InfoAdex

Internal product + AI · 2025–2026

InfoAdex is the go-to source on advertising activity in Spain: it collects and analyzes all the advertising in the country’s media. Its day-to-day work runs on internal tools that took at least a year to master. We lead UX/UI design and product management on the project that’s rebuilding them, with AI built in.

Design pieces from the InfoAdex project: the 2026 report and data cards
Client
InfoAdex
Year
2025–2026
Our role
UX/UI design and product management
Platform
Internal tools and client software
Status
Phase 1 done · phase 2 underway
−75%
learning time on the tools
+425,000
more ad placements processed every month
40 → 10%
of managers’ time spent on support

Context

01

Measuring a whole country’s advertising.

The goal: rebuild the internal tools and the client software, and take on more information and more data enrichment with the same team.

One key area is creative review: the team sorts the creatives coming in from suppliers into 28 sectors, each with its own fields.

The challenge

02

More data, same team.

The pain points we found in discovery:

  1. 01The database architecture had to change.
  2. 02More and more volume and enrichment to process.
  3. 03Training someone on the tools took at least a year.
  4. 04Middle managers spent at least 40% of their time on support.
The workflowCreatives come in, several AI models handle the repetitive work, and a quality check reviews what people and AI have enriched.

How we worked

03

The process, step by step

  1. 01

    Discovery

    The four pain points, pinned down with the team.

  2. 02

    Planning

    Two phases (internal tools and client software), the project methodology, and support for the new in-house dev team as it ramped up.

  3. 03

    Design and specs

    Wireframes for each view and one ticket per view with goal, scope, and acceptance criteria, in the client team’s template.

  4. 04

    Applied AI

    A workflow with multiple AI models for the repetitive work, quality control over what people and AI enrich, and a copilot for the team.

  5. 05

    Phase 2

    The client software, in progress.

Problems & fixes

04

Problems we ran into and how we fixed them

  1. A year to learn one tool

    Problem
    Training on the internal tools took at least a year.
    Fix
    A UX/UI redesign. In creative review: a read-only supplier card, an editable creative card, and no more than five fields per row.
    Result
    75% less time to learn.
  2. Managers turned into a help desk

    Problem
    At least 40% of middle managers’ time went to answering questions.
    Fix
    A copilot agent that answers the team’s questions.
    Result
    Support drops to 10% of their time.
  3. More volume with the same team

    Problem
    Every month brings more volume and more data enrichment.
    Fix
    Multiple AI models handle the repetitive work, and a quality check reviews what people and AI enrich.
    Result
    425,000 more ad placements processed every month.
  4. Divvying up the work without doing it by hand

    Problem
    Assigning each reviewer the placements still waiting to be classified.
    Fix
    Task templates with recurring or one-off scheduling, filters, and reviewer assignment, specced out in four tickets.

What we built

05

Pieces of the project

  • Workflow diagram: planning, cross-media view, and AI-powered reporting
    01

    The workflow

    Planning, a cross-media view, and AI-powered reporting.

Results

06

What we pulled off

  • −75%

    learning time

    On the internal tools.

  • +425,000

    placements a month

    More ad placements processed with the same team.

  • 10%

    of managers’ time on support

    Down from at least 40%.

Services

  • Product
  • UX/UI
  • Applied AI
  • Product management

Built with

  • Multiple AI models
  • Data quality control
  • Copilot agent