Part 1 The problemWhy teams need this
01 · THE PROBLEM
When support lives in three inboxes, urgent tickets wait behind easy ones and easy ones wait for a human. One router clears the routine questions, around 40% of the volume, and gets the rest to the right person fast.
Part 2 How it worksWhat it does, step by step
02 · WHAT IT DOES
Every support message, wherever it arrives, becomes one ticket. AI sorts it by urgency and topic, answers the routine how-to questions from your help centre, and sends everything else to the right team in Slack with the SLA clock already running.
03 · HOW IT RUNS
Step by step, as built.
One ticket shape
Email, chat and WhatsApp messages are normalised and merged.
Classify
Each ticket is sorted: urgent, how-to, billing or other.
Answer what's routine
An agent answers how-to questions from the help centre, and only a confident answer is sent.
Escalate the rest
Urgent, billing and uncertain tickets go to the right team in Slack.
Track the SLA
Every ticket is logged with its SLA clock; a check every 15 minutes escalates tickets close to breach.
04 · WHERE A PERSON STAYS IN
The machine drafts. A person decides.
The AI only replies when it's confident. Every other ticket goes to a person, with the SLA clock visible.
05 · TOOLS AND APPS
Built around the systems already in the process.
06 · WHEN SOMETHING BREAKS
Failure is designed in.
- 8nodes retry automatically when an external API fails.
- 1decision point (IF or Switch) checks the data before it moves on.
- ✓Anything unhandled triggers our central error workflow, so a crash gets reported instead of failing silently.
Standard on every build
- Schema validation before downstream writes
- Retry and error routes for external API failures
- Duplicate-safe processing and idempotent updates
- Human approval where the action carries business risk
- Execution logging for support and audit review
Part 3 The impactWhat it's worth, and how we'd build yours
07 · PROJECTED IMPACT
What it should change in the business.
Projections for a typical deployment. The calculation below shows the math, and you can put in your own numbers.
08 · ROI CALCULATION
How it pays back in your business.
The starting numbers are a hypothetical deployment sized to the projections above. Change any of them to your own volumes and costs, and the math updates underneath.
- Full-time equivalent freed
- 1.1 people
- Gross value
- $76,800
- Running cost
- −$1,440
- Return per $1 of running cost
- $53.3
The math: 3,000 support tickets × 8 min × 12 months × 40% ÷ 60 = 1,920 hours a year × $40/hour = $76,800. Net value = gross value − $1,440 running cost a year.
09 · HOW WE'D BUILD YOURS
How we'd build yours.
- Discover: map the current process, systems, volumes, owners and exceptions.
- Design: define the canonical data model, approvals, retries and system boundaries.
- Build: implement credentials, nodes, validation and observable error routes.
- Prove: run controlled data through success, duplicate and failure scenarios.
- Operate: publish runbooks, ownership and measurable service levels.
