1 of 11

CASE STUDY · MERKLE (dentsu) : GLOBAL DESIGN PRACTISE TRANSFORMATION

Human Led, �Agent Powered Design Capability Transformation.

A modern Design Delivery model, scaled for the speed of customer needs that I built and implemented globally for all studios. Replacing the double diamond for 350 designers.

Chad Borlase : SVP Experience Design Lead

2 of 11

The problem was the traditional design model was �broken for a brave new world.

The classic double diamond philosophy wasn't failing because the designers were bad. It was failing because the model was built for a different world that no longer existed and Merkle was falling behind.

Linear by design

01 · THE TIME SENSITIVE CHALLENGE

Chad Borlase : SVP Experience Design Lead

Research fed define. Define fed develop. Develop fed deliver. By the time work arrived, the brief had moved.

No parallel execution

One phase had to complete before the next began. In an AI-enabled world, that sequencing is a competitive liability.

Handoffs as failure points

Every transition between design and engineering lost context, fidelity, and time. The double diamond assumed clean handoffs exist.

No agent support layer

Volume work — research synthesis, journey mapping, variant generation — consumed senior designers. Agents could handle all of it.

2

2

2

3 of 11

What I built to replace it.

A new philosophy

1.

02 · THE SOLUTION

A new framework

2.

A new delivery model

3.

Chad Borlase : SVP Experience Design Lead

Human-Led. AI-Enabled. Agent-Powered. Humans own direction, judgment, and accountability. Agents handle execution, automation, and scale. Not AI-first. A specific position on where the line is.

X4 IGNITE. Three phases running in parallel: Ignite the Engine, Concept at Scale, Forge in Parallel. Infrastructure live before concepts begin. Design and engineering collapsing into a single motion.

Pod-based teams cast for the engagement, not the org chart. Human-to-agent ratio set by complexity. 15+ AI agents defined, built, and integrated across the full design lifecycle.

3

3

3

4 of 11

X4 IGNITE : DELIVERY IN AN AGENTIC DESIGN WORLD

THREE PHASES : RUNNING FASTER THAN THE DOUBLE DIAMOND COULD RUN ONE

03 · THE FRAMEWORK

Chad Borlase : SVP Experience Design Lead

PHASE 01

Ignite the Engine

2 weeks

The infrastructure is live before design begins. Tokens, tooling, AI context. Everything the sprint needs to move without friction is in place on day one.

PHASE 02

Concept at Scale

3 weeks

North Star vision translated into composable XD concepts. AI prototyping, AI concept testing, AI-facilitated user stories — all running in parallel with human concepting.

PHASE 03

Forge in Parallel

8-10 weeks

Pods form around capabilities, not handoffs. Design and engineering collapse into one motion. Every component production-ready. Change management runs concurrently.

4

4

4

5 of 11

2-WEEK�SPRINT

2-WEEK�SPRINT

2-WEEK�SPRINT

EXPAND

EXPERIMENT

EVOLVE

Forge in Parallel

Pods form around capabilities, not handoffs. Design and engineering collapse into a single motion — every component composable, every sprint production-ready. Change management and training run in parallel, so the org is ready the moment the system is.

Concept at Scale

Ignite the Engine

The engine doesn't wait for design to start — it starts before it. In two weeks, we lay the design and engineering foundation that makes every moment that follows faster, more consistent, and harder to derail. Tokens, tooling, and context. All of it live before concepts begin.

EXPLORE

MLP

Translating the vision into tangible XD Concepts, defining high impact patterns across key journeys. By applying Atomic Design principles, we ensure these concepts are broken down into a scalable framework optimized for a composable stack, that can be used across digital experiences, dropping dev debt later.

NORTH STAR VISION

2 weeks

3 weeks

8 - 10 weeks

Merkle’s continuous human-led, agent-powered approach

Human

Agent

Human

Agent

X⁴ IGNITE is how design engineering gets delivered.

Rooted in efficiency and impact, we unify Strategy, Design, Engineering and Research to turn creative vision into a scalable system.

The result is an ecosystem designed to scale, adapt, and continuously evolve.

From vision to system with �X4 IGNITE

THE AI POWERED DESIGN & ENGINEERING APPROACH

TRAIN AGENTS

Agent

Human

2-WEEK�SPRINT

  • AI prototyping
  • AI concept testing and validation
  • AI Facilitated User Stories
  • Automated Design Token Extraction
  • Intelligent Road mapping and Planning
  • Human design concepting

  • Brand guidelines
  • Journey maps
  • Heuristics
  • Primary research
  • API and Integration Inventory
  • Content Taxonomy
  • Functionality Discovery

  • Claude
  • Figma MCP
  • Askable
  • Dovetail
  • Perplexity
  • Claude
  • Askable
  • Figma Make
  • Galileo
  • Relume
  • Claude Code
  • Lovable
  • Cursor Bolt
  • Writer
  • Claude
  • Askable
  • Figma Make
  • Axel
  • Claude Code
  • Lovable
  • Cursor Bolt
  • Maze
  • Firefly
  • Co-Pilot
  • V0
  • Chromatic

INPUTS

AI TOOLING

  • Human detailed design
  • AI prototyping
  • AI A/B concept testing ongoing
  • Automated CMS Component Scaffolding
  • AI Code Assistance and Review
  • AI Assisted Functional, Performance, Security and Visual Testing
  • Intelligent data transformation and migration

5

5

6 of 11

The right mix for the right job

Human craft and judgment remain at the core, with AI continuously enhancing efforts at varying levels - all on a unified, scalable experience layer. This spans from people augmented by AI tools to autonomous agents operating within defined guardrails under human oversight

