Good week for builders. An async version of the aggregate analytics API, guide generation from transcripts in AI Studio, and, at long last, scheduled triggers for workflows. Here's my read on each.
TL;DR
- Submit big analytics queries as a queued job instead of chopping them into interval-sized chunks. Same request body as the synchronous aggregate endpoint; results when the job finishes.
- Point AI Studio at a queue and it turns 30–1,000 real transcripts into structured, step-based guide drafts.
- Run workflows on a clock (interactive schedule or cron expression) without an external scheduler or a data action hack.
Platform and API: async aggregate queries
The aggregate metrics endpoint has always been the workhorse of Genesys Cloud reporting integrations, and its biggest pain point. Pull a wide date range at high granularity and you had to split the work into a pile of smaller synchronous requests, then stitch the results back together around rate limits. That code is a pain to write and a bigger pain to maintain.
The async variant flips this to request-and-wait: you submit the query, the platform queues it, and you collect results when the job completes. The input contract is the same as the existing aggregate endpoint, so migrating is mostly swapping the call pattern (submit, poll, fetch), not rewriting your query logic. I'd move nightly warehouse loads and big historical backfills over first. Keep the synchronous endpoint for dashboards where you need an answer now.
Self-service and automation: guides drafted from your transcripts
Genesys AI Studio can now generate draft guides directly from real customer–agent conversation transcripts. You pick a single queue, a channel (voice or message), a language, and a date range containing at least 30 transcripts. The system analyzes up to 1,000 of them, clusters the conversations by topic, and produces a structured, step-based draft that mirrors how your agents resolve those issues in practice.
This attacks the blank-page problem in automation design. Instead of interviewing supervisors and reverse-engineering desktop procedures, a designer starts from a draft grounded in real interactions, then edits and publishes.
The prerequisites: Virtual Agent enabled in the org, AI Insights enabled at the program level, AI Studio guide permissions, and queue view permission. Licensing is Genesys Cloud 4 or the AI Experience add-on; if you're on neither, skip this section. Generation can take several minutes on larger transcript sets. It's a design-time tool, not something you run in a loop.
Treat the output as a first draft in the strictest sense. Transcripts capture what agents do, including workarounds and policy drift, so review each generated guide against your intended process before it gets anywhere near a production Virtual Agent flow.
Self-service and automation: scheduled triggers for workflows
Triggers in Genesys Cloud have always been event-driven. This release adds a Scheduled Triggers tab (Admin > Orchestration > Triggers) where a workflow can run on a recurring schedule, either through an interactive picker (times, weekdays, days of month, months) or a five-field cron expression. An explicit time zone is required at creation.
Two constraints. Frequency is capped at two executions per hour per trigger, so this is for hourly and daily housekeeping, not near-real-time polling. And execution lands anywhere up to five minutes after the configured time (the offset is consistent per trigger), so don't build anything that assumes to-the-second precision. Scheduled triggers also pass no inputs to the workflow; the flow has to work out its own context when it wakes up.
Even with those limits, this closes a real gap, and it should have shipped years ago. Teams have been faking cron with external schedulers calling the API, or with awkward in-flow loops. Nightly data syncs and morning readiness checks can now live entirely inside the platform, governed by the Process Automation > Trigger permission set.
What this means for your contact center
The common thread this week is Genesys moving work you used to do around the platform into the platform. The async aggregate endpoint removes a whole category of client-side batching code from reporting integrations. Scheduled triggers remove the external cron box, along with its credentials and monitoring. Both shrink the amount of custom glue you own, which is where most Genesys Cloud integrations break.
Guide generation is the more strategic one. If you're on Genesys Cloud 4 or AI Experience and haven't adopted guides because authoring them is labor-intensive, a pilot just got a lot cheaper. Start with your highest-volume, best-transcribed queues; topic clustering does its best work on repetitive conversations.
One caution. Anything that generates content from transcripts inherits the quality of your transcription and program configuration. If transcription accuracy is poor on a target queue, fix that first or the drafts will reflect it.
What to check
- Inventory the reporting jobs that split aggregate queries into chunks to dodge limits, and plan to move the biggest ones to the async endpoint.
- Prototype the submit/poll/fetch pattern against one existing query before committing; the input contract is the same.
- List every external cron job or third-party scheduler that exists only to kick off a Genesys Cloud workflow. Those are retirement candidates.
- Check that the two-runs-per-hour cap and the five-minute execution window fit your use case before migrating anything time-sensitive.
- Pick one high-volume queue with at least 30 transcripts in a recent window, generate a draft guide, and review it against your documented process.
- Confirm who holds Process Automation > Trigger permissions and AI Studio guide permissions before turning either feature on broadly.
Full details are in the official Genesys Cloud release notes for April 20, 2026. Questions about your deployment? Email me. Reading these notes closely is my job.