Stay Ahead with the Ultimate Human + AI Design Process
I've spent almost 20 years working in digital products. When our intern asked about the Double Diamond framework from her course, I thought it would be useful to share my real experience of how we run projects today.
The Problem with Traditional Design Frameworks
The Double Diamond framework has shaped how many organisations approach design. People use it because it offers a clear structure—but this structure is both its strength and weakness. Many teams apply it too rigidly rather than using it as a flexible guide.
The framework aims to help teams think broadly, then focus.
However, its linear design often leads to rigid steps that, when misinterpreted, look like this:
Discover: Stakeholders have an idea and gather some research
Define: The team creates a detailed specification document
Develop: Designers receive the specification and create designs
Deliver: Developers receive the designs and build the product
In large organisations, this process can take more than a year. Even when divided into smaller parts and called "sprints", this approach creates:
Small waterfall processes
Knowledge gaps between teams
Artificial barriers to progress
Frustration at each stage
Hard to iterate and improve
Strong resistance to changes after each step
By the time teams reach the delivery stage, the original vision has changed. No one wants to make changes because starting again feels too expensive and difficult. This is not how teams make good products.
Richard Eisermann, the Design Council's previous Director of Design and Innovation, has noted:
"The question is whether the Double Diamond is still, given the tremendous changes in design practice over the last two decades, fit for purpose? The answer: probably not. The ascendance of fast-paced digital design, along with the complexities of the challenges designers are currently addressing with services and systems, have left the Double Diamond a bit short of breath."
I agree — in 2025, tech allows us to collaborate and iterate using data faster than ever…
…So what's next?
The goal of any design process should be simple: move from idea to actionable feedback as quickly as possible. Then improve.
That's it.
In 'Design Thinking', IDEO's founder Tim Brown advocated for early and frequent prototypes with key stakeholders, saying to:
"Encourage teams to create a prototype in the first week of a project".
This is the essence of what the Double Diamond tries to achieve in the Discovery phase. The best way to achieve this, in my experience, is through effective collaboration and communication.
Designers play a crucial role as translators in this process—they have the unique ability to convert abstract ideas into visual language. While others might struggle to express concepts in words, designers can transform those verbal descriptions into tangible visuals that everyone understands.
This translation skill should be leveraged throughout the entire process, with designers actively replacing ambiguous words with clear visuals at every opportunity.
Why Many Designers Get Stuck
Many designers still prefer working alone with a predefined problem and presenting a solution later. Let a client hand you a list of problems they defined without you and off you go. I understand this—many designers are introverts (me included) and there's comfort in going to our isolated safe spaces with a problem and coming up for air after our eureka moment.
This places too much responsibility on the designer. There is little shared creation, which results in long feedback emails and design review meetings. The designer then has to "sell the idea" to others.
Talk to any seasoned designer about their junior design days and they'll tell you how that process goes… and how quickly you become an extended pencil for whoever is in charge.
Or go to most design forums and you can read pages of designers complaining about their clients "not getting it" or "suggesting ridiculous changes".
These are all because there was a failure in effective communication.
I used to think the best designers were those who could clearly communicate their ideas and "win over" decision makers. While that might move a project forward, the end result misses opportunities and kicks the can down the road. Often for developers to clean up.
Now I believe the best designers are people who can; facilitate discussion, filter out noise, capture important insights, and voice them back clearly in visual formats—always with the view that what they created isn't theirs, it's ours.
The Science-Backed Visual Co-Creation Approach
The most effective approach we've found at Caboodle involves intense workshops that create visuals with stakeholders. These should be quick sketches, flows or wireframes—not finished designs. Something transient, not precious. Written notes and specifications are not enough.
When people can see ideas taking shape visually in real-time, it activates different cognitive processes than when they're just verbal discussions or reading specifications.
Words can be misinterpreted and create multiple different 'visions' in each person's mind. A visual created together, however, does quite a few magical things.
We've seen the magic happen in real-time, but it's also backed by a lot of research:
Picture Superiority Effect (Paivio & Csapo, 1973)
Our brains process and remember images better than text
Visual information is encoded twice - visually and verbally
Explains why stakeholders understand and retain visually-made decisions better than written specifications
Real-time Sketching Benefits (van der Lugt, 2005)
Enhances idea generation in design meetings
Allows team members to build upon each other's ideas more effectively
Supports reinterpretation and linking of concepts between participants
Visual Collaboration Advantages (Eppler & Platts, 2009)
Breaks down knowledge silos during strategic planning
Creates shared understanding among diverse team members
Helps identify edge cases that text-based approaches miss
Complex Problem Solving (Cash et al., 2014)
Visualisation helps teams navigate complex design problems
Makes patterns and relationships visible that would be missed in text-only approaches
Supports the identification of innovative solutions
Practical Implementation
Who Should Be Involved
The best process includes these key roles throughout:
Decision maker: Someone who can approve the project
Subject expert: A researcher, sales person, or customer support team member who understands the users
Technical expert: Someone who understands what is technically possible
Designer: Someone who converts ideas into visuals
Facilitator: Someone who keeps the workshop on track
In smaller companies, one person might fill multiple roles.
