Get a Daily Support-Metrics Summary in Slack, Automatically

That morning summary doesn't need a human. It's a fixed query against your helpdesk and a fixed message into a channel — the textbook case for a scheduled agent. With Macha you build an agent that wakes up at 9:00 AM, pulls the previous day's metrics from Zendesk, formats them, and posts a clean summary to the channel your team already reads. No dashboard to log into, no report to pull, no one to remember.

This post walks the whole build: what the agent does, how the scheduled trigger works, the exact guardrails that keep it from misbehaving, and the honest list of cases where you should not reach for this. Macha is an AI agent layer that sits on top of the helpdesk you already run — Zendesk, Freshdesk, Gorgias, or Front — so nothing here replaces your support stack; it just automates the reporting layer on top of it.

What the agent actually posts

Here's the output we're aiming for — a single, scannable card in Slack every morning:

📊 Daily Support Summary — March 4 - Tickets created: 47 - Tickets solved: 52 - First response time (avg): 12 min - Resolution time (avg): 2.4 hours - CSAT: 94% - Top topic: Order tracking (18 tickets) ✅ Cleared backlog — 5 more solved than created

That's the shape of it: a handful of metrics that matter, plus a one-line read on the trend ("cleared backlog"). Because an AI agent — not a static template — composes the message, that closing line is generated from the actual numbers. On a rough day it might say "backlog grew by 9 — created outpaced solved" instead. The summary reads like a teammate wrote it, because the agent reasons over the data instead of dumping it.

The two ingredients: a schedule and two tools

Under the hood this is one of the simplest agents you can build in Macha. It needs exactly one trigger and two tools:

Both Zendesk and Slack are first-class Macha connectors, so you authorize each once and the agent gets the relevant actions. The Slack connector's Send Message action is what lets the agent write into a specific channel; the Zendesk connector's Search Tickets action is how it reads yesterday's volume, response times, and satisfaction.

You can browse the full action set on each connector's page — the Slack integration and the Zendesk integration — to see exactly which reads and writes are available.

Building the scheduled trigger

The interesting part is the trigger. Most Macha agents fire on an event — a new ticket, an updated priority, a Slack mention. A metrics summary is different: nothing happens to set it off. It runs on the clock. That's what Macha's scheduled triggers are for.

A scheduled trigger runs your agent on a recurring interval — hourly, daily, or a custom cron expression — entirely independent of ticket activity. You point it at the metrics agent, set it to daily at 09:00 in your timezone, and you're done.

All of your scheduled agents live in one place, so you can see what's set to run, when it last fired, and toggle any of them off without touching the agent itself.

The guardrails that matter

Scheduled execution is exactly where automation can go wrong — a runaway loop hammering your helpdesk API, or a broken agent quietly posting garbage for a week. Macha builds the guardrails in so you don't have to think about them:

• Minimum 1-hour interval. You can't accidentally schedule something every minute. A daily summary is well within bounds.
• 24 runs/day cap and a 3-run org concurrency limit. Even if you stack up several scheduled agents, they can't stampede.
• Auto-disable after 5 consecutive failures, with an email notification. If your Zendesk token expires or the channel gets archived, the trigger pauses itself and tells you — it doesn't keep failing silently.
• A 60-second polling loop with stale-lock cleanup at 30 minutes, so a single hung run never blocks the next one.

These aren't toggles you configure — they're defaults that ship with the feature. (One small UX detail: the debounce control that appears on event-driven triggers is intentionally hidden on cron triggers, because "debouncing the clock" makes no sense.)

Plan limits

Scheduled triggers are a paid-plan capability. The Professional plan includes 3 scheduled triggers; Enterprise includes 20. A single daily-summary agent uses one of them, which leaves room for the obvious companions — a weekly refund report or a trending-topics alert. See the pricing page for current plan details; there's a 7-day free trial, no credit card required if you want to build one before you commit.

Writing the agent's instructions

The trigger handles when; the agent's instructions handle what. This is plain-language configuration, not code. You're telling the agent something close to:

Every morning, search Zendesk for all tickets from yesterday. Report: tickets created, tickets solved, average first response time, average resolution time, CSAT, and the single most common topic. Post it to #support-daily as a short summary with one line on whether we cleared or grew the backlog.

Because the agent reasons over what it pulls back, you can ask for things a static report can't easily do — "flag CSAT in red if it dropped below 90%," or "call out the top topic only if it's more than 20% of volume." The agent decides per-run, from the real numbers.

The metrics agent is exactly this shape: instructions that describe the summary, a Zendesk Search Tickets tool to read yesterday's volume, a Slack Send Message tool to post it, and the scheduled trigger from the section above wired in.

If you'd rather not write the instructions by hand, Macha can build the agent through conversation — describe the summary you want in the AI builder and it wires up the tools, the trigger, and the instructions for you, then you review before it goes live.

