---
title: AI Workflows
seoTitle: Automated Call Review and AI Compliance Monitoring
description: Write your policies in plain language and Argus Archive checks every call against them, after the call or live, and alerts the right people in Microsoft Teams.
subnav: true
finished: true
---

> [!herorow]
> AI Workflows
>
> # *Automated call review* against your own policies
>
> Describe your business and what to check in plain language. Argus Archive reviews every call against it, after the call or live with the Real-time add-on, and sends each alert to the right person in Microsoft Teams.
>
> [Get a demo](/request-a-demo/) [See pricing](/pricing/)
>
> ![The Argus Archive visual workflow editor for detection rules, alerts and follow-up actions](/media/uploads/2026/09/argusarchive-workflow-editor.jpg)

> [!tourleft]
> Steps
>
> ## From your policy to the right person
>
> Four steps in one editor, with no code at any point.
>
> ![The business context and review instructions of an Argus Archive policy](/media/uploads/2026/10/17-policy-context.jpg)
>
> ### Describe your business
> Add your products, your terms and the standards your reviewers apply. Every rule in the policy uses it.
>
> - Terms read the way your firm uses them
> - Written once for the whole policy
>
> ![An Argus Archive policy as a flowchart: each plain-language question, its response type, condition and alert](/media/uploads/2026/10/18-policy-rules.jpg)
>
> ### Ask your questions
> Each rule is one question about the call, such as "Did the adviser read the risk warning?" The answer can be yes or no, a number, a list, or the exact quotes found.
>
> - Write it yourself, or describe it in one sentence and AI writes it
> - Check your own people, the customer or both
>
> ![Setting an alert condition and its actions in Argus Archive](/media/uploads/2026/10/19-alert-actions.jpg)
>
> ### Choose what happens
> Pick when an alert is raised from simple menus, how serious it is, and what follows: a tag, a message to a Teams channel or an event to your own systems.
>
> - Conditions picked, never typed as code
> - Follow-up that runs on its own
>
> ![Test results for an Argus Archive policy before it is published](/media/uploads/2026/10/20-policy-test.jpg)
>
> ### Test, then publish
> Try the policy on past calls or on test conversations the AI drafts for you, check each answer, then publish it.
>
> - Problems found before they reach a reviewer
> - Every change saved as a version

> [!image]
> Review
>
> ## Review that stands up to scrutiny
>
> Each alert has an owner, and every decision on it is recorded with a reason, so you can show how it was handled.
>
> - **The exact words** from the call that raised it
> - **Assigned** to a reviewer, with a status and notes
> - **Escalated** with a reason, such as possible regulatory exposure
> - **Closed** with a reason, such as a false positive or a confirmed breach
> - **Shared** with a Teams user or channel
> - **Every step** kept in the audit log
>
> ![An Argus Archive rule: a plain-language question, who it applies to, and its alert and severity](/media/uploads/2026/10/12-alert-policy.jpg)

> [!highlights]
> Uses
>
> ## What teams check with it
>
> The same editor serves compliance and customer service. Common first policies:
>
> ### Risk warnings
>
> Required warnings and disclosures given before the customer agrees.
>
> ### Conduct risk
>
> Pressure selling, promised returns or risks played down.
>
> ### Off-channel moves
>
> Asking to carry on by mobile or a messaging app.
>
> ### Supervisor requests
>
> Customers who ask for a manager or say they will complain.
>
> ### Service standards
>
> Greeting, identity checks and next steps agreed on each call.
>
> ### Calls for a case
>
> Matching calls filed into an open case by their alerts.

> [!faq]
> FAQ
>
> ## Frequently asked questions
>
> Still have questions? Ask the team that built it.
>
> [Talk to us](/contact/)
>
> ### How do I write a rule?
> Describe your business and your review standards once, then add each rule as a question in plain language. Define Rule with AI writes a rule from a short description. Alert conditions are chosen from menus, so there is no code.
>
> ### How do I know a policy works before it goes live?
> Test it on a sample transcript, on past calls or on test conversations the AI drafts, and check each answer. Publish only when you are happy. Every change is saved as a version with a note, and AI Assurance keeps the testing on record.
>
> [AI Assurance](/product/ai-assurance/)
>
> ### Does it check calls live or after they end?
> You choose for each policy. After the call is included; live checks need the Real-time add-on.
>
> ### What can happen when a rule matches?
> Argus Archive raises an alert at the severity you set. It can also tag the call, post the alert to a Microsoft Teams channel and file the call into an open case. Sending an event to your own systems through the API is coming soon.
>
> ### How can I see why an alert was raised?
> The alert shows the rule, the AI's answer and the exact words from the call, with the recording and transcript one click away.
>
> ### Who can change a policy?
> Only people whose role allows it. Publishing makes a new version live, and earlier versions stay on record.

> [!articles]
> Articles
>
> ## Worth reading
>
> How AI review works, how it is tested, and what it changes for compliance.
>
> - /blog/ai-assurance-explained/
> - /blog/real-time-compliance-monitoring/
> - /blog/extending-microsoft-365-copilot-with-governed-communications-intelligence/