A rep can handle a roleplay call flawlessly and still take four extra minutes on the CRM update afterward, log the wrong disposition, or skip a required field entirely. Until now, nothing in a training platform could see that part of the job at all.
Why this matters
Roleplay training has always stopped at "goodbye." But for support, sales, and account management teams, a large share of handling time and error happens after the conversation ends: updating records, selecting a disposition, entering notes into the right fields in the right order. That process is often as scripted as the call itself, tied directly to average handling time (AHT) and resolution quality, yet it's rarely trained on with any precision. Reps learn it by shadowing someone, or by making mistakes in production.
How it works
With workflows, launched in v3.4 on April 14, 2026, an admin or manager records the ideal version of a business process directly in the real system, Salesforce, a support console, or any browser-based tool, using the Outdoo Workflow Chrome extension. That recording captures both a video of the session and the underlying events: clicks, field entries, and the order they happened in.
The admin attaches this recording to a roleplay agent. When a rep practices that roleplay, they hold the conversation and complete the same process in the same target system at the same time, exactly as they would after a real call. That only works if the rep has the same Chrome extension installed in their own browser, since it's what captures their clicks and field entries for comparison.
Outdoo then compares the rep's run against the admin's ideal recording and scores it on average handling time, whether the process reached a valid resolution, and which steps were skipped or done out of order. Step-level accuracy shows exactly which parts of the process trip reps up most often, and repeat attempts show whether that improves over time.
Before a recording can be published for reps to practice against, it has to clear a few checks: site access granted, screen permission granted, a video longer than 30 seconds, and a file size over 5MB, so an incomplete take can't accidentally become the standard everyone is measured against.
Full setup steps are in Create a workflow.
What this unlocks
A support team can now train new hires on the exact disposition logic used in their help desk, not a generic version of it, and catch which specific field gets skipped most often before it becomes a data quality problem. A sales team rolling out a new CRM field for deal registration can measure how long the update actually takes reps in practice, not in theory, and fix the instructions or the field itself if everyone is getting it wrong the same way.
For distributed support teams working across time zones without a single trainer to shadow, this replaces "watch someone do it once" with a measurable, repeatable check against a defined standard.
A telecom retention team can record the exact save-code and billing-adjustment sequence agents are supposed to log after a cancellation call, then see which specific field new hires skip most often, before it turns into bad data in the billing system three months later.
The takeaway
The call was never the whole job. Now the part after it can be trained and measured with the same rigor as the conversation itself.
See it in Action
Tell us what you're working on, whether that's sales calls, support tickets, or leadership coaching, and we'll show you how it fits.
Talk to Sales