Aliena Cai AI-Augmented Design Method

Last updated: 19 September 2026

Apply AI strategically across the Double Diamond design process to produce more rigorous, faster, and business-credible design work — without letting AI hollow out the human judgment that makes you irreplaceable.

// TL;DR

The Aliena Cai AI-Augmented Design Method is a decision framework for applying AI strategically across the Double Diamond design process — problem space and solution space — without letting AI replace the human judgment that makes designers irreplaceable. Use it whenever you start or are mid-stream on a UX/product project (portfolio, client, or startup) and need to decide where AI accelerates your work (market analysis, PRDs, content generation, prototyping) and where human-led process must be protected (primary research, empathy, synthesis). It combats the most expensive mistake in design: solving a problem that doesn't exist.

Framework

UX design AI tools Double Diamond Figma design process product design prototyping design career

// When should you use the AI-Augmented Design Method?

Use this skill whenever you are starting or mid-stream on a UX/product design project — whether personal portfolio work, client work, or startup product building — and need a principled decision framework for where to deploy AI tools and where to protect human-led process.

// What do you need before applying this method?

Project typerequired
Is this a personal portfolio project, a client/employer project, or a startup/side-business product?
Design stagerequired
Which diamond are you currently in — problem space (First Diamond) or solution space (Second Diamond)?
Design idea or briefrequired
The raw idea, product concept, or existing brief you are working from.
Stakeholder context
Who are the stakeholders? Have the problem and solution already been pre-defined by the business before you joined?
Target market / user group
Who are the intended users, and is internationalisation a factor?

// What are the core principles behind AI-augmented design?

Expand the First Diamond

The single most dangerous failure mode in design is spending months solving a problem that does not exist. Use AI aggressively to widen the First Diamond — market analysis, revenue modelling, competitor mapping — before a single screen is sketched.

AI Assists Primary Research; It Never Replaces It

Primary research means first-hand learning — it is not searching online and it is not asking an AI. Use AI to draft better recruiting emails, stronger interview questions, and sharper notes — but the nuance, judgment, and synthesis belong to you.

Content Is Design, Not Decoration

Lorem Ipsum is a design smell. Realistic content communicates what a product actually does. Replace placeholder text with contextually accurate copy as early as possible, using AI-assisted content generation inside your design tool.

Documentation Is Not Optional

Every undocumented iteration is a liability — it forces rework when stakeholders ask 'Have you tried X?' and it produces weak case studies. Use AI note-taking and structured annotation to capture what changed, why, and who requested it, on every major iteration.

Diverge Before You Converge (Both Diamonds)

Both the problem space and the solution space demand a diverge-then-converge cycle. Never get emotionally attached to the first idea — whether that is the first business concept or the first AI-generated prototype. Generate multiple variants intentionally before selecting a direction.

The Flight of the Bumblebee Workflow

In high-velocity startup contexts, the new design workflow is: idea → prompt → prototype → evaluate → iterate. This is deliberately non-linear and chaotic, but it is navigated by following a fixed order of concerns: structure first, then content, then visuals, then test fast — targeting 7 or 8 out of 10, not perfection.

Being Human Is a Durable Competitive Advantage

AI will never replace who you actually are. Empathy with difficult clients, levity in tense meetings, and leaving people feeling at ease or inspired are the qualities people remember — and the qualities that surface your name when opportunities arise. AI amplifies what you bring; it cannot substitute it.

// How do you apply the AI-Augmented Design Method step by step?

