AI Product Design

Human-Centered AI

Making Generative AI Accessible Through Familiar Workflows

I designed an AI content creation platform that helps creative teams generate professional product imagery without requiring deep technical knowledge of AI.

Role

Lead Product Designer

Platform

Web application

Coming soon

01 · Problem & context

Product photography was the bottleneck.

Producing imagery for large product collections can involve extensive planning, physical production, models, locations, photography, and post-production.

Generative AI created an opportunity to rethink this process. The challenge was that many target users were not comfortable with advanced AI tools and had difficulty adapting to unfamiliar interaction patterns.

The goal wasn't simply to build an AI image generator. It was to make powerful generative AI accessible within a workflow that already made sense to users.

Months → Hours

reduced production time by months thanks to AI

€10k+

estimated average savings from photo shoots

0 of 12

interviewed teams comfortable with raw prompting

02 · Approach

Meet users where they are

Rather than asking users to learn how to prompt an AI model, I designed around workflows they already understood.

We studied how users interpreted different parameters and used familiar terminology throughout the experience. Guided controls replaced open-ended prompting where it could lead to inconsistent or ambiguous results.

Research also revealed an important shift in the user's mental model. They didn't want to create one image at a time. They wanted to define a model and visual style once, establish a set of poses, and generate an entire collection.

That insight shaped the product around bulk creation rather than individual image generation.

03 · Key decisions

Three decisions carried the product.

Guided prompting

Open-ended prompts gave users too much responsibility for describing complex visual requirements. Structured controls gave them more predictable results while still allowing fine-tuning.

Define once, generate many

Users could establish a model, style, and set of poses once, then apply them across an entire collection.

Consistency with creative freedom

The system needed to support different brand identities while maintaining enough visual consistency for generated images to feel like part of the same campaign.

04 · Outcome

A new way to produce product imagery

94%

reduction in time from brief to first usable image

more campaign variants produced per quarter

0

prompt-engineering training sessions required

The platform enabled an entire collection to be generated within a few hours, transforming a traditionally physical production process into a digital workflow.

It generated commercial interest from large organizations, while client-specific results remain confidential. The product also received industry recognition for innovation and responsible AI development.

Most importantly, the project demonstrated that advanced generative AI could become accessible without requiring users to become experts in the technology.

Marco Manuel

© 2026 Marco Manuel. All rights reserved.

Marco Manuel

© 2026 Marco Manuel. All rights reserved.

Marco Manuel

© 2026 Marco Manuel. All rights reserved.