Fifty-seven years of history running from the Sony Trinitron to today's service design work, and a scale almost nobody else can match, with more than 2,000 people across 35 studios. frog sits inside Capgemini Invent, which means it arrives attached to systems integration, data and technology delivery. For a large organisation trying to change a customer journey that touches six business units and four legacy systems, that combination is the point. Fahrenheit 212, Idean and June21 were absorbed into the brand, and Doing joined in Italy in 2026, bringing the Milan teams to around 300 people alongside a new design centre of excellence.
For enterprise • 10 firms
The 10 Best Enterprise Product Design Agencies
At enterprise scale the design question is rarely whether a firm can do the work. It is whether the work survives your organisation. Most large programmes fail on adoption, governance and staffing continuity rather than craft, and those are structural problems that come down to how the agency is owned and how it staffs.
Every entry below states who owns the firm, because at this level that single fact predicts more about your engagement than the portfolio does.
Start with ownership
Best forScale, systems integration, global delivery and one accountable vendor — with consultancy rates, months-long procurement and rotating staff. Right when the problem spans technology and design.
Real scale and capability with a parent that has its own targets. Ask how your account is prioritised against larger ones — and get the answer in the contract.
Senior people stay on the work, contracting is faster, and the firm can decline a bad-fit brief. The limit is capacity — none can staff 200 people next quarter.
Unusually stable senior staff — the people delivering own the business — with multi-market coverage and no holding-company politics.
The Complete Rankings
Full agency profilesDifferent proposition from a transformation consultancy. Work & Co builds the thing customers actually touch, and has done it for IKEA as agency of record, for Apple, Google, Pfizer, Mercedes and the PGA Tour. Around 425 people across eight offices in the United States, Europe and Latin America. Accenture completed the acquisition in January 2024 and the team now operates inside Accenture Song, which brings global reach and a consultancy's contracting process. The craft reputation that made the studio's name predates the acquisition, so ask how staffing works now rather than assuming it is unchanged.
Founded in Brooklyn in 1999 with IKEA as its first client, and for two decades one of the largest digital agencies anywhere, now around 1,000 to 1,500 people. Interpublic sold the business to AEA Investors in December 2024 and combined it with Hero Digital, and Josh Campo became chief executive in January 2026. Private equity ownership at this scale cuts both ways, bringing investment discipline and margin pressure in the same package. The CX capability and the global footprint are intact. What is still settling is the shape of the combined offer, so put continuity commitments in the contract rather than the relationship.
04
Designit
designit.com ↗
Best for
Service design across physical and digital touchpoints in a complex organisation.
Founded in Aarhus in 1991 and still carrying the Scandinavian research heritage that came with it, which shows up as an unusually strong ethnographic and service design practice rather than screen craft alone. Wipro acquired the firm in 2015 and it now operates as Wipro's experience innovation arm with more than 500 designers, researchers, strategists and technologists across fourteen locations. Cooper, the firm that helped invent interaction design as a discipline, was folded in during 2017. The recurring enterprise engagement is building internal design capability, including design labs inside Fortune 500 organisations, which is a different sale from delivering a product.
05
Clay Global
clay.global ↗
Best for
Enterprise-quality work without an enterprise programme wrapped around it.
The useful thing here is what is missing. Clay works with Meta, Google, Amazon, Sony, Toyota, Samsung, UPS, ADP and VMware, but it does it with senior remote teams led by the co-founders rather than a hundred-person programme and a governance layer. For an enterprise team that has a defined problem, a real budget and no appetite for a six-month procurement cycle, that is a genuinely different option from everything above it on this list. Brand and product are handled together, which matters for large organisations where the two have drifted apart across business units. Sector depth sits in B2B, SaaS, AI and fintech.
Large organisations tend to be good at launching things and bad at getting people to come back, and this studio's entire proposition addresses the second problem. Game design principles applied to software that is not a game, aimed at retention rather than task completion. Ford, Qantas, American Express and Jaguar Land Rover are all enterprise clients with genuine consumer-facing products. Employee-owned and a certified B Corp, running since 2004 across six studios, which gives multi-market coverage without a holding company deciding where the senior people go.