What it costs to run

A scheduled agent consumes credits the same way every other Macha agent does: per AI action, not per report. Credit cost depends on the model you pick — roughly 0.5–9 credits by model, with the default GPT-5.4 Mini at 1 credit. A daily summary is a lightweight job — one read, a bit of reasoning, one write — so the per-run cost is small, and there are only ~30 runs in a month. For a recurring report that would otherwise eat a lead's time every single morning, the math is not close.

Crucially, deflection vs. resolution vs. automation are different things in Macha's pricing. This use case is pure automation — Macha isn't resolving a customer ticket, it's orchestrating a report — so you're paying for the actions the agent takes, full stop. No per-resolution fee, no outcome pricing.

How this compares to the usual approaches

You have other ways to get numbers into Slack, and it's worth being straight about them:

• Native Zendesk → Slack triggers push individual ticket events into a channel ("ticket #123 updated"). This is real and well-documented — you add the Slack app to a channel, then build a ticket trigger that notifies the Slack integration. It's great for live awareness and useless for a daily roll-up — triggers fire per event and don't aggregate or summarize.
• BI/dashboard tools like Geckoboard can schedule a snapshot image of a dashboard into Slack daily, weekly, or monthly (Geckoboard's own docs walk through the setup). That works, but it's a picture of a chart, not a written summary you can read on your phone in two seconds — and it's another tool and subscription to run.
• Workflow builders like n8n or Zapier can absolutely stitch Zendesk → format → Slack on a schedule; n8n even ships a Zendesk ticket summarizer that runs daily at 10am and posts to Slack. The trade-off is that you build and maintain the workflow — the trigger, the formatting, the LLM node, the credentials — and keep it patched as APIs change.

Macha's angle is that the same agent platform already wired into your helpdesk for live ticket work also runs the report — and an AI agent, not a hand-built template, writes the summary. One system, and the prose adapts to the day. If you're standardizing on Zendesk specifically, the Macha on Zendesk page covers the rest of what the layer does.

Watch-outs: when not to reach for this

A scheduled Slack summary is genuinely useful, but it's not the right tool for everything. Be honest with yourself about these:

• You need real-time, not daily. If the point is to catch a spike as it happens — an outage, a flood of "can't log in" tickets — a once-a-day summary is the wrong shape. Use an event-driven trending-topics alert instead, which watches volume continuously and pings the channel when something jumps.
• You need deep, sliceable analytics. This is a glanceable roll-up, not a BI replacement. If you want to pivot CSAT by agent, by tag, by hour, across a quarter, that's a dashboard job (or a Macha Study over your ticket history) — not a Slack one-liner.
• Your metric definitions are contested. "Resolution time" and "first response time" mean different things in different orgs (business hours? calendar hours? reopens?). The agent reports what Zendesk reports — make sure the team agrees on the definitions before you anchor a daily ritual to them.
• The channel is already noisy. A daily post into a busy firehose channel gets ignored. Give it a dedicated, low-traffic channel (#support-daily) so the summary actually gets read.

None of these are reasons to skip the agent — they're reasons to scope it right. For most teams, a single clean morning card in a quiet channel is exactly enough.

FAQ

What metrics can it include? Anything Zendesk's Search Tickets action can surface for a date range — tickets created and solved, first response and resolution times, CSAT, ticket volume by tag or topic, backlog. You define the list in the agent's instructions, and the agent composes the summary from what it pulls.

Can I change the time or frequency? Yes. The scheduled trigger supports hourly, daily, or a custom cron expression, with a minimum 1-hour interval. Set it to 9:00 AM, twice a day, weekday-only — whatever fits your standup.

Does it work with Freshdesk instead of Zendesk? The pattern is identical; you'd swap the Zendesk read for the Freshdesk equivalent. Macha connects to Zendesk, Freshdesk, Gorgias, and Front as helpdesk connectors, plus Slack for delivery.

What happens if it breaks — say my Slack token expires? The trigger auto-disables after 5 consecutive failures and emails you, rather than failing silently or spamming the channel. You fix the connection and re-enable it.

Is this a Zendesk replacement? No. Macha is an AI agent layer that runs on top of your existing helpdesk. The metrics still live in Zendesk; Macha just reads them on a schedule and writes the summary to Slack.

How much does it cost to run? Credits are consumed per AI action (roughly 0.5–9 by model; default GPT-5.4 Mini = 1). A daily summary is a single lightweight run, so monthly cost is minimal. See pricing for plan details, and start a 7-day free trial, no credit card required to build one.

Build your morning summary

If your team's day already starts with someone pulling yesterday's numbers by hand, this is the first agent worth building — small, safe, and immediately useful. Start a 7-day free trial, no credit card required, connect Zendesk and Slack, point a scheduled trigger at a metrics agent, and let it post your support summary at 9:00 AM tomorrow. The docs have the step-by-step, and there are plenty of other use cases — refund reports, VIP alerts, trending-topic warnings — that follow the exact same scheduled-agent pattern.

Written by Abbas (Customer Support & AI, Macha) · Reviewed by Ankeet Guha (Co-founder & CTO) · Published 2026-06-24 · Last updated 2026-06-24.