Intelligence Brief №01 — AI Fluency — June 2026

BCG has four levels.
We have seven.

The firm just launched its AI Fluency Ladder. Design Studios already has something more sophisticated — and the comparison reveals exactly where our advantage lives, and what's at risk if we don't protect it.

7 DS Fluency Levels
4 BCG Firm Levels
2 Levels BCG Doesn't Map
~400 DS People In Scope
01 — Context

The firm moved.
We were already there.

In June 2026, BCG launched the AI Fluency Ladder — a firm-wide, four-level certification programme open to all BCGers. Tool-agnostic, role-specific, tied to BCG's 2030 strategy. 60% of the firm already uses AI habitually. The ladder pushes everyone further.

Design Studios has been running parallel to this for months. The DS AI Fluency Framework — seven levels across three bands — was built for a creative studio workforce, not a consulting firm. It goes further at the top, is more specific in the middle, and more honest at the bottom.

These aren't the same ladder. The question is whether DS positions its framework as distinct and deeper, or allows it to be absorbed into a four-level model that doesn't have language for what our best practitioners actually do.

"If it ever becomes a promotion gate, the rungs become a lie. Keep it descriptive. Keep it public. Keep it out of HR systems."
DS Creative AI Ladder — Design Principle
The Competition Problem

MSP leadership: "every MSPer achieves at least Level 3 by fall." There's a leaderboard. A prize. DS's pulse survey framework specifically predicts this: self-assessment inflation is concentrated in L1–L3 — the exact levels where gamification produces false signal.

02 — Level Map

Where the ladders align — and where they don't.

BCG's four levels map to roughly DS L1–L5. DS's top two levels — Pioneer and Orchestrator — occupy territory BCG doesn't name, measure, or certify.

BCG AI Fluency Ladder — Firm-Wide
DS AI Fluency Framework
no equivalent
no equivalent
Enterprise Track E1+
AI Builder
Production-grade AI systems in client or BCG environment. Requires L1–L4 + coding assessment.
BCG L4
AI Multiplier
Self-contained AI systems within engagement scope — not integrated into BCG or client tech stacks
BCG L3
AI Automator
Design AI workflows and automations for personal or team productivity
BCG L2
AI Practitioner
Agents, chatbots in cohort — specific workflows for measurable impact
BCG L1
AI Explorer
Foundational knowledge and basic use cases — expected of all BCGers
DS only
Shared
Shaping — L7
Orchestrator
Organisational AI policy. Studio-wide decisions about adoption, governance, and limits. The framework others move through — you helped build it.
Shaping — L6
Pioneer
Frontier practice. Testing unreleased tools. External reference material. Teaching and shaping the field's standards.
Making — L5
Architect
System-level thinking before picking any tool. Modular, tool-swappable design. Agents run without you.
Making — L4
Builder
Workflow-first, then tool selection. Build things others operate. Tool-agnostic: Claude vs ChatGPT, Kling vs Higgsfield.
Making — L3
Composer
Design how you work with AI before starting. Repeatable personal systems. First Claude skills — slash commands, saved approaches.
Using — L2
Experimenter
Iterate with purpose. Reliable approaches across multiple task types. Building instincts, even if you can't always explain why they work.
Using — L1
Explorer
Daily use, accepting outputs. One or two tools, one or two task types. Finding the edges of usefulness.

DS L6 and L7 — Pioneer and Orchestrator — exist in territory BCG doesn't name, measure, or certify.

The DS Advantage
03 — Key Divergences

Four places where the frameworks split.

01
Certification vs Self-Report

BCG tests.
DS trusts.

BCG's ladder is a knowledge assessment — pass it, get a badge, it goes somewhere on your record. DS explicitly designed its framework as the opposite: "if it ever becomes a promotion gate, the rungs become a lie." Both have internal logic. They are philosophically incompatible. DS staff now carry two frameworks with entirely different stakes attached.

02
Gamification vs Honest Measurement

Prizes create inflation.

The MSP competition — leaderboard, office happy hour for top completion — creates exactly the incentive DS's pulse survey was designed to detect and filter for. The DS handover warns explicitly: "self-assessment inflation is concentrated in L1–L3." A race to L3 by August will produce a leaderboard, not a fluency picture.

03
Creative Depth vs Consulting Breadth

BCG's ceiling is DS's middle.

BCG L4 (Multiplier) is the top non-enterprise level. DS has three levels above what BCG counts as advanced. And BCG has no language for taste, craft, voice inside AI output — things that define a designer's work at L3–L4 in DS's model. Tool-specific fluency (Kling vs Higgsfield, Figma Weave, Claude skills) is invisible in BCG's framework entirely.