03 · THE FRAMEWORK

Chad Borlase : SVP Experience Design Lead

Human in the Loop

Creative and strategic work with AI augmentation

AI Agent

  • Human drives�AI accelerates
  • Least risk

Human on the Loop

Agents working with real-time human review

  • Agents execute�Human reviews and intervenes in real time to reduce risk

Human above the Loop

Agents operating autonomously, humans setting direction and guardrails

Guardrails

AI Agent

AI Agent

AI Agent

  • Human sets direction�Agents operate within pre-determined guardrails to reduce risk

AI Agent

AI Agent

6

6

6

7 of 11

One size fits none – each POD is cast with the right HUMAN & AGENT mix

Blending human judgment with AI execution

03 · THE FRAMEWORK

Chad Borlase : SVP Experience Design Lead

Human

Hybrid

(Human + Agent)

AI Agent

Product

Owner

Experience

Strategy

ACD UX/ UI

Lead

Architect

Business

Analyst

UX Designer

Tech

Strategy

UI Designer

Developer

QA

Copywriter

Developer

Content

Creative

Research

ACD UX/ UI

Lead

Architect

Tech

Strategy

UI Designer

Developer

Product

Owner

Lead

Architect

QA Lead

ACD UX/ UI

QA

Copywriter

Developer

Content

Creative

Research

Product

Owner

QA

UX Designer

Pod A: Build something complex together

Pod B: Execute a defined scope efficiently

Pod C: Scaled production

Copywriter

Research

EXAMPLE POD STRUCTURES

7

7

7

8 of 11

15+ AI agents. Built, not theorized (see appendix for details)

04 · DELIVER FAST AND TEST, AND GROW

Brief Distillation

Messy briefs in. Structured strategy brief out. Faster alignment from day one.

Research Synthesis

Qual + quant ingested. Insight clusters and unmet needs produced in days, not weeks.

Journey Mapping

Multi-persona maps with evidence citations, running in parallel. Hours, not weeks.

UI Variant Generation

20+ on-brand screen variants from approved concepts. Designers curate, not draw.

Design to Code

Figma MCP to production-grade code. Accessibility checked. PR ready for review.

System Governance

Design system health monitored continuously. Drift flagged automatically. Scales past manual review.

Chad Borlase : SVP Experience Design Lead

8

8

8

9 of 11

The value obtained.

05 · OUTCOMES

BEFORE · Double Diamond

AFTER · X4 IGNITE + Agents

Linear phases. One finishes before the next begins.

Senior designers doing research synthesis and journey mapping.

Handoffs between design and engineering lose context and fidelity, increasing Tech dept and eating into budget.

One team model for every engagement.

No AI in the delivery workflow.

Parallel execution. Research, concepting, and build run simultaneously.

Agents handle volume work. Designers make the calls that matter.

Design & engineering in a single pod. One motion, one production-ready output.

Pods cast by complexity. Human-to-agent ratio set per engagement.

15+ agents integrated across the full lifecycle. Every step human-reviewable.

Chad Borlase : SVP Experience Design Lead

9

9

9

10 of 11

What this means for your organization

06 · THE ADVANTAGE

  • I built a new delivery model for a 350-person global practice — not a pilot, a full practice transformation.

  • I designed the agent library from scratch: 15+ AI agents with defined inputs, outputs, and team ownership.

  • If your design function needs to integrate AI without losing quality or creative authority, �I have already built the answer.

Chad Borlase : SVP Experience Design Lead

Most design functions are still running the double diamond and calling it agile. �That gap is widening.

10

10

10

11 of 11

AGENT: Brief Distillation

Takes messy client briefs, SOWs, or kickoff notes and produces a structured strategy brief: problem, audience, constraints, success metrics, and open questions.

Key Outputs

  • Structured one-page strategy brief
  • Stakeholder assumption log
  • Clarifying questions for client

Value by team

Strategy – Spend time on thinking, not formatting

XD – Clearer constraints before design begins

All — Faster alignment from day one

AGENT: Competitive Research

Analyzes competitor digital experiences against a defined rubric (UX patterns, CX moments, personalization signals, GTM positioning, etc.) and identifies whitespace opportunities and saturated territory.

Key Outputs

  • Differentiation opportunity summary
  • Feature/pattern benchmarking matrix
  • Experience gaps flagged by priority

Value by team

XD — Audit coverage at scale without pulling senior designers

Design Systems — Signal on which components cause friction

All — Credible accessibility posture before client review

AGENT: Research Synthesis

Ingests qual and quant research (interviews, surveys, analytics) and surfaces themes, tensions, and opportunity spaces. Supports North Star visioning, roadmap definition, and stakeholder alignment.

Key Outputs

  • Insight clusters with supporting evidence
  • Unmet need statements ranked by frequency
  • Draft "How Might We" prompts

Value by team

Strategy — Rapid synthesis, getting to insights faster

XD — Designs from evidence-backed insights with traceable rationale, making it easier to defend decisions

All — Insight turnaround in days, not weeks

APPENDIX : Agent Candidates

These support north star activities

All

Strategy

XD

Design Systems

AGENT: Journey Mapping

Takes research transcripts, analytics logs, and support ticket data and produces structured journey maps showing user goals, actions, friction moments, and emotional states at each stage. Runs against multiple personas in parallel.

Key Outputs

  • Multi-persona journey maps with evidence citations
  • Friction heatmap ranked by user impact
  • Opportunity moments tagged by experience stage

Value by team

Strategy — Evidence-backed journey work that holds up in debate

XD — Journey artifacts ready for critique in hours, not weeks

All — Consistent journey framing across every workstream