Triage RingCentral voicemails and alert Slack on urgent ones

Every 30 minutes, triage my new RingCentral voicemails so the urgent ones get seen in minutes instead of at the end of the day.

Start with RingCentral "List Messages" against my message store, filtered to voicemail message type, unread read status, and a date range starting at the previous run. Keep track of the voicemail ids you have already handled and skip any that appear again, so an overlapping window never produces a duplicate alert. Do not change read status in RingCentral: my team uses unread as their own worklist, so leave it exactly as they left it.

For every new voicemail, call RingCentral "Get Message" to pull the full detail, including the caller number, the timestamp, the duration, and the voicemail transcription text.

Read each transcription and classify it on three axes. First, what the caller actually wants, in plain words. Second, how urgent it is. Third, which bucket it belongs in: support problem, sales enquiry, billing question, or spam and robocall.

Treat a voicemail as urgent when the caller describes something broken or down, is blocked from using the product, sets a same day or next morning deadline, threatens to cancel or escalate, sounds seriously angry, reports anything with a safety, security, or payment failure angle, or is an obviously large inbound sales opportunity. Everything else is routine, including general questions, callbacks with no deadline, and vendors.

Post anything urgent to Slack immediately using "Send a Message" to my urgent voicemail channel. Include the caller number, the time it came in, the bucket, a summary of no more than two lines, and an explicit line explaining why it was flagged urgent. One message per urgent voicemail, posted as soon as you have judged it, not held for the digest.

Do not ping per voicemail for the routine ones. Collect every non urgent voicemail from this run into a single digest post to my voicemail digest channel using Slack "Send a Message", grouped by bucket, with one line each covering the caller number, the time, and what they want. If there are no routine voicemails in a run, post nothing at all rather than an empty digest.

For genuine sales and support callers, look the caller number up in HubSpot with "Search Contacts", trying the number in a few common formats before concluding there is no match. Then attach the follow up with HubSpot "Create Task": subject naming the caller and what they want, body containing your summary plus the relevant part of the transcription, associated to the matched contact when you find one. Make support tasks due the same day and sales tasks due the next business day, and set high priority on anything you flagged urgent. If there is no matching contact, still create the task with the caller number in the subject so an unknown caller never gets lost.

Silently drop the noise. Obvious robocalls, spam pitches, automated collections and warranty messages, empty or silent voicemails, and dead air with no usable transcription get no Slack post, no digest line, and no HubSpot task. The one exception is a voicemail that clearly has real speech but no transcription available: mention that one in the digest as needing a manual listen, with the caller number and the time.

Alert fatigue is what kills this workflow, so be strict. If you are genuinely torn between urgent and routine, put it in the digest rather than the urgent channel.