04
Responsible AI as Layer vs Enduring Skill

One skill vs an architecture.

BCG lists Responsible AI as one of six enduring skills. DS's framework treats it as an intensifying layer from L3 through L7 — growing in weight and consequence at every level. By L7 it becomes organisational AI policy and studio-wide governance. That's not a skill you certify. It's a posture you build over years.

04 — The DS Framework In Full

Seven levels. Three bands.
One intensifying responsibility.

Built for creative practitioners. Each level is descriptive, not evaluative. The Responsible AI layer begins at L3 and deepens at every step.

Using — Daily use, building instincts L1 – L2
L1 — Using
Explorer
You use AI tools every day. You take what AI gives you, tidy it up, and move on.
One or two AI tools, one or two task types. Light prompt shaping — adding context, specifying tone — but largely accepting first or second outputs. You know AI tools are useful. You're still finding the edges of that usefulness.
L2 — Using
Experimenter
You iterate with purpose. You have reliable approaches. You're building instincts, even if you can't always explain why they work.
Try different approaches, notice what shifts, adjust. Personal shortcuts — framing a brief, setting context, sequencing a prompt — that consistently produce better results than a cold ask. Colleagues occasionally ask you for tips.
Making — Systems, workflows, agents L3 – L5
L3 — Making
Composer
You design how you work with AI before you start. Repeatable personal systems. You've begun building, not just using.
Two or three workflows where AI plays a consistent, intentional role. First Claude skills — structured prompts, slash commands, saved approaches. Exploring node-based canvases: Figma Weave, Runway Space. Workflows built for you; not yet designed for others to run.
Responsible AI

Active choices about where AI fits and where it doesn't. Developing judgment about quality, attribution, and when human craft is the point.

L4 — Making
Builder
You design workflows first, then choose AI tools to fit. You build things others operate. You think in inputs, outputs, and handoffs.
Tool-agnostic in practice: Claude vs ChatGPT, Kling vs Higgsfield for camera control. Node-based pipelines. At least one agent built — a system that runs without you executing each step. L1–L2 people use what you create.
Responsible AI

Deciding what goes into shared AI systems — data, models, outputs. Thinking about downstream consequences. Guardrails are your responsibility now.

L5 — Making
Architect
You build AI systems others run without you. Tool-swappable by design. When a better tool comes along, you update a node — not the whole structure.
System-level thinking before picking any tool. Modular workflows: logic separated from AI tooling so a better model can replace an older one. End-to-end pipelines where the human role is direction and judgment, not AI execution.
Responsible AI

Actively designing for where AI shouldn't be. What stays human, what requires sign-off. Restraint is architecture, not instinct.

Shaping — Field-defining. BCG has no equivalent. L6 – L7
L6 — Shaping
Pioneer
You operate at the frontier. Testing what others haven't released yet. Your work becomes reference material — inside BCG Design Studios and beyond.
Track AI model releases before announcement. Run early access and beta programmes. Share findings through writing, teaching, workshops, public frameworks. People outside BCG Design Studios reference your thinking on AI practice.
Responsible AI

Forming and sharing positions on what the AI field should and shouldn't do. Teaching at the frontier carries weight. Your voice shapes the field's standards.

L7 — Shaping
Orchestrator
You design how the organisation learns, adapts, and decides about AI. You determine where AI goes and where it doesn't.
Not the best individual AI practitioner — the person who designs conditions for others to improve their fluency. Structural decisions about AI capability across BCG Design Studios: which tools adopted, what governed, what trained for, what is off-limits and why.
Responsible AI

Organisational AI policy. Studio-wide decisions about adoption, governance, and limits. As focused on where DS draws lines as on what DS builds.

05 — Strategic Brief

DS doesn't need to compete with the ladder. It needs to own the floor above it.

01
Make BCG L1 the DS baseline expectation

BCG certification is now the floor, not the ceiling. DS should communicate clearly: L1 is expected of everyone; L2 is the studio expectation within six months. Encourage completion without making it a development gate. DS's framework sits above and alongside it — and goes much further.

02
Name the DS-exclusive territory

Pioneer and Orchestrator are the levels BCG doesn't map. These are where DS practitioners and leaders have the most to offer the firm — and the most to lose if the framework gets absorbed into a model that doesn't see them. Make this visible. Give those levels language the firm can recognise even when its ladder doesn't reach them.

03
Run the pulse survey before the competition corrupts the data

The MSP leaderboard will inflate self-assessments across the firm. DS's pulse survey — designed specifically to get honest data — should run before August, while responses aren't shaped by competitive incentives. After the leaderboard closes, everyone at BCG will have a reason to claim L3. That signal will be noise.