Product walkthroughs, demos and short explainers from AnnexSys on digital and AI trust, security and governance.
Length 1:06
Upload a customer's security questionnaire and AnnexIQ drafts the answers from your own trust content and evidence, with sources shown. In formal review, a person approves every answer before anything goes out.
Watch
Length 1:03
Signed-in buyers download public documents freely. Confidential ones wait for an owner's approval, and the PDF a buyer downloads carries their name, email and download time.
Watch
Length 1:06
Buyers see a live compliance ring showing the pass rate of your own control tests and when they last ran. It drops when a test fails and recovers when the control is fixed and retested.
Watch
Length 1:07
Review clocks show which trust content is on track, due soon or overdue. An AI agent drafts the refresh and a named person approves it before anything is published.
Watch
Length 1:08
See which trust content buyers open and which visits turn into access requests, from the last thirty days of visitor activity. Counted, not sampled. It shows interest, not closed deals.
Watch
Length 1:07
Choose which AI systems to disclose. Signed-in customers see a curated summary, while risk assessments, incidents and vendor details stay internal.
Watch
Length 1:07
A buyer's after-hours question is answered from your published documents, with the sources listed. Only documents that buyer is allowed to see are used.
Watch
Length 1:07
Expired and expiring evidence is flagged early, and an AI assistant drafts reminders for a person to approve. Outside auditors get one scoped, time-boxed workspace.
Watch
Length 1:08
Map one control to every framework requirement it satisfies, and size a new framework before you start. Test once, and every mapped standard can rely on the result.
Watch
Length 1:08
Rules pair a control with a plain check on data from your connected systems. A control that drifts shows up at the next run, not at the next audit.
Watch
Length 1:04
Your team logs a regulatory change and sees the requirements, controls and AI systems it touches. One action schedules re-tests, and a review record is kept for the regulator.
Watch
Length 1:05
A failed or partial control test cannot be saved without a finding. Each finding gets an owner, overdue work shows on the dashboard, and a reviewer closes it with history kept.
Watch
Length 1:07
Start a review and the list builds itself from your identity system, with risk flags to focus the reviewer. Each approve or revoke is recorded, and the completed review is filed as audit evidence if linked to a control.
Watch
Length 1:05
See what compliance costs, built from recorded work, and what a failure could cost as a dollar range. Both are flagged as system-calculated until a person approves them.
Watch
Length 1:09
Find unregistered AI across your connected sources, with each item shown next to its source. Then give each item a decision: dismiss it, or register it into the governance lifecycle.
Watch
Length 1:07
AI ideas follow one path to production. Every request is logged, four checks must be cleared before approval, and there is a record of who approved what.
Watch
Length 1:02
See assessed AI systems on one heatmap built from your own risk matrix. Open any rating to see how it was reached, and track every mitigating action to closure.
Watch
Length 1:09
See each AI system's standing against the EU AI Act, ISO 42001 and NIST AI RMF. AI-drafted tests are approved by a person, and evidence is kept with expiry dates.
Watch
Length 1:09
Give each vendor trust and sovereignty scores computed from declared data. Sovereignty gaps are flagged as advisories, and each vendor can be re-reviewed on a cycle.
Watch
Length 1:09
Map what each agent can reach from recorded links, and see what depends on it. Set a limit that suspends a system in the register when reviewed red-team findings reach it.
Watch
Length 1:01
AI incidents get their own types, owners and a defined path to closure. A crossed monitoring threshold raises a candidate incident, and a full history is kept for leaders and auditors.
Watch
Length 1:08
Break AI spend down by AI system, business unit and user, and open any total into the calls behind it. Unattributed spend stays on a visible line.
Watch
Length 1:06
Set budgets at every level, get notified as thresholds are crossed, and have spend spikes and agent loops flagged by system. It alerts and reports, and never blocks a call.
Watch
Length 1:07
Monthly chargeback statements per business unit are drafted from metered usage, reviewed by finance and frozen once issued. Disputes are settled as adjustments on the next period.
Watch
Length 1:06
Record the real provider invoice and compare it with your metered usage for the month: matched, or the gap shown and never quietly corrected. Rate variance flags models that cost more than planned.
Watch
Length 1:07
See where a cheaper model could cut the bill, with the saving computed on your own volumes. A model is recommended only where a passing evaluation is on file. Advice only.
Watch
Length 1:09
Show metered AI spend beside the results you record, with cost per outcome, a return multiple and a headcount equivalent. Every assumption is stated and versioned.
Watch
Length 1:08
See AI spend grouped by risk tier, including systems not yet assessed or overdue for review. One risk-weighted score and vendor trust and sovereignty scores sit beside cost.
Watch
Length 1:04
Upload a model file and AnnexIQ reads its structure without running it. A clean scan is logged, a risky file is flagged with the exact reason, and a person decides.
Watch
Length 1:06
Run adaptive, multi-turn attack tests on a registered AI system from your release pipeline, through your own AI provider. Full evidence is kept and a person reviews the findings.
Watch
Length 1:05
Your application sends AI output, such as SQL or commands, to AnnexIQ and gets a verdict before it acts: allow, strip, rewrite or block. Fixed rules decide the clear cases, and an AI second opinion handles unclear ones.
Watch
Length 1:06
Define what counts as sensitive, and a match in an AI chat tool becomes a finding with the tool, category and severity for a reviewer to confirm or dismiss. It reports and never blocks.
Watch
Length 1:05
Repeat alerts of the same kind on one AI system are grouped into one ranked cluster with a plain reason. A serious cluster can raise one incident for a person to review.
Watch
Length 1:05
Compare each key, agent or user with its own history and flag a sharp rise in activity or a new tool. Agent sessions can be reviewed for drift. Nothing is revoked automatically.
Watch
Length 1:06
Record each tool's package and version and check them in one click against a curated list of known-vulnerable packages. Then see which agents and systems connect to a flagged tool.
Watch