Contact usHire us

[Agentic AI UX design]

AI Fit: Design AI that keeps users in

AI Fit is our 10-day strategic session to help you work out where AI belongs in your product and make sure people actually use it.

Your users don’t really care about AI…

…no matter how many tech evangelists tweet about it. They care if something in your product got easier, faster, or clearer. And if AI doesn’t do that, well, your adoption charts will say so.

  • “I don’t have a reason to use this”

    The AI does something users already find easy or solves a problem they don’t have. They try it once and move on.

  • “That’s not how I work”

    The feature sits in its own tab or panel, so when users are actually in the task it could help with, they never think to open it.

  • “I’m not sure what it can do”

    The empty state says “Ask me anything.” Users try once, get a mediocre answer, and don’t come back.

[Where agentic AI UX design breaks down]

Why do AI features fail at UX?

Every AI feature has to answer four questions. When it can’t, users leave. Behind all four, there’s a bigger question: does AI belong here at all?

  • AI made a recommendation. Why?

    Users see the output, but nothing tells them how the AI got there.

  • The user wants to adjust AI’s output. Can they?

    The interface doesn’t put the user in control. They watch AI make decisions.

  • AI returned an answer. Is it correct?

    The output shows up without sources or confidence signals.

  • AI got it wrong. Now what?

    If AI gets it wrong, users hit a dead end with no way to recover.

More on why users never try AI features twice

[Our approach]

How AI Fit works

Two layers of work for one outcome: an AI feature your users reach for.

  • First, we work out where AI belongs

    Not every feature should be AI. Before we design anything, we map where AI could genuinely improve your product.

    • AX opportunity map: where AI would help your users
    • Design feasibility per flow: how to design UX around that feature
    • Visual concepts: 1–3 main AX flows
    • 3-month AI product roadmap with priorities
  • Then, we pressure-test that feature against four things

    We design the feature against four dimensions from what we call the AI Feature Design Fit framework.

    • Explainability
    • Control
    • Trust signals
    • Failure design

[AI Feature Design Fit framework]

How we design AI features users love

We check each AI feature against our AI Feature Design Fit framework. It has four dimensions.

  1. 01

    Explainability — Can users understand AI’s actions?

    With no context, most users either accept AI’s answer blindly or ignore it, usually the latter. The fix: show what shaped the answer — the factors, steps, reasoning.

  2. 02

    Control — Do users stay in the loop?

    When the AI completes an action without a way back, users may lose trust in the whole product. The fix: let users review, edit, or undo anything the AI does that has consequences.

  3. 03

    Trust signals — How can users tell a confident answer from a shaky one?

    Without small cues like sources, confidence indicators, and consistent behavior, every answer feels like a guess. The fix: show the evidence behind AI’s answer.

  4. 04

    Failure design — When the AI is wrong, what do users do next?

    The AI will be off sometimes, maybe often. In most products, that means the user hits a dead end. The fix: give the user a next step — try again, edit the input, or get a human.

Read the framework deep-dive

[The Habitat AI design system]

We’ll prototype your AI feature in 10 days. How?

When we design your AI feature, we adapt patterns from our own Design System, built with Agentic Layout at its core. That’s what gives us both speed and quality.

  • Senior designer runs the process

  • AX patterns already tested

  • Delivery in 10 days

  • Figma and code in sync

[Case in point]

10 minutes to prototype AI search for a B2B travel platform

What happened
Travel Point had an AI agent named Lucy and wanted to build smarter search down the line. In 10 minutes, we adapted the agent components from our Design System to their product. By the next call, they were looking at a live demo of their AI search.
“This is a great direction for search. I really want to see where this goes.”
Yevhen, CEO · Travel Point

Shipped AI features that aren’t landing? Planning AI features and want to get it right?

  • Already shipped

    We find where UX broke down and why users are bouncing.

  • Still planning

    We analyse where AI actually belongs in your product.

[AI Feature Blueprint]

Design AI users love. Blueprint starts at $4,900

Start with the 10-Day AI Feature Blueprint. We help you figure out where AI belongs in your product and how to design it so users stay. You leave with visual concepts and a 3-month roadmap your team can build from.

$4,900

  • 10 days
  • 1 fixed price
  • 3-month AI roadmap
  • Take it in-house or keep going with us
  • AX opportunity map by Day 6

    Where AI would matter to your users, and where it wouldn’t. Ranked by impact.

  • Visual concepts for 1–3 AX flows by Day 8

    The most important AI flows designed in enough detail to greenlight or rethink.

  • 3-month AI product roadmap by Day 10

    What to build first, and in what order.

  • Ready to build by Day 10

    Take the roadmap in-house, or continue with the Habitat team.

From the founders who worked with us

Clutch52 reviews

“They combined original ideas with a strong work ethic and consistently high standards.”

Dmytro Babat, Director

PlusOne Moving

“We are extremely lucky to have worked with them and would highly recommend them to anybody.”

