Design Studios in the Age of AI

For a generation, the design studio was how serious companies bought taste. You did not hire “a designer.” You hired IDEO, or Pentagram, or frog, or Wolff Olins. The invoice was a strategy deck, a product, a brand system, and a room full of people who could explain why the corner radius mattered.

Then three things happened in overlapping waves: consultancies bought the studios, corporations hired the designers in-house, and generative AI made a first draft free. The names are still on the door. The business under them is not the same business.

This is a map of that twenty-year arc — with the figures that are public, and the ones that are reported, labeled as such.

The twenty-year boom, in four moves

2005–2012: the product becomes the brand. Apple’s second act made “design” a CEO word. Studios that could ship objects and interfaces — IDEO (the Apple mouse, in the origin myth), frog (Sony, Lufthansa, the industrial-design century), Lunar in Silicon Valley — sold more than identity. They sold the thing.

2013–2018: everyone buys a studio. Accenture acquired Fjord in May 2013. McKinsey bought Lunar (2015) and Veryday (2016). WPP already owned Landor; Omnicom already owned Wolff Olins. Flextronics had bought frog in 2004; the firm later passed through Aricent and Altran into Capgemini Invent (integrated 2021, alongside Cambridge Consultants, into a 10,000-person innovation bench). In 2016, Hakuhodo DY’s Kyu collective acquired IDEO. The independent consultancy became a capability inside a much larger invoice.

2012–2019: the client staffs up. IBM is the blunt instrument. Phil Gilbert’s program took the company from about 375 designers in 2012 to 1,100 by 2015 and 1,600 by 2017, across 44 studios, with tens of thousands of employees badged in “design thinking.” Designer-to-engineer went from roughly 1:72 to 1:8. Airbnb, Google, Capital One, and a dozen banks ran smaller versions of the same play. You no longer had to call Palo Alto to run a workshop.

2018: McKinsey puts a number on the religion. The Business Value of Design tracked 300 public companies. Top-quartile scorers on the McKinsey Design Index grew revenue 32 percentage points faster than industry peers over five years, and total shareholder return 56 points faster. For a decade that slide justified the retainers.

Then the workshops were a commodity, the in-house teams existed, the holding companies had absorbed the brand names — and a prompt could produce a first draft.

IDEO: the design-thinking era, billed and then copied

IDEO is the firm the MBA case studies were about: David Kelley, the Stanford d.school, “design thinking,” the shopping cart, the mouse, a methodology that escaped the studio and became a corporate training product (IDEO U, The Field Guide to Human-Centered Design).

It also became a victim of its own success. Fast Company’s October 2023 account is the public record of the contraction. CEO Derek Robson told the company there would be layoffs. IDEO later confirmed to Fast Company it was eliminating 32% of headcount that year — all levels. Offices in Munich and Tokyo closed; footprints shrank in the U.S., London, and Shanghai. Headcount had already fallen from about 725 in 2020 to roughly 500 by the all-hands. A former employee told the magazine revenue had gone from about $300 million to $100 million over four years. That last figure is reported, not an audited filing. Treat it as a signal of magnitude.

The causes were not a single villain:

• Clients built the capability IDEO had taught them.
• “Design thinking” was copied by every consultancy on earth, then mocked for workshops that did not ship.
• Kyu ownership (2016) sat the firm inside a Japanese holding structure just as Western tech budgets tightened.
• Climate and DEI work, however sincere, was rumored internally at a few million dollars — not a replacement for Fortune 500 product retainers.
• Fees sat near McKinsey altitude without McKinsey’s ongoing implementation annuity.

By 2025 the firm had another CEO, Mike Peng, talking about teaching clients to innovate without a standing IDEO team — which is another way of saying the old billable hour is not coming back at 2016 rates. Peng, to Fast Company-adjacent coverage: customer-centricity is “table stakes now.”

Pentagram: the structure that did not sell

Pentagram is the counter-example, and it is not subtle about it. Founded in London on 12 June 1972 by Theo Crosby, Alan Fletcher, Colin Forbes, Mervyn Kurlansky, and Kenneth Grange. Unique among large studios: the owners are the designers, and they are the primary client contact. The 2023 two-volume Pentagram at 50 was a victory lap. The live about page still says 24 partners, all practicing, working “in friendship.”

Paula Scher is still a partner. Michael Bierut — Mastercard, Hillary’s H, the Yale and Saks work — stepped back from partner in October 2024 to an advisory “strategist at-large” retainer, going into the Flatiron office once a week (Fast Company). That is succession, not a fire sale.

Pentagram never sold to WPP. It never became a “design thinking” training company. It sells identities, type, environments, packaging, the Public Theater posters Scher has done for decades. AI shows up in the work the way a new printing process would: a tool inside a partner’s studio, not the product. When a Pentagram AI-assisted public project draws online scorn, that is the profession arguing about craft. It is not a 32% layoff memo.

