Industry 04 / 05

UX design for AI products and agents

delbueno™ Studio designs AI-native products: agent workflows, copilots and explainable model output. We focus on trust, with human-in-the-loop controls, clear confidence signals and recovery paths, so users know what the system did, why it did it and how to correct it. Buzz Bunny, 1to1000 and Shay are our AI work.

Book a 15 minute call See our services

Kickoff today. First concepts in 48 hours.

Typical problems

What goes wrong in AI.

  1. Output users cannot inspect

    People do not trust a result they cannot question. Explainability has to be designed for each audience, from the end user to the auditor.

  2. Agent actions that run without review

    The question is where a human belongs in the loop. We sort actions by reversibility and blast radius and design approvals around that.

  3. Uncertainty that is hard to communicate

    Intelligent software is sometimes wrong. Confidence signals, previews and recovery paths keep that honest without slowing the user down.

  4. Models and workflows nobody can read

    AI dashboards for models, pipelines and results need the same clarity as any complex SaaS screen, or the team cannot act on what the system shows.

What we design

The screens that carry the weight.

Related work

Work behind this page.

Four Buzz Bunny app screens showing top recommended stocks, a price forecast chart, top gainers and a portfolio growth ring

Buzz BunnyFintech

91%

onboarding completion, up from 62%

Three 1to1000 onboarding screens asking where the user works, which style they prefer and what matters most, with swipeable photo cards

1to1000Proptech

45 days

from concept to market-ready MVP

Shay loyalty app landing page with the headline Tus premios favoritos, siempre en tu bolsillo and a phone showing active loyalty cards

ShayRetail and loyalty

+37%

customer retention

From the journal

Go deeper.

AI & UXDesigning explainable AI UX to Build Trust in Intelligent InterfacesExplainable AI UX design helps users understand how decisions are made. Learn strategies to build trust, transparency, and confidence in AI-driven systems.AI & UXAI Agent UX: Designing Human-in-the-Loop ControlsHow to design human-in-the-loop approvals for AI agents: preview before, interruptibility during, audit and undo after. A UX framework for trust.AI & UXDesigning AI first product design patterns for intelligent systemsA field guide to AI-first product design: UX patterns for intelligent systems, ML interface design and interfaces that earn trust and speed decisions.

All journal posts

Frequently asked questions

How do you design human-in-the-loop experiences?

We define which actions need approval, show the agent plan before it runs and make every step reviewable and reversible.

Have you shipped AI products?

Yes. Buzz Bunny is an AI stock recommendation app, built in Flutter, where onboarding completion rose from 62% to 91%. 1to1000 is an AI real estate matching app that went from concept to MVP in 45 days. Shay pairs a loyalty app with an AI-powered CRM.

Can you prototype AI features quickly?

Yes. We use AI-assisted prototyping to put a working version in front of users early, as we did for the diagnostic wizard in the Vzari store.

Other industries

Also in our field.

Cloud

Interfaces for cloud platforms, cost tools and infrastructure consoles.

Cybersecurity

Interfaces for security operations, posture management and threat response.

DevOps

Interfaces for pipelines, observability and developer platforms.

Fintech

Interfaces for payments, banking, investing and financial operations.

Next

Let’s build intelligent products powered by AI.

Book a call

or hello@delbuenostudio.com