07
Fantasy
fantasy.co ↗
Best for
Flagship interfaces at global brands where the design has to set a category benchmark.
Independently owned since 1999, which at this tier of the market is close to unique, and it changes the economics. No holding company overhead means senior people stay on the project rather than moving to whichever account is under pressure. The work is cinematic and high-polish, with Netflix, Spotify, Samsung, Ford, Disney and Microsoft on the roster, and the studio has moved firmly into AI-native product design including generative platforms and AI-driven operating systems. Two US studios, San Francisco and New York.
Sanofi's Elements design system went from nothing to delivered in ten months across a global pharmaceutical brand, which is the specific credential that matters, because component libraries are easy and governed systems adopted by multiple teams are not. Toronto since 2013 with a second office in Amsterdam, and a client list including Nike, Google, Staples, UNIQLO, The North Face and Kenvue. The firm restructured in 2026 around agentic development, running one to four senior engineers per engagement supported by AI agents, so delivery is faster and considerably more engineering-led than it was.
09
Elsewhen
elsewhen.com ↗
Best for
Regulated enterprises where the interface has to satisfy legal before it satisfies users.
London, running since 2011, now positioned as an AI consultancy with strategy, design, transformation and engineering under one roof. The reason to shortlist them is that in regulated markets the design constraint is rarely aesthetic. It is the collision between what the regulator mandates, what the legacy platform allows and what a user can actually complete, and resolving that needs consulting capability. Engagements run in defined blocks rather than open-ended programmes, which is unusual at this end of the market and useful when you need a decision on a date.
10
Instrument
instrument.com ↗
Best for
Brand and product moving together at consumer technology scale.
Portland and Brooklyn since 2006, with a long Google relationship behind much of the work people recognise. Teams are structured so designers who code and developers who design sit together, which removes the handoffs where enterprise projects usually lose their edge. The company runs an all-women C-suite. Instrument belongs to Stagwell's Code and Theory network, and for an enterprise buyer that cuts both ways, bringing network scale and network priorities that are not yours. Worth asking directly how your account is resourced and what happens if a larger network client needs the same people.
Compare the ten
Facts, not scoresNo rankings by number here — at this scale ownership predicts more than craft. Compare who owns the firm, how it delivers, and where it operates.
| Firm | Best for | Ownership | Scale & base |
|---|---|---|---|
 frog | Transformation programmes | Capgemini Invent | 2,000+, 35 studios |
 Work & Co | Flagship consumer apps | Accenture Song | ~425, 8 offices |
 Huge | Multi-market CX programmes | AEA Investors (PE) | ~1,000–1,500, global |
| Designit | Service design & capability | Wipro | 500+, 14 locations |
| Clay Global | Enterprise work, no programme | Independent | Remote-first, SF HQ |
 ustwo | Consumer repeat use | Employee-owned | 6 studios |
| Fantasy | Category-benchmark flagships | Independent since 1999 | SF + NY |
 Rangle | Governed design systems | Independent | Toronto + Amsterdam |
| Elsewhen | Regulated enterprises | Independent | London |
| Instrument | Brand + product together | Stagwell (Code & Theory) | Portland, Brooklyn |
Consultancy, network, or independent
The decisionAt enterprise scale this is the decision that shapes everything else, and it is worth making deliberately rather than discovering it in month four.
Consultancy-owned
frog inside Capgemini Invent, Work & Co inside Accenture Song, Designit inside Wipro. You get scale, systems integration, global delivery and the ability to hand a programme to one accountable vendor. You also get consultancy rates, a procurement process measured in months, and a staffing model where people rotate. Right when the problem genuinely spans technology and design and you need one throat to choke.
Network or private equity owned
Huge under AEA Investors, Instrument inside Stagwell's Code and Theory network. Real scale and real capability, with a parent that has its own targets. The question to ask is how your account is prioritised against larger ones in the same portfolio, and to get the answer in the contract.