Replacing Written Feedback Loops
We ditch written feedback and emails entirely. Every time one of these emails gets sent, someone is left out of the process, or something gets missed. That's how knowledge silos and costly mistakes develop later.
Instead, we replace written feedback with walkthrough videos:
These work because people need time to digest information at their own pace. We send these videos to all stakeholders, ensuring everyone stays informed. Since the direction was set in the first co-creation workshop, nothing should come as a surprise.
If more discussion or a significant decision is required, we simply schedule another workshop. This approach needs to be communicated before the project starts to ensure people will have availability for 90-minute workshops at key stages throughout.
Tools That Enable Collaboration
This remote collaborative approach is supported by powerful tools:
Figma/FigJam for real-time design and wireframing
Loom or Tella for video walkthroughs
Meet / Appear for video calls
This matters because older processes forced teams to work separately. Bringing people together in physical spaces is expensive and time-consuming. It does not support quick improvements.
The Human Element Remains Central
I haven't mentioned AI yet. It helps speed up routine tasks when used strategically, but true innovation comes from human minds working together. When different people who understand a problem share ideas, they create new solutions.
As Napoleon Hill observed:
"No two minds ever come together without, thereby, creating a third, invisible, intangible force which may be likened to a third mind."
The creative stage works best when done together, not separately. I believe co-creation is now a necessity.
But…
Tools like v0, Bolt, Loveable give us the power to see a working, coded prototype in a modern UI like ShadCN. They can even hook into your database and design system. I believe as these tools evolve they will become a powerful tool for a designer to translate and communicate ideas in the most realistic format.
When it becomes quicker to mockup and iterate in these AI tools than with Figma, with a collaborative angle — then the shift away from tools like Figma is likely.
Of course, you've got to know what you want to prompt which is where the human collaboration still plays a vital role. And developers will still be needed to ensure the code is sound and secure.
But the role of AI in the design process is coming, rapidly.
The Process That Works
Written specifications feel comfortable but do not lead to innovative products or easier iteration.
As soon as something becomes visual, the specification often changes. New options arise. Fringe cases emerge. This is why "change requests" appear during design and development.
These late discoveries often extend the budget, delay deadlines and add unnecessary stress for all people involved.
Our approach looks like this:
Explore many ideas using visuals with key stakeholders
Narrow down options, again using visuals
Do this in real-time during workshops
At the end they come away with key assets to move forward:
a detailed primary persona
a journey map for said persona
a set of wireframes made to turn a stranger into a champion
If a client comes with a long spec doc, we:
use AI to turn it into stickies in Figjam
help them remove unnecessary features (that's usually about 80% of the features)
translate those features into wireframes in real-time
This will often remove even more features and replace them with higher impact features they may not have considered. Often it will surface use cases or tech limitations that could be problematic.
Results and Benefits
When this approach works well, you'll see:
Shared accountability with no 'design selling' required
Fewer, if any revision cycles needed
Better alignment between the team
Increased excitement and engagement during ideation
Faster time to prototype and implementation
We have used this method for both complex projects with multiple features and tiny projects, all with great success:
Designed new features for Fortune 100 companies that needed user testing
Entirely replaced "here's what we need designed, can you design it" requests with visualisation workshops
Saved weeks of work and frustration
Brought real energy and passion back to our design process
Reduced client feedback time significantly
Challenges and Considerations
This process can feel uncomfortable at first. You need a facilitator and a skilled designer who can translate ideas into visuals quickly. Without a Lead Designer, this process could fall into the design-by-committee trap. Aim for stakeholders providing insights as opposed to directives.
We've had pushback spending even 90 minutes doing this. People still love to send a document as a means of communication.
The process must remain user-centred. Industry knowledge and early data and research form the foundation of the workshop. In practice, we use personas and journey mapping as part of the process and then frame it around converting strangers into champions. This helps visualise the commercial value.
Looking to the Future
Web products are beginning to adapt automatically to users. Future products might offer completely personalised experiences driven by AI.
Getting user testing results quickly to feed into the design process is important. Current methods still feel slow and expensive. Are AI Researchers the future? They'd offer multiple simultaneous sessions, no breaks required. People seem happy to share their deepest thoughts and feelings with AI already.
A Lead UX Researcher's main role then is ensuring the right questions get asked, spotting emerging patterns and iterating the questions on the fly.
With new coding tools like loveable.dev or bolt.new (no affiliation), it feels like the future is coded prototypes created directly from visualisation workshops. This would allow fast improvement based on user testing and deeper insights from AI analysis of eye tracking, emotions, and more. The AI tools could work from a shared codebase to ensure consistency and aid with iteration.
As we move forward, the essence remains the same: collaborate visually, test quickly, and improve continuously.
