WorkServicesStudioInsightsContact Start a project ↗
Index/ Work/ 005 · ADREP AI
CLIENT · ADREP  ·  GLOBAL · 2026

AI ad analysis
that writes
the report for you.

Adrep is a GPT-powered ad-analytics platform that turns Meta, Google, and LinkedIn campaign data into a written, shareable performance report with ROAS / ROI / conversion recommendations — in three steps and under a minute.

Client Adrep
Year 2025 — 2026
Practices AI · Software · Mobile
Live at adrep.ai ↗
Adrep AI — product screenshot

/01   THE BRIEF

Freelance marketers and agencies were spending half their week building looker-studio dashboards for client reports. Adrep was launched to compress that into a single prompt.

Three steps: paste your campaign metrics, the AI engine writes the analysis, the report goes out via a PIN-protected share link. Mobile apps for iOS and Android plus a web app, multilingual EN/TR, four pricing tiers from free to enterprise.

Time per report <1min From data paste to shareable PDF.
Channels 3+ Meta · Google · LinkedIn supported at launch.
Free tier 3/mo Free reports per month, no card required.
Languages EN · TR Full localisation from day one.

/02   THE APPROACH

We treated the report itself as the product — not the dashboard, not the integrations. Marketers want the document; everything else is plumbing.

The Inertia + Vue 3 web app shares its design system with the iOS + Android apps. The AI engine is provider-agnostic — Adrep can swap from one frontier model to another via a single config line, with eval coverage on the report-quality regressions.

"

We bill clients for the insight, not the spreadsheet. Adrep made the insight the deliverable.

Performance Marketer
Independent agency · Adrep customer
WEEK 01 — 02
Diagnose

Interviewed 12 marketers across Türkiye and EU. Tore down their current weekly reporting flow.

WEEK 03 — 06
Prototype

Three-step report builder + first AI-generated PDF. Internal pilot with two partner agencies.

WEEK 07 — 16
Build

Production web app, iOS + Android apps, billing, share-link auth, multi-language, observability.

WEEK 17 — ∞
Operate

Add channels (TikTok next), tune the eval set, optimise per-report cost as models evolve.

WEB Laravel
FRONTEND Inertia · Vue 3
STYLE Tailwind
MOBILE React Native
LLM OpenAI
EVAL Promptfoo
DB Postgres
QUEUE Redis
BILLING Stripe
OBSERVE Sentry
NEXT CASE STUDY · 005 / 010
Metonnect
BROWSER-NATIVE VIDEO CONFERENCING

Have a brief like Adrep's?

We'll reply within one working day.

Start a project