Thorben Scheidegger, Co-Founder

FinTech Company

“Their designs connected perfectly with what we wanted, and the collaboration was very pleasant.”

Lucas Bonavita, Product Manager

Quicargo

“We’re getting everything we want and more.”

Seif Salama, Senior Product Marketing Manager

AngelList Venture

“They really care about the client’s needs.”

Peng Cheng, Founder & CEO

PauseAble

“Habitat Creative Company was smooth and had good deliverables compared to other agencies. They took feedback early and ensured the designs were translatable to actual development work.”

Kara Amantoy, Co-Founder & COO

Pickle

“The team was a great group of communicators and clear thinkers who were capable of empathy.”

Aboudy El Farkh, Founder

Aura Labs

“The team is dedicated to the project and focused on their client’s success.”

Greg Short, Co-Founder

Feeds Inc.

“They delivered the project on time and made sure my needs were fully satisfied.”

Tzur Rupin, CMO

Dots

“We saw a 340% increase in website traffic, 2.7x more time on site, and a 38% drop in bounce rate.”

Idan Benishu, Co-Founder & CEO

Bagel AI

“The designs they built were extremely tasteful and contemporary. They were fast in their work and communication.”

Jesse Johnson, Founder & Director

STORE 5a

“Habitat had a multi-step process for delivering business value but were flexible and considerate of our timelines.”

Mike Faber, CEO

Vendo

“They combined original ideas with a strong work ethic and consistently high standards.”

Dmytro Babat, Director

PlusOne Moving

“We are extremely lucky to have worked with them and would highly recommend them to anybody.”

Thorben Scheidegger, Co-Founder

FinTech Company

“Their designs connected perfectly with what we wanted, and the collaboration was very pleasant.”

Lucas Bonavita, Product Manager

Quicargo

“We’re getting everything we want and more.”

Seif Salama, Senior Product Marketing Manager

AngelList Venture

“They really care about the client’s needs.”

Peng Cheng, Founder & CEO

PauseAble

“Habitat Creative Company was smooth and had good deliverables compared to other agencies. They took feedback early and ensured the designs were translatable to actual development work.”

Kara Amantoy, Co-Founder & COO

Pickle

“The team was a great group of communicators and clear thinkers who were capable of empathy.”

Aboudy El Farkh, Founder

Aura Labs

“The team is dedicated to the project and focused on their client’s success.”

Greg Short, Co-Founder

Feeds Inc.

“They delivered the project on time and made sure my needs were fully satisfied.”

Tzur Rupin, CMO

Dots

“We saw a 340% increase in website traffic, 2.7x more time on site, and a 38% drop in bounce rate.”

Idan Benishu, Co-Founder & CEO

Bagel AI

“The designs they built were extremely tasteful and contemporary. They were fast in their work and communication.”

Jesse Johnson, Founder & Director

STORE 5a

“Habitat had a multi-step process for delivering business value but were flexible and considerate of our timelines.”

Mike Faber, CEO

Vendo

Where AI Feature Blueprint leads

You leave Blueprint with a 3-month roadmap. Take it in-house or continue with Habitat.

Not every feature should be AI

And not every AI feature should be a chatbot. Let’s work out where AI would make the most sense for your product.

Best reads from our blog

More resources

Have questions?

  • Regular UX and agentic AI UX design aren’t the same thing. AI features come with their own failure modes. For example, a chatbot that confidently makes things up, or an agent that acts without asking. AI Fit is a specific diagnostic for AI features, run against the four dimensions we developed for exactly this problem.

  • That’s what Blueprint is for. Not every feature should be AI, and not every AI feature should be a chatbot. Before we design anything, we map where AI would genuinely help your users and where it wouldn’t. If the answer is “AI doesn’t belong here yet,” that’s a useful thing to leave with.

  • We design the UX. We don’t build or tune the model. But we do check whether the model can actually support the feature we’re designing. There’s no point mapping out a great experience onto a model that can’t deliver on it. If it can’t, we flag it early. If it can, we design against what it can do.

  • We work with what you have. If your design system doesn’t cover agentic patterns yet (most don’t), we bring ours in for those specific parts and match them to your visual language. If you’d rather build the patterns into your own system, we can help with that too.

  • Nothing formal. A founder or PM interview on Day 1 covers what we need. If you already have research, competitor references, or a draft of the AI features you’re considering, bring them. If not, that’s fine too.

  • Yes. If you continue with Habitat on Design Embed or Deployed Prototype after Blueprint, the $4,900 applies toward that total. You pay once for the strategy work, and everything after runs on it.

Let our designers take a fresh look at your product. For free

We’ll go through your product and come back with 3-4 specific things we’d change, along with why we’d change them. It’s the kind of practical feedback you’d normally pay for, and we’ll get it back to you within 48 hours.

Please insert a URL to your product or Figma file and your email below ↓