Work · Independent — engaged directly, not through the LLC

Verizon

Forty-two designers had every AI tool. Nine months later, nothing had improved.

So we stopped asking how to get people to use AI, and asked where the pain actually lived — then built the systems that sat exactly there.

One-hour interviews, about yesterday’s work — not about AI
16
Friction points mapped, then ranked by severity, business impact and AI opportunity
40+
Designers first, then cross-functional stakeholders, across six brands
42→80+
New tools bought. The environment allowed two, and that was the brief.
0
THE WORKFLOW, UNCHANGED BRIEF PLAYBOOK JOURNEY DESIGN PROTOTYPE HANDOFF QA LAUNCH WHAT WE PUT UNDERNEATH IT Organisational memory — every meeting, decision and rule, answerable in plain language Edge-case scan — the cases that used to surface after handoff Behavioural audit — validation where there was no budget for any EACH BAND STOPS AT THE PHASE ITS PROBLEM USED TO SURFACE IN
Drawn for this page. The systems themselves sit behind an enterprise login and are not shown.

01

The diagnosis came before the prescription.

Nine months of high adoption and no operational improvement said the problem was not the tools. Sixteen interviews about daily reality, not about AI, made the map everything after it was built against.

02

AI amplifies the architecture it finds, broken or not.

Fragmented knowledge returns fragmented answers faster. So each system sat in the workflow at the point its failure lived, not beside it as one more tool to adopt — nobody had to change how they worked, which is why nobody had to be persuaded.

03

The constraint was the brief, not the obstacle.

No custom models, no API access, no budget, and a restricted environment. Everything was built from what the organisation already allowed, which is why it survived contact with the place it had to run in.

Client
Verizon — forty-two designers, then eighty-plus people across six brands
Role
Diagnosis, system design, the build, a pilot before anything was rolled out, and training the teams that now run it
What we made
A central project knowledge system, a set of role-specific agents, and an AI-assisted heuristic evaluation workflow
Alongside
The post-login experience for prepaid
Reported back
Fewer decisions resting on who happened to remember, and less rebuilt late — described by the teams, not measured against a baseline

If your teams have the tools and nothing has changed.

Find where the pain actually lives before anything is bought or built, then put the systems at those points, inside the constraints you already have. What we would need from you is an hour with each of the people doing the work. We reply from New York.

hello@creativityiseverywhere.com

All nine engagements