AI client reporting

AI client reporting that actually knows your clients.

Generic AI summarises a chart. Scout makes the AI know the client - its strategic priorities, decision rules, voice, and this week’s live numbers - so the report is right, not just fluent.

Claude
Wednesday · 2:41pm
Quick gut check before my call - how’s Driftline tracking, and should we push more into cold this week?

Solid week - but I’d hold on cold. Three things:

memoryDriftline weighs ROAS over volume - efficiency first. (Strategic priority, reinforced 6×)
live dataCold ROAS softened to 1.9× this week, down from 3.1× last week.
memoryYour rule: pause the bottom creative before adding spend when cold ROAS holds under 1.8×.

So: pause Surfboard Hero v4, hold cold budget, and let the Welcome flow keep carrying email ($8.40/recipient). Want me to draft the client note?

What Scout did underneath
01Loaded Driftline’s memory

Eight facets, decision rules, and voice v7 - the agency’s, not one account manager’s Claude account.

02Pulled live data

This week’s Meta and Klaviyo numbers, fetched on the spot - no dashboard, no exports.

03Answered in your voice

Short, efficiency-first, and it offered the next step - exactly how your team writes.

The problem with bolting AI onto a dashboard is that the AI has no context - it can describe a number but not what it means for this client.

Scout is the context layer. It loads each client’s structured memory and your agency’s voice into Claude before a word is written, so the output reads like it came from someone who knows the account.

Context, not just numbers

Eight structured facets per client - priorities, rules, quirks - feed every answer before live data does.

Memory that compounds

Every conversation and edit updates the client profile. The AI gets sharper every week.

Answers in your voice

Not Claude’s default - your agency’s phrasing, emphasis, and the next step it would offer.

How it works

From connected accounts to a sent report, without the busywork.

01
Connect each client once

OAuth into each client’s Shopify, Klaviyo, Meta, Search Console and GA4 - read-only, agency-wide. About five minutes per client, then you never touch it again.

02
Scout assembles the week

It pulls live data from every connected platform, loads the client’s structured memory, and applies your agency’s voice profile - all before anyone asks.

03
Claude drafts the report

Ask in Claude and the update is there: headline performance, KPI commentary against the client’s targets, per-channel reads, and recommendations - written the way your team writes.

04
You edit and send

Tighten a sentence, publish to a private link, or send it to the client. Scout writes; the account manager always ships.

Always included

Every Scout plan, not an upsell.

Structured per-client memory, owned by the agency - not one person’s Claude account
An agency voice profile that sharpens with every edit you make
Daily briefings and real-time signals running quietly in the background
One shared workspace, so every account manager works from the same source of truth

Common questions

Which AI does Scout use?

Claude. Scout runs as an MCP server that gives Claude the context - your client memory, voice, and live data.

Does my data train the model?

No. Your data is never used to train models, and is tenant-isolated per agency.

Is Scout a dashboard?

No - deliberately. Scout writes the report. Charts only render inline in a published report when they earn it; the deliverable is the written update, not a dashboard your team has to build and maintain.

Does Scout write to my clients’ ad accounts?

Never. Every integration is read-only, via OAuth per client account. Scout reports; humans make the changes.

What does it cost?

Free for one client, forever. Pro is $149/month for unlimited clients - no per-client fees, no per-seat charges. Beta agencies lock a lifetime discount.

See it on your clients.

A lifetime beta discount on Pro. Onboarding in under an hour.