They do not publish revenue. Ignore the scraped “$25 million” figures that float around vendor databases. The unit that matters is the partner. Twenty-four people who can still get a CEO in the room is a different organism from a 700-person methodology firm.

The rest of the famous names — who owns them now

A useful way to read the last twenty years is not “which studio is best” but who signs the paycheck.

What the boom years were worth — and what broke

McKinsey’s 2018 index was not wrong, exactly. Companies that treated design as an operating system outperformed. The mistake was thinking the studio would capture that surplus forever.

Once IBM has 1,600 designers, the surplus stays inside IBM. Once Accenture owns Fjord, the surplus stays on Accenture’s digital slide. Once every product manager has run a sticky-note workshop, “design thinking” is a line in onboarding, not a reason to fly six people to Chicago.

Then AI hit the remaining billable work that looked like production: decks, first-pass identities, research summaries, UI variants, the junior layer of every studio. We covered the template and generation numbers in What AI did to design templates. For studios the equivalent is: the thing you used to staff a team of eight for can be staffed with two people and a model — until the work has to be a system, not a picture.

Fast Company, around the IDEO cuts, reported product-design job postings down about 18% and graphic-design postings down about 57% in that downturn year. U.S. BLS still projects graphic-designer employment at 2% growth through 2034, slower than the all-occupation average, and names generative AI as a demand constraint. That is not “design is dead.” It is “the middle of the studio pyramid is thinner.”

The specialized-design-services market is still large and still growing — The Business Research Company puts it at $166.7 billion in 2025 and $177.1 billion in 2026 (~6% CAGR). Architecture giants like Gensler sit in a different, construction-tied bucket. The pain is not “there is no design spend.” The pain is who collects it: in-house teams, Accenture Song, Capgemini Invent, WPP brand arms — and a thinner set of independents who still sell a signature.

Where some of them are now

Pentagram — still the world’s largest independent design consultancy by its own billing. Partners in London, New York, Austin, Berlin. The work is identity and culture, not a 12-week innovation sprint. Succession is happening in public (Bierut) without a private-equity tombstone.

IDEO — smaller, Kyu-owned, arguing for a teaching/enabling model. The d.school is still a pilgrimage site. The 700-person global studio that charged like McKinsey is not coming back in that shape.

frog — a Capgemini Invent brand. Access to 35+ studios and a huge delivery machine. The word “frog” on a case study now means a node in a 270,000-person company, not a Palo Alto loft.

Fjord / Accenture Song — design as a chapter in a digital transformation. Scale that no 24-partner firm will match. Homogeneity that no 24-partner firm would want.

Landor, Wolff Olins, Interbrand — the brand consultancies inside holding companies. Still making identities. Competing with in-house brand studios and with Canva for the work that used to be a junior pitch.

Collins, CGH, a long tail of 8–40 person shops — the surviving independent pattern: small enough that the principal still touches the work, famous enough that the phone still rings for the thing AI cannot fake on a letterhead.

The in-house orgs — IBM, Airbnb, the banks. They hired through the boom, laid off with tech in 2022–24, and kept the design system. They are both the studios’ alumni network and their substitute.

What a studio is for, after the prompt

A Pentagram identity still has to be a system: type, color, applications, a twenty-year public. An IDEO-style product still has to survive manufacturing and a nasty user. frog still has to make a device someone will hold. None of that is a Midjourney poster.

The work that vanished is the work that was always a picture of a process: the 80-page “opportunity landscape,” the first-pass logo board, the deck that restated the brief in nicer type. AI ate the picture. It did not eat the system.

That is the same split we keep finding in templates. How DesignForever templates differ is the file-format version: a working layout versus a JPEG of competence. Studios that still matter sell the first. Studios that sold the second — workshops, first drafts, “innovation theater” — are the ones whose headcount charts go orange.

Taste did not die in 2023. The retainer for simulating taste did.

Sources

• Fast Company, Mark Wilson, Oct 2023 — IDEO 32% cut, Munich/Tokyo closures, ~725 (2020) → ~500, former-employee revenue comment
• Pentagram, About — 24 partners, independent partnership; Bierut partner-at-large from Oct 2024
• Fast Company, Lilly Smith — Bierut steps back from partner
• Fortune / IBM Design (2017) — ~375 (2012) → 1,600 designers (2017), 1:72 → 1:8
• McKinsey, The Business Value of Design, 25 Oct 2018 — +32 pp revenue, +56 pp TRS, top-quartile MDI, n=300
• Accenture newsroom, 7 May 2013 — Fjord acquisition
• Capgemini Invent, 23 Jun 2021 — frog + Cambridge Consultants integration
• Wired, May 2013 — “Big Corporations Are Buying Design Firms in Droves”
• The Business Research Company — specialized design services $166.72B (2025) → $177.1B (2026)
• U.S. BLS Occupational Outlook Handbook — graphic designers, 2% employment outlook 2024–2034
• PRINT Magazine / Pentagram at 50 (2023) — founding date 12 June 1972, structure