Your web apps change. Sentinel AI watches.
Sentinel AI continuously exercises, analyses and validates your App Recipes — helping detect workflow changes, security concerns and failures before they affect your users.
Continuous AI assurance for every web app.
Continuous application assurance for every App Recipe.
FirmBrowser doesn't simply automate your web apps
It continually checks that the automation, security and application behaviour can still be trusted. Each App Recipe carries a monitoring schedule, and on that schedule Sentinel AI agents execute the Recipe and inspect its operation.
Each App Recipe carries its own monitoring schedule. Business-critical applications can be exercised every hour; lower-risk Recipes can be validated daily or weekly. The schedule is set centrally, alongside the rest of the Recipe's security policy.
A responding web page is not the same as a working, governed application
Sentinel AI validates the whole FirmBrowser workflow — authentication, transformation and security policy — not simply whether a server answered.
“Is the website online?”
- A page responded
- A certificate is valid
- A status endpoint returned 200
“Can our people still securely use this application exactly the way our FirmBrowser policy expects?”
- The login page still loads
- The expected username field still exists
- The password workflow still operates
- Push MFA approval is still enforced
- Authentication completes successfully
- The correct post-login page is reached
- Expected transformations are still applied
- Restricted elements remain inaccessible
- URL Rules still operate
- Table Rules still identify the intended data
- Credential Lockdown remains effective
- The final state matches the Recipe outcome
An AI agent runs the Recipe exactly as a user would
At the scheduled time, the agent executes the App Recipe in a controlled test environment. It follows every stage, determines whether each expected outcome occurred, records the result, analyses abnormalities and decides whether the Recipe remains healthy.
- Step 01Scheduled AI run
The Recipe's monitoring schedule triggers an agent run.
- Step 02Launch App Recipe
The agent executes the Recipe in a controlled test environment, exactly as a user would.
- Step 03Observe every step
Each stage of the login, transformation and policy sequence is watched and recorded.
- Step 04Validate expected behaviour
Every expected outcome is compared against the Recipe's established baseline.
- Step 05Analyse changes & security signals
Differences are interpreted — what changed, and whether it actually matters.
- Step 06Pass, warning or failure
The run resolves to a Recipe Health status the administrator can act on.
- Step 07Administrator alert + AI analysis
A written explanation of what happened, not a raw stack of logs.
Login, transformation and security policy — all continuously verified
Sentinel AI validates the different kinds of control a FirmBrowser App Recipe carries.
Login Assurance
- The application can still be authenticated successfully
- Credentials are still handled correctly and never exposed
- Push MFA is still enforced where policy requires it
- The user arrives at the expected application state
App Transformation Assurance
- Element, URL and Table Rules are still being applied
- Restricted elements remain hidden or inaccessible
- Tables continue to identify the intended information
- Navigation and components behave as the Recipe expects
Security Policy Assurance
- URL restrictions continue to operate
- Credential Lockdown controls remain effective
- No unexpected authentication paths have appeared
- Application behaviour has not undermined firm policy
AI that understands whether a change actually matters
Vendors change HTML, DOM structures, scripts, authentication workflows, URLs, layouts and session handling constantly. Some changes are harmless, some break automation, and some quietly weaken a security control. Sentinel AI explains the difference in plain language.
[03:21:04] recipe: legacy-finance-portal [03:21:07] step 7 FAILED [03:21:07] element not found: #signin [03:21:07] run aborted
A technician still has to work out what changed, whether it matters, and what to do about it.
Login workflow changed overnight
Sentinel AI detected that the application's login workflow changed at approximately 3:20 AM. The previous Sign In button is no longer present. A new Continue button now appears before password authentication. The current Recipe requires review.
Selector changed, replacement identified
The login button selector changed, but Sentinel AI successfully identified the replacement element and completed the Recipe within the expected baseline.
New third-party authentication domain
The Recipe completed successfully, however the application now contacts a new third-party authentication domain during login. Administrator review is recommended.
Transformation could not be confirmed
An expected App Transformation could not be confirmed on the post-login page, and a previously hidden application control is now visible to the test role.
Every App Recipe carries a health status administrators can act on
A single view across the application estate — what is healthy, what has changed, what needs attention and what has failed.
- Every 6 hours HealthyMYOB
All expected steps completed within baseline.
- Hourly HealthySalesforce
Login, MFA approval and transformations verified.
- Hourly Change detectedLegacy Finance Portal
Login workflow changed — additional Continue step observed.
- Daily WarningInEight
A restricted element was visible for part of the session.
- Every 6 hours FailedSupplier Portal
Recipe could not reach the expected post-login page.
All expected steps completed and behaviour remains within the established baseline.
The Recipe completed, but a material change was observed and should be reviewed.
Part of the workflow produced an unexpected result or a security concern.
The Recipe could not successfully reach its expected outcome.
Know before your users do
Sentinel AI proactively exercises your critical App Recipes and identifies changes before they become tomorrow morning's support tickets.
Reactive
- 1The application changes overnight
- 2Users hit a failure at 8:45 AM
- 3Support tickets arrive
- 4A technician starts investigating
- 5The firm loses a morning
With Sentinel AI
- 1The application changes overnight
- 2A scheduled Sentinel AI run exercises the Recipe
- 3The change is detected and analysed
- 4An administrator is alerted with an explanation
- 5The Recipe is corrected before users arrive
Your applications change. Your security shouldn't silently break when they do.
Controls are only controls while they are still working
FirmBrowser already governs how users interact with web applications. Sentinel AI adds continuous verification that those controls remain effective — credential protection, Push MFA enforcement, Credential Lockdown, App Transformation, URL restrictions and unexpected authentication paths.
The intelligent security and automation layer between your people and their web applications
FirmBrowser does not merely store credentials, automate a login, transform an application or enforce a control. It continually observes, tests, analyses and verifies that those systems still operate as intended.
Continuously prove that the applications your firm depends on still work — and still work securely.
See Sentinel AI, Recipe Health and application change intelligence in a live FirmBrowser walkthrough.