CASE STUDY · GAMING & ESPORTS
Nine admin hours a week returned to coaching staff
A live youth esports franchise runs our Discord-to-CRM architecture in production. Parent questions, roster changes, and session signals now land in one record instead of three inboxes.
How to read these numbers
This client asked not to be named. The figures below were measured inside the system we built and confirmed by the client, they are not estimates. (Client-confirmed · measured in-system · name withheld)
~9 hrs
Admin hours returned weekly
Measured against a pre-build baseline
1 record
Per household, not per channel
Discord, email, and billing resolved
Same day
Follow-up on flagged intents
Previously multi-day or missed
The staff stopped being the routing layer. Parent questions and roster changes land in one place now, and nobody is digging through three inboxes between sessions.
Paraphrased from the engagement and shown without a named person while we wait on client sign-off. We will not attach a name or a title to a quote the client has not approved.
Context
- A multi-team youth esports franchise coordinating parents, coaches, and players across Discord, email, and spreadsheets.
- Coaching staff were the de facto operations team, triaging messages between sessions.
The problem
- Every meaningful signal, a parent asking about a make-up session, a player going quiet, a roster gap, arrived in a channel nobody owned.
- Nothing connected a Discord handle to the paying parent record, so follow-up depended on memory.
What we did
01 · Define
Mapped the signals already being generated and agreed one metric: staff hours spent on admin per week.
02 · Design
Identity resolution from Discord handle to household record, with a confidence label on every probabilistic match.
03 · Build
Agentic classification of inbound messages into intents, routed to the CRM as owned tasks with observability from day one.
04 · Measure
Weekly comparison of admin hours before and after, reviewed with the franchise owner.
Tooling shown is what was used in this engagement, not a required stack. We build to the platforms and systems you already run.
What we would do differently
We would instrument the baseline for two full weeks before building. We had enough to prove the gain, but a longer baseline would have made the seasonality clearer.
NEXT STEP
Recognise the pattern?
Tell us where your signals live and we will tell you whether this shape fits.