Melissa Cameron
Melissa Cameron
Senior Product Designer
Face Swap Cards
Platform - App
Face Swap was thortful’s first AI-powered personalisation experience, combining emerging technology with creator-led designs to give customers a more unique and thoughtful way to personalise cards.
As the sole Product Designer, I led the experience end-to-end, from discovery and concept development through to testing and launch, helping deliver 134K+ card sales and an estimated £565K in gross sales.
Metric - Conversion
Impact
The launch of Face Swap successfully demonstrated the commercial value of AI-powered personalisation. Year to date we have sold over 270K cards, this is a great result given we have no marketing spend to push this further, so this high intent organic customers.
Cards sold to date
134K
% increase in AOV for single card orders
5%
Incremental revenue
+ £27K
DISCOVERY
Customer Problems
Customer Insight
Customers wanted more fun and memorable ways to personalise cards beyond simply uploading a photograph.
-
Existing photo cards lacked novelty and surprise.
-
Many customers wanted cards that felt more unique and shareable.
-
Customers were unfamiliar with AI-powered experiences, meaning the journey needed to feel incredibly simple and trustworthy.
-
There were concerns around privacy and how uploaded photos would be used.
Problem definition
Customer feedback highlighted a growing gap in thortful’s personalisation experience. Customers wanted their cards to feel considered and chosen specifically for the recipient, something that showed they had put thought and effort into the gesture, rather than picking up a generic card from a supermarket.
While deeper personalisation could help address this need, our legacy platform made significant changes to the existing card editor complex and costly, limiting our ability to offer the level of customisation available elsewhere in the market.
Hypothesis: We believed that introducing AI-powered Face Swap would enable customers to create more personal and considered cards, increasing engagement, conversion and order value.
Solution Discovery
Research
Competitor benchmarking
At the start of the project, I conducted competitor benchmarking to understand how other products were introducing AI-powered personalisation and the interaction patterns customers were already familiar with. This helped identify opportunities to differentiate thortful’s experience, establish best practices around photo upload, generation and retry flows, and ensure we were creating an experience that felt intuitive while still being distinctly thortful.
Customer Research
Following the competitor benchmarking, I spoke with 20 customers across two distinct segments: top-tier lapsing customers and customers who had only placed 1–2 orders.
Some of the key questions we explored included:
-
What type of AI-generated content would customers actually enjoy?
-
Would customers trust AI with their personal photos?
-
How should we communicate the feature without overwhelming users?
-
How would this impact thortful’s creator community?
-
Was this a scalable long-term proposition?
Assumption: Different customer segments would be motivated by different benefits. A successful loyalty proposition therefore needed to balance retention of high-value customers with the opportunity to increase frequency amongst lower-engagement customers.
Key research insights:
Worried about Ai
Insight: Customers struggle to justify paying upfront for a greeting-card membership unless the financial benefit is obvious.
Breadth of content
Insight: Customers struggle to justify paying upfront for a greeting-card membership unless the financial benefit is obvious.
Needs to be easy to follow
Insight: Customers struggle to justify paying upfront for a greeting-card membership unless the financial benefit is obvious.
DELIVERY
Ideation
Early ideation sketches allowed me to quickly explore different ways customers could discover, create and refine a Face Swap without committing to a solution too early. By mapping alternative journeys and interactions on paper, I could identify potential friction points and narrow the concepts towards the simplest experience for the MVP.
Concept Validation
Option 1 - Ai generated Cards
The first concept explored was giving customers complete creative freedom by allowing them to generate an entirely new greeting card using AI prompts. Customers could enter a description such as “Create a birthday card of my dad riding a dinosaur in the style of an 80s action movie” and the AI would generate both the artwork and potentially the accompanying copy.
POC using Claude Ai to get early buy in
Option 2 - Creator led Faceswap templates
The second concept focused on enhancing existing creator-designed cards rather than replacing them. Customers could upload a photo and use AI to seamlessly swap the face of a celebrity, character or illustrated figure with the face of a friend or family member.
This approach combined the excitement of AI with a familiar and intuitive personalisation journey. Rather than asking customers to create something from scratch, they simply selected a design they loved and transformed it into something uniquely personal. It also preserved the role of thortful’s independent creators by building on their existing artwork instead of generating entirely new designs.
Proposed Solution
Option 2: Creator-Led Face Swap Templates
It complemented thortful’s existing marketplace rather than competing with it. By allowing customers to insert faces into artwork created by our independent creators, we could harness the excitement of AI while preserving the creativity and personality that make thortful unique.
This approach also enabled creators to continue producing topical, humorous, and culturally relevant designs that resonated with customers throughout the year, while AI acted as an enhancement to their work rather than a replacement. From a customer perspective, the concept was immediately understood, required very little learning, and delivered a fun, personalised experience that felt authentic to the thortful brand. From a business perspective, it strengthened our creator ecosystem, reduced the risks associated with fully AI-generated content, and provided a scalable platform for introducing new Face Swap designs across occasions and trends.
Final Designs
Before exploring technical feasibility, I wanted to validate whether this was a problem worth solving.
I conducted interviews with loyal customers alongside workshops with Marketing and Customer Happiness to understand customer expectations, potential concerns and the wider business opportunity.
Some of the key questions we explored included:
-
What type of AI-generated content would customers actually enjoy?
-
Would customers trust AI with their personal photos?
-
How should we communicate the feature without overwhelming users?
-
How would this impact thortful’s creator community?
-
Was this a scalable long-term proposition?
The results
On Web+0.2ppt improvement to ecom conversion. Through the remainder of 2026 this is worth roughly:
-
~3k incremental orders
-
~£27k incremental revenue
Within the variant around 11% of purchases are through Apple Pay Express CheckoutOn iOS+0.5ppt improvement to ecom conversion. Through the remainder of 2026 this is worth roughly:
-
~2.7k incremental orders
-
~£24k incremental revenue
Within the variant about 15% of purchases are through Apple Pay Express CheckoutAcross both platforms the estimated rest-of-year impact is ~£51k incremental revenue