ANML is an AI product design agency based in San Jose, California. Since 2012 we have designed AI features and AI-company brands, including AI meal plans and ingredient recognition for Yummly, used by more than 50 million people; the Umi AI wellness coach for PanasonicWell; and the brand and website for Enfabrica, the AI networking company whose team and technology Nvidia brought in for more than $900 million.
AI only works when people understand it. We partner with teams to turn advanced models into clear, human centered products that people can trust, use, and rely on in real contexts.
UX and UI design for AI and machine learning products
AI-native product design, where the model is the product
AI features added to existing products
Data visualization and interaction design
Concept modeling, prototyping, and user validation
Interfaces for AI agents, copilots, and assistants
Patterns for communicating confidence, uncertainty, and system limits
Brand and product integration for emerging technologies
We translate technical complexity into clarity. Our team works closely with product, data science, and engineering to understand model behavior, data inputs, constraints, risk surfaces, and trust requirements. That context lets us design flows and interfaces that are honest about how the system works and what it can and cannot do. Every decision is grounded in real use cases, so people can interpret outputs, act with confidence, and recover when things are uncertain or wrong.
Thoughtful AI design builds trust.
When people understand what the system is doing, why it produced a result, and how to influence it, they are more likely to rely on it in meaningful ways.
By balancing logic with empathy, we create interfaces that help people see what is happening, why it matters, and how to take the next step with confidence.
When the model is the product, we design around what it can and cannot do from the first concept. When an established product adds AI, we fit the new capability into flows people already trust.
For Yummly, we designed AI meal plans and AI ingredient recognition inside an app used by more than 50 million people. For PanasonicWell, we designed Umi, an AI wellness coach.
Our designers work in code with Claude Code alongside Figma. Claude Code reads our Figma files and design system, and a shared project knowledge base keeps the strategy, requirements and decisions in context. We deploy functional prototypes that clients review in a real browser, and the prototype becomes the spec.
On our last static-design project, 1 in 4 QA tickets existed only to make the build match the design. On the Salesloft rebrand, where we designed in code, that dropped to 1 in 9. Read how we design with Claude Code.
Examples of the outcomes we focus on:
AI platform dashboards that make decision logic and system status clear at a glance
Data driven internal tools that help teams explore, validate, and act on insights
Assistant and copilot experiences that feel reliable, controllable, and explainable
Monitoring and operations views that reveal model performance and system health
AI products have to account for uncertainty, transparency, and trust.
We design interfaces that explain system behavior clearly, help people interpret outputs, and make intelligent features feel intuitive in everyday use.
We create patterns that surface confidence levels, inputs, and reasoning in ways real users can understand.
That clarity supports better decisions and stronger adoption.
Yes. We often begin while models are still evolving.
We help shape the product experience early so teams build with user clarity, constraints, and value cases in mind.
Yes. We specialize in visualizing system behavior, uncertainty, and data flows so people can explore, compare, and act without needing to be experts.
Always. Deep technical discovery is part of our process and not an add-on.
It ensures the experience aligns with model behavior, limitations, and edge cases.
Yummly (2012 to 2023): ANML was Yummly's core design team through its acquisition by Whirlpool, designing AI-generated meal plans and AI ingredient recognition as the platform grew to more than 50 million users.
PanasonicWell: ANML extended the brand system and supported UX and go-to-market for Umi, an AI wellness coach for families.
Enfabrica: ANML designed the website, launch campaign, and product configurator for the AI networking company whose team and technology Nvidia later brought in for more than $900 million.
Cognita: ANML designed the web presence for an AI radiology company.
Yes, in two ways. ANML is an AI product design agency that designs AI-native products and AI features inside existing products, including AI meal plans and ingredient recognition for Yummly and the Umi AI wellness coach for PanasonicWell. AI is also part of how we work: our designers use Claude Code to turn design work into functional prototypes in code that clients review in the browser.
Yes. Our designers use Claude Code, connected to our Figma files and design system, to build functional prototypes in code, with a shared project knowledge base that keeps project context in every session. On the Salesloft rebrand, the prototype became the spec, and QA tickets that existed only to make the build match the design fell from 1 in 4 on our last static-design project to 1 in 9.
Yes. ANML was founded in San Jose, California in 2012 and works from Silicon Valley with US and international teams.
Last updated November 2026