AI pet grooming preview generator to boost sales

Realeza Pets is a grooming business in Curitiba, Brazil, in which I am an investor. It runs its own WordPress-based digital ecosystem, focused on community, services and personalized customer care.

Problem:

Selling a haircut is an experience built on trust. So how do you make that visual? There was also a problem we didn't even know we had: the feature became a reason to "revive" old customers, and a great icebreaker.

Solution:

We built an AI grooming generator that lets pet owners preview different styles on their own pet.

• Upload or use the pet photo already on file
• Apply grooming styles with AI
• Asynchronous generation, so performance isn't affected
• Direct integration with the pet's profile The feature was implemented as part of the Realeza AI system, using image models and custom prompts for each style.

Technologies used:

• WordPress + custom plugins
• Google Gemini (image generation)
• Action Scheduler (asynchronous processing)
• ACF (pet data)

A product-oriented architecture with a long-term vision

The core problem was not just "choosing a haircut". It was the lack of a system connecting decision, visualization and service delivery. Instead of treating this as an isolated feature, we structured the solution as part of a pet data ecosystem, where each interaction feeds future decisions.

• We modeled the pet as the central entity (structured, reusable data)
• We built a style system based on configurable prompts (ready to scale later)
• We implemented asynchronous generation to avoid infrastructure costs and freezes
• We limited generation per style to control usage and cost This approach allowed us to:
• Reduce dependence on manual service
• Avoid rework and misunderstandings
• Create a reusable base for recommending services and products We also anticipated something that usually goes unnoticed: the customer's decision happens before the appointment, not during it. The solution moves that decision into the digital environment, reducing operational friction. Grooming editing screen

Development journey and experience flow

Development was guided by the real flow of use, not by the technology. We started with a simple question: "How does a pet owner decide on a haircut day to day?" From there:

• We removed any need for technical navigation
• We placed the feature directly in the pet's profile
• We built a linear flow: choose a style, generate, preview We also noticed that this is a great icebreaker for reactivating old customers. You can simply send a "hi" followed by "look, I used the feature on our site to see how Aurora would look with different haircuts." The main decision here was to treat the feature as a continuous experience, not a one-off action. Pet page with the generated haircuts

Features and product evolution

The solution started as a simple generator but evolved as we watched how people used it:

• Grooming styles applied with AI on the real pet
• A persistent gallery of generated images
• A smart limit per style to control usage
• An automatic watermark for branding
• Logging and operational monitoring During development, we spotted additional opportunities:
• Using the images as a conversion trigger for booking appointments
• Potential for automatic service recommendations
• Future integration with the history of real grooming sessions This evolution reflects an entrepreneurial approach: not just delivering the feature, but turning it into a strategic asset for the business. Real message sent to a customer

Results

Business impact

• Higher engagement on the platform
• Stronger purchase intent before the appointment
• A clear competitive edge in the local market

Operational gains

• Less time spent explaining things during service
• Fewer misunderstandings between customer and groomer
• Better service predictability

Lessons learned and next steps

• Visual tools speed up decisions in physical services
• AI needs to be connected to the user's real context
• Next steps:
  • Automatic grooming recommendations
  • Integration with scheduling
  • Using history for advanced personalization