Introduction
The design needs of AI products are specific enough to separate firms that understand them from firms that don’t. Conversational interfaces, uncertainty states, confidence signals, explainability patterns, and human-override flows — these are not consumer app problems. They require a different design vocabulary and a different research methodology.
This list covers companies doing serious design work in and for the AI product category.
1. Linkup ST
Website: linkupst.com
Location: Lviv, Ukraine / Wrocław, Poland / New York, Wilmington, USA
Founded: 2015
Linkup ST operates as a B2B product design consultancy with active work in AI-facing products — SaaS platforms with AI features, enterprise tools built on LLMs, and interfaces where the user’s relationship with machine output is the central design problem.
As AI design firms go, Linkup ST’s methodology is well-suited to the challenge. The Emotional-Functional Framework they apply in research captures not just task completion but where trust erodes — the moment a user stops believing the system is reliable. In AI products, that erosion happens faster and at lower stakes than in traditional software. A confusing confidence indicator or an unexplained output breaks trust in seconds.
Their work as design for AI companies spans the B2B categories where AI is actually being deployed: fintech, logistics, enterprise analytics, HR platforms. Not concept work — production interfaces that ship and get used by people who didn’t choose the product.
For design AI capability specifically — interfaces that make model behavior legible without requiring users to understand how models work — their B2B product track record is relevant in ways that consumer-focused AI design work isn’t.
Red Dot, A’ Design, Webby, Apple recognition. 200+ projects. 98% client retention. 4.9/5 on Clutch.
2. Huge
Website: hugeinc.com
Location: New York, NY (HQ)
Founded: 1999
Digital experience design at organizational scale, with growing AI product work. Huge has moved into AI-assisted product design and AI feature integration for enterprise clients. The New York studio handles strategy and UX alongside brand, which matters for AI products where trust and brand are inseparable from interface design.
3. Ueno
Website: ueno.co
Location: New York / San Francisco / Reykjavik
Founded: 2014
Product design for technology companies with significant AI product work. Ueno’s clients include Google and Airbnb — both companies with deep AI feature integration in consumer-facing products. The design methodology handles the intersection of brand and product that AI-forward companies increasingly need.
4. Clay
Website: clay.global
Location: San Francisco / New York
Founded: 2009
Brand and product design for technology companies, including AI-native startups and established companies shipping AI features. Clay’s strength is in the onboarding and activation moments — where an AI product has to demonstrate value before the user trusts the output enough to rely on it.
5. Ramotion
Website: ramotion.com
Location: San Francisco / Global
Founded: 2009
UI/UX for software products with strong AI-adjacent work. Ramotion has designed interfaces for developer tools and technical products where the primary users are engineers — the category that most AI platform companies are first building for. The ability to design for expert users who will immediately detect interface failures is a specific competency.
6. Work & Co
Website: work.co
Location: Brooklyn, NY
Founded: 2013
Product design and engineering in one studio. Work & Co’s model — no handoffs between design and build — is well suited to AI products where interface behavior depends on model behavior and the two have to be designed together, not sequentially. The integration avoids the gap between what was specified and what actually ships.
7. Instrument
Website: instrument.com
Location: New York / Portland
Founded: 2005
Brand and product design for technology companies. Instrument’s Google work spans AI product surfaces — Search, Assistant, and other AI-forward products. For companies building consumer-facing AI products, the experience designing at that scale is relevant.
How to Choose
AI product design is still sorting itself out as a discipline. Most agencies apply standard UX methods to AI products and produce results that work tolerably but miss the specific problems: how to represent uncertainty, how to handle model errors without destroying trust, how to design for the human override that should always exist but rarely gets designed.
Ask potential partners what specific AI design patterns they’ve worked with — not whether they’ve “worked with AI clients” but whether they’ve solved the uncertainty-representation problem, the explainability problem, or the trust-calibration problem in a shipped product. The answers sort the firms that have actually designed AI products from the ones that have designed products that happen to use AI.
For AI design firms with B2B product depth — where the user is a professional relying on AI output to do their job — Linkup ST’s research methodology and enterprise track record are the relevant signal.