Independent
Fantasy, Clay, Rangle and Elsewhen. Senior people stay on the work because nobody upstairs is reallocating them, contracting is faster, and the firm can decline work that does not fit. The limit is capacity. None of them can put two hundred people on a programme next quarter, and pretending otherwise would be the failure mode.
Employee-owned
ustwo is the only one here. Unusually stable senior staff, since the people delivering own the business, and multi-market coverage without holding company politics.
How enterprise design programmes actually fail
Failure modesRarely on craft. Almost always on one of these.
The pitch team was not the delivery team
You met four excellent people and got two juniors and a rotating cast. This is the single most common complaint at this tier and it is entirely preventable with named individuals and continuity clauses in the contract.
Nobody planned for adoption
A design system shipped with no governance, no advocate inside the organisation and no budget for the second year is a very expensive PDF. Adoption is a change management problem, and it needs an owner on your side before the work starts.
The pilot never scaled
A brilliant proof of concept in one business unit dies because the other six were never consulted and have their own roadmaps. Involve them early even though it slows the first phase down.
The agency never learned your domain
A month of teaching a vendor your business is a month you paid for at consultancy rates. Weight domain experience heavily, and ask candidates how long onboarding takes and who bears the cost.
What to insist on in the contract
Before you signNamed individuals with seniority levels, and what happens when one leaves. Full handover of source files, design tokens, components and documentation in a form your engineers can use without the agency present. A defined governance model for anything system-shaped. Clear IP ownership. And a measurable outcome that is not a deliverable, because a programme judged on whether artefacts arrived will produce artefacts.
Frequently asked questions
FAQWhat does an enterprise design engagement cost?
A defined product engagement with a specialist firm runs $250,000 to $1 million. Multi-year transformation programmes with a consultancy-owned firm reach several million and are usually priced as part of a wider technology engagement. Design systems for large organisations run $150,000 to $500,000 depending on governance scope. Independent studios sit meaningfully below consultancy rates for comparable design work.
Should we hire a design firm or the design arm of our systems integrator?
If the problem genuinely spans technology and experience, and you need one vendor accountable for both, the integrator's design arm is the pragmatic answer. If the problem is design quality and your technology delivery is already handled, an independent firm will usually produce better work faster and for less. The mistake is buying a transformation programme when you needed a product team.
How long does enterprise procurement usually take?
Three to six months is normal for a consultancy-owned firm, including security review, vendor onboarding and legal. Independents can often start in four to six weeks. If your deadline is fixed, factor this in before shortlisting, because procurement time has killed more enterprise design projects than budget has.
Will an agency work inside our existing tooling and process?
Ask directly, because the answers vary more than you would expect. Some firms need to run their own sprint cadence and rituals and will struggle inside yours. At enterprise scale your process is not moving, so a mismatch surfaces in week three and costs you a quarter.
How do we handle security and compliance review?
Start it in parallel with the commercial conversation rather than after. Ask candidates for their existing certifications, their data handling model, and whether they have passed review at organisations of your size in your sector. Firms that work regularly with regulated clients will have this packaged. Smaller studios often will not, which is a real cost rather than a dealbreaker.
Can we run a pilot before committing to a programme?
Yes, and you should, but scope it as a genuine test with a decision at the end rather than a discount on phase one. Define in advance what result would make you continue and what would make you stop. Pilots that cannot fail teach you nothing.
What happens to the work when the engagement ends?
Whatever you negotiated at the start, which is why it belongs in the contract. At minimum you want source files, a documented design system your team can extend, and a handover period with the actual delivery team rather than an account manager. Ask what post-launch support is included and whether there is a stated warranty period, since several firms at this tier offer neither by default.
Does agency ownership really change the work?
It changes who is on your account, how quickly they can start, what they charge, and whether they can turn down a brief that does not suit them. It does not reliably change craft quality, which is why the ownership question belongs in your evaluation rather than at the top of it.