1
Audit whether the First Diamond has been genuinely worked
If the problem and solution were pre-defined by stakeholders before you joined, or if you were tempted to jump straight into Figma, the First Diamond has been skipped or severely shrunk. Flag this explicitly before proceeding.
2
Run AI-assisted market and business validation on your idea
For personal/startup projects: prompt ChatGPT (or equivalent) with your idea and request at minimum: (1) a market analysis, (2) a revenue model design, and (3) a revenue prediction. Repeat this for multiple ideas before committing to one. The goal is business sense, not just user sense. A portfolio project that ignores competitive reality signals to hiring managers that you lack business acumen.
3
Draft the three stakeholder emails using AI (client/employer projects)
Email 1: To the project owner — request inclusion in the discovery meeting before the PM finalises the PRD. Email 2: Meeting invite to all relevant stakeholders (PM, content designer, data analyst, engineers, legal, UX researchers) — this is the kick-off alignment meeting where everyone shares ideas before the problem is locked. Email 3: To the UX research team and PM — propose a short primary user interview plan to get approval before design begins. Ask AI to draft each email from the perspective of the business stakeholder to make the argument more convincing.
4
Protect and conduct primary research — use AI only in support roles
Do NOT use AI-generated research as a substitute for first-hand user learning. AI-appropriate tasks here: drafting recruiting emails, generating stronger interview questions, preparing faster. AI note-taking tools (e.g. Granola) are acceptable as backup, but you must personally review, polish, and analyse the notes. Nuance and meaning are easy to miss; the judgment is yours.
5
Produce a Product Requirements Document (PRD) using AI
A design brief alone is insufficient — it carries only the user perspective and omits business goals and technical constraints, making personal projects feel fake. Use AI to draft a PRD that defines: business objectives, success metrics, scope, constraints, dependencies, and risks. This document becomes the anchor for all solution-space work and is especially critical for startup and portfolio projects.
6
Replace all placeholder content with realistic, context-accurate copy
As soon as you begin laying out screens, eliminate Lorem Ipsum. Use Figma AI's Replace Content feature to describe what each text element represents and generate contextually accurate copy. Select multiple elements simultaneously to batch-replace — this is faster than any copy-paste workflow with an external AI tool. If designing for international markets, use Figma's Translate To feature to stress-test layouts in target languages before engineering handoff.
7
Prototype complex interactions using Figma Make rather than debating feasibility
When an interaction is too complex to prototype cleanly in Figma (e.g. sortable and draggable table columns), do not present a static annotation to an engineer and invite a 'that's too hard' response. Instead: select the frame in Figma, right-click → Send to Figma Make, describe the interaction in plain language, and generate a functional prototype in minutes. This shifts the conversation from 'is it possible?' to 'does this feel right?' — a position of far greater design authority.
8
Apply the Flight of the Bumblebee workflow for startup or early-stage projects
Paste your PRD directly into Figma Make and generate a prototype. Immediately start two to four additional Figma Make projects with the exact same prompt — each run produces different output. Try different AI models per run to maximise divergence. Do not commit to the first result. Evaluate all versions and identify what you like in each before converging. Iterate in this order: (1) structure and happy path, (2) content and copy, (3) visuals and polish. Target 7–8 out of 10 quality — the goal is to test fast, not to perfect.
9
Document every major iteration with AI-assisted annotations
Keep the latest version at the top of your Figma file. Allocate one page per major iteration. On each page, annotate in large, clear text: what changed, why you made the change, and who gave the feedback. Use ChatGPT to generate clean point-form annotations from rough spoken or written notes — this protects your hands from repetitive strain and your project from undocumented dead ends. Paste AI meeting summaries (e.g. from Granola) into the file as the minimum baseline.
10
Test with real users before polishing further
Internal testing with teammates role-playing as users is acceptable but secondary. The best test is always with real users. For startup contexts, user testing also doubles as early waitlist building and demand validation. Do not over-polish before testing — 7 or 8 out of 10 is the right quality threshold at the test stage.

// What does the AI-Augmented Design Method look like in practice?

A designer wants to build a portfolio project and is tempted to redesign an existing popular consumer app.

Before touching Figma, run at least three to five product ideas through AI-assisted market analysis, revenue model design, and revenue prediction. Select the idea with the strongest business case. Draft a PRD — not just a design brief — that captures business objectives, success metrics, and technical constraints. This signals business sense to hiring managers and avoids the 'fake project' read that kills job applications.

A designer joins a project after the PM has already written the PRD and stakeholders are aligned, with pressure to deliver designs immediately.

Draft the three stakeholder emails using AI: (1) request entry into the discovery process before requirements are finalised, (2) propose a kick-off alignment meeting with all stakeholders, and (3) propose a short user interview plan to the research team and PM. Frame all three from the business stakeholder perspective — AI helps you write more persuasively to that audience. These emails can reopen the problem space and expand your influence before a single screen is designed.

A startup designer needs to propose a new feature on a page owned by another team, without an existing design file.

Take a screenshot of the existing page. In Figma, start a new Figma Make project, paste the screenshot, and describe the proposed interaction in plain language. Generate an interactive prototype in minutes. Demo it live in the stakeholder meeting. If moving forward, bring the design back into a Figma file, annotate how the interaction works, and make clear that the demo is a reference — not a spec — and that the final product follows the design system.

