Fresh Coast AIClient case study

Stanton & Company

Stanton & Company sees every client's KPI pace each Monday.

A weekly digest whose numbers come from code, not AI, built in Stanton's own Google accounts and run by their team.

“Before the digest we weren't tracking KPIs consistently. Someone would sit down every so often and spend an hour piecing together where each account stood. Now it's clear right away how we're tracking against our targets.”
Emily MichelsDirector of Operations, Stanton & Company

01 / The outcome

One view for leadership.
A system the team runs.

Every account, every Monday

Leadership sees where every account stands in one place on Monday morning, against that account's targets and goal period.

Rolled out by Stanton

A team member volunteered to own the system and rolled the full client roster onto the digest through August.

Managed by the internal owner

She now adds clients, adjusts targets, and edits the recipient list herself. When Emily joined in September, FCAI walked her through how a tracker becomes a digest row.

02 / The problem

Good records.
No weekly answer.

Stanton's account teams already kept good records. Every client had a tracker in Google Sheets, KPI targets set at the start of the engagement, and a priorities email sent by the account manager on Monday morning.

Each tracker answered for its own account. Leadership needed one weekly answer: are we on pace against what we promised each client? Seeing every account side by side meant opening each tracker and doing the math by hand, which could take hours.

Amy Stanton wanted that picture every Monday, built so her team kept working the way they already did and owned the result.

03 / The build

Meet KPI Bob.

FCAI built a weekly digest the team named KPI Bob. It reads each client's tracker and the master KPI sheet, then works out where every account stands against its goals for its own goal period.

Accounts are grouped as On track, Watch, or Needs attention, each with a one-line reason. Early-stage accounts show as Building. The digest also checks whether the week's priorities address the KPIs that need the most help.

It sends Monday and Tuesday at 9:30 a.m. Pacific. Monday is the read. Tuesday picks up late inputs and holidays, and shows what changed since the day before.

Leadership gets two formats: a short email listing only the accounts that need a look, and a full browser view with every account and every KPI.

A look at the output

The weekly picture,
at a glance.

The example below uses illustrative accounts and sample data. It shows the format, not Stanton's client information or measured results.

Weekly KPI Digest Monday3 accounts shown. Sample week.
Illustrative, sample data
Illustrative weekly KPI digest using sample accounts and sample data
AccountBandRead
Harborline NaturalsOn trackMonth 4 of 12. 5 of 6 KPIs at or ahead of pace.
Larkfield Paper GoodsWatchMonth 7 of 12. 3 of 5 KPIs on pace. Feature placements at 78% of pace.
Meadowfield KitchenNeeds attentionMonth 2 of 6. 2 of 4 KPIs on pace. This week's priorities skip both lagging KPIs.

04 / Under the hood

Familiar tools.
Clear checks and ownership.

The data path

The only new habit is one CC on an email the team already sends. Account managers keep updating their trackers and copy the digest's mailbox on their Monday priorities email.

A sheet lists every account, tracker, status, and target. The digest reads live from those files on every run, so fixing the sheet fixes the next digest.

The numbers check

Ordinary code calculates every figure, band, and flag. AI reads the week's priorities and writes the bottom-line summary.

Every number in AI-written text is checked against the computed values before sending. If the AI is unavailable, a plainer template ships instead.

The ownership model

Everything runs in Stanton's Google Workspace account and Google Cloud project, billed to Stanton and administered by their IT partner. The AI uses Stanton's own account, where prompts are not used to train Google's models.

Stanton received the code, an operations guide, an IT infrastructure guide, a failure playbook, and a one-button rerun. Run cost is a single Workspace seat plus metered AI usage.