evidence remains visible instead of becoming a polished claim.

Exact-state authority

A material change to the payload, evidence, scope, or governing criteria clears dependent approval.

Scoped applicability

Commercial, enterprise, and federal criteria remain separate unless the case proves that a requirement actually applies.

AssuranceOps can map work to customer-selected security, supplier-risk, vulnerability, and compliance criteria. It does not replace an independent assessment, certification, accreditation, authorization, legal judgment, or buyer decision when those are required.

Authority boundary: AI may classify, compare, summarize, and draft. Named people retain authority for assertions, exceptions, risk acceptance, attestations, and external responses.

Start an AssuranceOps fit review

Describe the systems or AI agents you need to assess, your scanning or monitoring needs, current tools, decision owner and required report or evidence package. THONIS will reply with a fit assessment and next step. A call is optional. Do not send sensitive evidence through first contact.

Start an AssuranceOps fit review