A designer is building their own SaaS product on the side and needs to move from idea to testable prototype in days, not months.

Apply the Flight of the Bumblebee workflow: paste the PRD into Figma Make and simultaneously generate three to five prototype variants using the same prompt, varying the AI model per run. Evaluate all variants, extract what works in each, and iterate in fixed order — structure and happy path first, then content, then visuals. Target 7–8 out of 10 quality, then test with real users and use that feedback to build an early waitlist before investing more time.

// What mistakes should you avoid when using AI in design?

Skipping the First Diamond entirely and jumping straight into solution design — this is the most expensive mistake, costing months of work or millions of dollars on a problem that does not exist.
Using AI-generated research (chatbot answers, online synthesis) as a substitute for primary research — real user understanding requires first-hand, direct contact with users.
Keeping Lorem Ipsum in designs past the earliest layout stage — placeholder text hides what a product actually does and weakens both the design and the handoff.
Deleting or hiding past design iterations instead of documenting them — undocumented explorations force rework when stakeholders ask 'Have you tried X?' and make case studies impossible to justify.
Presenting a static annotation to engineers for complex interactions and inviting a 'that's too hard' veto — prototype the interaction in Figma Make first to shift the conversation.
Getting emotionally attached to the first AI-generated prototype and not running multiple parallel variants — always diverge before converging, even in the solution space.
Designing entirely in AI tools without bringing the final interaction back into a proper Figma file with design system adherence and clear specifications for engineering.
Aiming for perfection before testing — the Flight of the Bumblebee workflow targets 7 or 8 out of 10 quality at the prototype stage; over-polishing before validation wastes time.
Treating a design brief as sufficient project documentation — a brief captures only the user perspective; a PRD is required to also capture business objectives, success metrics, scope, constraints, dependencies, and risks.

// What are the key terms in the AI-Augmented Design Method?

First Diamond: The problem space phase of the Double Diamond Design Process — beginning with divergent primary and secondary research, then converging through synthesis to a well-defined problem. Aliena's core argument is that this diamond is chronically shrunk in practice and must be actively expanded using AI.
Second Diamond: The solution space phase of the Double Diamond Design Process — diverging across possible solutions, prototypes, and ideas, then converging through testing and refinement to a validated, engineering-ready design.
Severe Shrinkage of the First Diamond: The common failure pattern where business stakeholders pre-define both the problem and the solution before designers are involved, leaving designers with only a visualisation role and no ability to influence direction or impact.
PRD (Product Requirements Document): A document that goes beyond a design brief to define business objectives, success metrics, scope, constraints, dependencies, and risks. Aliena recommends using AI to draft a PRD for every project — including personal portfolio work — to make designs feel realistic and demonstrate systems thinking.
The Three Stakeholder Emails: Aliena's specific AI-assisted tactic for regaining design influence on pre-defined projects: (1) email to project owner requesting discovery meeting access, (2) kick-off meeting invite to all stakeholders before the problem is locked, (3) email to UX research team and PM proposing a short primary user interview plan.
Replace Content: A Figma AI feature that replaces placeholder or Lorem Ipsum text with contextually accurate, AI-generated copy based on a description of what the text element represents. Can be applied to multiple elements simultaneously.
Translate To: A Figma AI feature that translates all text within a selected frame or set of frames into a chosen language in a single action — used to stress-test layouts for international markets before engineering handoff.
Figma Make: A Figma feature that generates functional, interactive prototypes from text prompts or existing Figma designs. Central to Aliena's startup workflow and complex-interaction prototyping strategy.
Flight of the Bumblebee Workflow: Aliena's named workflow for high-velocity startup design: idea → prompt → prototype → evaluate → iterate. Deliberately non-linear and chaotic, but governed by a fixed iteration order: structure and happy path first, then content, then visuals. Target quality is 7–8 out of 10. Test fast rather than polish.
Happy Path: The core user experience that 80% of users will go through — the first structural priority when iterating in the Flight of the Bumblebee workflow. Must be super intuitive before any other design concerns are addressed.
Primary Research: First-hand user learning — direct interviews, observations, and contact with real users. Aliena explicitly distinguishes this from online searching or AI-generated research, and treats it as the one research mode that must never be replaced by AI.
Fake Project: Aliena's term for a portfolio project (typically a redesign of a famous app) that ignores business reality, competitive context, and revenue viability — readable by hiring managers as evidence that a designer lacks business sense.

