AI as a Partner, Not (Just) a Tool

AI Disclosure: Core ideas are mine, AI was used to create the connective tissue between sections. Heavily edited by me for accuracy and voice. Example prompts and bullet points under each “AI as X Partner” section are mine.

Feed AI generic inputs and you're going to get generic outputs. When I first started using AI, I was impressed … until I tried to get it to write up user tests. It was so frustrating and felt like it took more time just modifying the script it gave. I might as well have done it myself! AI felt underwhelming until I started treating it as a collaborative partner across every aspect of your UX practice.

Here's how I've restructured my workflow to leverage AI as a strategic partner in six key areas:

AI as My Design Partner

Instead of starting from a blank canvas, AI helps me:

• Generate early concepts during ideation sessions

• Assess and rank my ideas against a matrix based on user research

• Critique my work objectively, catching blind spots I might miss

• Identify gaps in my design thinking before they become problems

• Brainstorm unhappy paths and edge cases I hadn't considered

• Map and analyze user journeys for potential friction points

AI as My Product Partner

Product strategy becomes more data-driven when AI helps me:

• Groom and prioritize backlogs based on user impact (imagine shorter grooming sessions!)

• Align features to KPIs and product roadmaps

• Conduct competitive analysis and assess market fit

• Define table stakes and MVP scope more systematically

• Navigate stakeholder dynamics—from managing pushback to securing resources and crafting compelling presentations

• Create comprehensive customer journey maps

• Transform insights into actionable artifacts that teams actually use

• Connect design decisions to acceptance criteria

• Challenge and debate me on features so we can narrow down the priorities

• Log decisions so we can reference them as pivots and challenges come up

AI as My Research Partner

Research becomes more thorough and systematic:

• Develop and refine interview questions and research plans

• Synthesize findings by identifying patterns and extracting key insights

• Rank insights by frequency and impact for better prioritization

• Create executive summaries and research debriefs that stakeholders can quickly digest

• Extract themes and patterns from user interviews

AI as My Development Partner

Handoffs become smoother when AI helps me:

• Anticipate developer questions before they arise

• Identify all potential edge cases and error states during design

• Brainstorm solutions for complex logic problems

• Create comprehensive handoff documentation that reduces back-and-forth

AI as My Brand and Content Partner

Content creation becomes more strategic:

• Craft compelling copy that aligns with brand voice and fit the design

• Develop brand narratives that resonate with users

• Plan content calendars based on user needs and business goals

• Stay current with market and content trends

• Build templates and prompt libraries for consistent output

AI as My Business Partner

Design decisions carry more weight when AI helps me:

• Align design work with OKRs and KPIs from the start

• Extract UX opportunities from town halls and earnings calls

• Build compelling ROI cases for design initiatives

• Identify high-impact projects that move the needle

• Learn from relevant case studies across industries

• Navigate complex stakeholder relationships more effectively

AI as My Accessibility Partner

Inclusive design becomes more comprehensive:

• Audit designs for accessibility issues before development

• Ensure GDPR, HIPAA, and KYC compliance from the design phase

• Identify blind spots in user consideration

• Role-play as users with disabilities to uncover usability barriers

• Expand inclusive design thinking beyond my own perspective

Example Prompts

There are many resources on prompt engineering but the best practices for prompts stick to this format: role, context, format, constraints. Feel free to copy and paste!

For accessibility audits:
"Role: Act as an accessibility and ADA compliance expert.
Context: Analyze this Figma prototype for WCAG 2.1 AA violations. The target user base is X. This will be built using ReactJS.
Format: List the top issues and suggest specific fixes in a table starting with the most impactful. Rate by impact and effort. Include ARIA tag suggestions.
Constraints: We need to launch this in 90 days and also have to build features A, B, C so we will need to pick the highest impact accessibility issues to fit into our sprint schedule."

For design critique:
"Role: Act as a Principal Product Designer with an eye for detail and microinteraction.
Context: Critique this design. The user is X. The user’s objective here is Y. Some rules: user cannot do Z. Assess this against UX best practices and similar tools.
Format: Tell me what is effective and what is ineffective. Ask questions about the design and user experience. Point out any gaps or edge cases and user requirements I’ve overlooked. Rank feedback starting with most critical. Provide your rationale for your feedback.
Constraints: I cannot change the colors, fonts, or button styles and any content changes will need to go through approval with the content strategy team."

For product alignment:
"Role: Act as a product owner.
Context: Assess this design against this product roadmap. I am concerned about our ability to fit all the functionality from the design to complete the features on time. Table stakes are A, B, and C and currently we are debating about the feasibility of features D and E.
Format: Give us a few options with pros and cons of each to help us deliver this product on schedule.
Constraints: We have X developers and Y time. Some other needs competing with our resources are bug fixes with the current system. Stakeholder Z is concerned about A, B, and C."

To reiterate: generic inputs create generic outputs. So give context when possible. (But be aware of the diminishing returns as context degradation can occur.

As Humans, Our Skill Becomes More Important Than Ever

As I continue integrating AI into my practice, I've realized something important: the technology amplifies both good and bad UX thinking. Your foundational skills need to be sharper than ever to effectively evaluate AI outputs, catch inconsistencies, and maintain design quality.

When you partner with AI instead of just using it as a tool, you’ve got to push back. Challenge It. Question it. Here’s a real example from a script I tried to create with DeepSeek.

Well, that doesn’t exactly inspire confidence …

So even though I said AI is best used as a partner … it’s more like an inexperienced intern who really wants to flatter and agree with you. And while there are prompts to circumvent this, just be aware and honest with its limitations. And always read the output! It’s alarming how often people don’t do this.

AI doesn't replace UX expertise, it demands more of it. The designers who thrive will be those who can seamlessly blend AI capabilities with strong design fundamentals and critical thinking.