// FREQUENTLY ASKED QUESTIONS

What is the Aliena Cai AI-Augmented Design Method?

It's a decision framework for applying AI strategically across the Double Diamond design process — using AI aggressively to widen the problem space (market analysis, PRDs, competitor mapping) while protecting human-led work like primary research, synthesis, and stakeholder empathy. The core goal is producing faster, more rigorous, business-credible design without letting AI hollow out the judgment that makes you irreplaceable.

What is the First Diamond and why does it matter so much?

The First Diamond is the problem space phase of the Double Diamond — divergent research followed by convergence to a well-defined problem. It matters because the most expensive mistake in design is spending months solving a problem that doesn't exist. Aliena argues this diamond is chronically shrunk in practice and must be actively expanded using AI for market and business validation before any screen is sketched.

How do I use AI in design without letting it replace my skills?

Use AI for support tasks — drafting recruiting emails, generating interview questions, market analysis, PRDs, content replacement, and prototyping — but never for primary research, synthesis, or the human qualities that people remember. Primary research means first-hand user learning, not searching online or asking a chatbot. AI amplifies what you bring; empathy, judgment, and nuance stay with you.

How do I make a portfolio project not look fake to hiring managers?

Run at least three to five product ideas through AI-assisted market analysis, revenue modelling, and revenue prediction before opening Figma, then select the strongest business case. Draft a full PRD — not just a design brief — capturing business objectives, success metrics, and technical constraints. This signals business sense and avoids the 'fake project' read that kills applications, especially for app redesigns.

How does this method compare to just using AI to generate designs?

Unlike generic 'AI generates my designs' workflows, this method uses AI to widen thinking, not replace it. It insists on primary research over AI-synthesized research, generating multiple parallel prototype variants before converging, documenting every iteration, and bringing final work back into a proper Figma file with design system adherence. AI accelerates the work; it never becomes the designer.

When should I use the Flight of the Bumblebee workflow?

Use it in high-velocity startup or early-stage contexts where you need to move from idea to testable prototype in days. The flow is idea → prompt → prototype → evaluate → iterate — deliberately non-linear but governed by a fixed order: structure and happy path first, then content, then visuals. Target 7–8 out of 10 quality and test fast rather than perfect.

What is a PRD and why do I need one for a portfolio project?

A PRD (Product Requirements Document) defines business objectives, success metrics, scope, constraints, dependencies, and risks — going beyond a design brief that only captures the user perspective. You need one even for portfolio work because a brief alone makes projects feel fake; a PRD demonstrates systems thinking and business acumen to hiring managers, and anchors all solution-space work.

What results can I expect from applying this method?

You can expect faster delivery, more business-credible design work, stronger case studies, and greater stakeholder influence. Expanding the First Diamond prevents months wasted on nonexistent problems. Documented iterations eliminate rework and make case studies defensible. Prototyping complex interactions in Figma Make shifts engineering conversations from 'is it possible?' to 'does this feel right?' — a position of real design authority.

How do I regain influence on a project where the problem was already defined?

Use the Three Stakeholder Emails, drafted with AI from the business stakeholder's perspective: (1) email the project owner requesting entry into discovery before the PRD is finalized, (2) invite all stakeholders to a kick-off alignment meeting before the problem is locked, and (3) propose a short primary user interview plan to the research team and PM. These reopen the problem space.

Should I still keep Lorem Ipsum in my early designs?

No — Lorem Ipsum is a design smell. Replace placeholder text with contextually accurate copy as soon as you start laying out screens, using Figma AI's Replace Content feature to batch-replace multiple elements based on what each represents. Realistic content communicates what the product actually does and strengthens both the design and the engineering handoff.

How do I prototype a complex interaction I can't build cleanly in Figma?

Select the frame in Figma, right-click → Send to Figma Make, describe the interaction in plain language, and generate a functional prototype in minutes. This is better than presenting a static annotation to an engineer, which invites a 'that's too hard' veto. A working prototype shifts the conversation from feasibility to whether the interaction feels right.

// GET THIS SKILL — FREE