AI continues to reshape the ways we work, and the ways we work together. 

Now the challenge isn't figuring out how to get AI to write an interview script—it’s deciding what to do with your team's schedule once the administrative busywork evaporates.

In the webinar Rethinking Capacity with AI in the Workflow, hosted by Dscout and the Design Executive Council, design and UX leaders laid out their real-world strategies for repurposing newly found time and prioritizing what matters.

Contributors include:

• Richard Dalton, VP, Head of Consumer Design at Verizon
• Jennifer Darmour, VP of Health Design at Oracle
• Brian Greene, UX Director of Design and Research Operations at Nationwide
• Anshuman Kumar, SVP, Head of Design at Datadog
• Brian Rice, Global VP and Head of Design at Newell Brands

1. Automate work effectively

Injecting AI into your operations requires a realistic look at your current workflow. Teams need to figure out exactly where speed and scale override the need for precision, and where human oversight is a top priority.

Build automated toolchains for evaluative studies

Anshuman Kumar (Datadog): Anshuman’s team handles routine evaluative testing with automated toolchains so that product teams can move faster. They do this through user tracking with Claude and an interview tool to auto-recruit participants and draft basic research plans. 

All of their work funnels into a UXR tool for AI-driven synthesis. The tools are reliable enough for everyday evaluative checks, keeping human researchers focused on deep, longitudinal studies.

Takeaways and advice for leaders:

• Automate the floor, not the ceiling. It helps to package and automate more tactical, evaluative research so that product teams can run basic tests on their own (without overrelying on UXR bandwidth).
• Keep researchers focused on depth. Protect your dedicated research team from routine usability tasks so they can own the more complex, longitudinal work where the most valuable discoveries happens.

Offload information processing

Richard Dalton (Verizon): Richard’s team hands the administrative side of research over to AI. Things like planning drafts, transcriptions, quote pulling, and first-pass analysis. 

However, his team strictly draws the line at synthesis and participant trust and keeps those human-led. "Good enough" AI works for fast, internal reads, but strategic decisions require high-confidence human validation.

Takeaways and advice for leaders:

• Reframe your RACI (responsible, accountable, consulted, informed) Models: AI could act as the worker churning out data summaries, but make sure a human researcher remains accountable for verifying the information and making the final business call.
• Stay human-led in strategy: When it comes to overarching strategy, AI can help with brainstorming, but humans still need to be in the driver's seat.

Streamline massive feedback channels in high-risk environments

Jennifer Darmour (Oracle): Jennifer’s team deploys AI to ingest massive customer feedback channels—like service tickets and spreadsheets—to instantly highlight recurring themes and help conquer backlogs.

In high-risk environments like healthcare, Oracle treats these AI readouts purely as drafts. Human judgment remains mandatory to contextualize trends, evaluate patient safety, and guide strategy.

Takeaways and advice for leaders:

• Reduce the backlog. Try using AI tools to structure messy, high-volume data into clean trend summaries that are much more scanable across the team.
• Treat AI syntheses as first drafts: To echo the earlier advice, never execute a corporate roadmap blindly based on an automated chart. Be sure to always verify automated trends against human risk models.

Use AI as an assistant when speed matters over precision

Brian Rice (Newell Brands): Brian’s team uses AI as a research assistant to eliminate extremely manual, hands-on work and help accelerate the team’s foundational understanding. 

For times when project speed is more valuable than precision, automated synthesis can be a great tool to track patterns quickly. 

The team even uses synthetic personas and virtual moderators to pressure-test early assumptions before funding additional studies (a note to use synthetic personas wisely!)

Takeaways and advice for leaders:

• Hone in on milestones where speed is most important. Identify low-risk project phases where quick alignment is extremely valuable and let AI help speed up those cycles.
• Pressure test early assumptions. Experiment with synthetic user personas to spot obvious design flaws before investing in live user recruiting, but do so with extreme care.

Keep people on the hook

Brian Greene (Nationwide): Brian’s team optimizes operations by offloading tasks like note-taking, coding, and summary creation to AI. 

One hiccup they’ve run into is that LLMs often stumble on technical insurance jargon. To help with this, Nationwide uses a "human on the hook" rule. AI handles the heavy processing, but the human practitioner is accountable for validating the accuracy of the final research product.

Takeaways and advice for leaders:

• Audit specialized jargon and terms. If you’re in a highly jargony field, be sure to actively check AI summaries for mistakes.
• Establish clear accountability: Remind your teams that while automated tools can speed up production, the human behind the AI bears the responsibility for the final output—so triple-check!

2. What to do with your team’s reclaimed time and capacity

Automating administrative overhead often results in reclaimed time. But leaders need to find the best ways to strategically reinvest that time into activities that elevate corporate strategy and business impact.

Maximize coverage through design taxonomies

Richard Dalton (Verizon): Richard’s team maps its entire ecosystem using a 150-box "design taxonomy" grid. 

Headcount limits once left major touchpoints completely unstaffed. But by using AI to automate quick, directional research, they reclaimed human capacity to deploy actual designers and researchers to cover those previously neglected blocks.

Takeaways and advice for leaders:

• Map your design blind spots. Try building a visual taxonomy of your entire user journey to pinpoint exactly which touchpoints lack human research coverage.
• Repurpose time for coverage. It’s critical to avoid the temptation to downsize when AI creates efficiency. Instead, use that free time to expand your team's footprint across unmonitored touchpoints.

Expand the team into unmet needs

Anshuman Kumar (Datadog): The Datadog team uses AI efficiency to expand its ambition and move upstream, closer to the customers. 

Reclaimed hours allow designers time to experiment with new tools and to move directly to design and building functional software code. In addition, freed-up time gives the team the opportunity to build internal tools that help everyone across the company.  

Takeaways and advice for leaders:

• Expand practitioner roles. Push your designers closer to code production and your researchers closer to long-term macro trend forecasting.
• Build proprietary tools. Task your research team with developing custom AI skills and internal prompt libraries to help cross-functional partners build with hard facts rather than "vibes".

Slow down to frame the right problems

Jennifer Darmour (Oracle): Oracle avoids the trap of using AI speed to just ship more deliverables, and instead they reinvest saved time into deep problem solving. 

For example, during a task management project, AI helped accelerate some of the production and discovery activities, letting the team linger in the problem space to realize doctors and nurses have opposite workflow mental models. This changed their entire architecture!

Takeaways and advice for leaders: 

• Linger in the problem space. Use AI efficiencies to buy your team more time to challenge early product assumptions and better understand and frame the true core problem.
• Value strategy over output: Spend your saved hours aligning leadership behind deep user insights, rather than just manufacturing more features.

Turn researchers into niche experts

Brian Greene (Nationwide): Brian advocates for a "shift left" strategy, utilizing automated time savings to turn researchers into heavy-hitting domain experts. 

In financial services, researchers must interview busy, niche professionals. Nationwide spends its reclaimed hours sharpening researchers' business acumen so they can gain deep context.

Takeaways and advice for leaders:

• Invest in industry education. When AI can free up some of the busy work, train your research staff to deeply understand the financial, regulatory, and operational drivers of your specific vertical. 
• Maximize rare interview windows: Ensure that when researchers interact with high-value, specialized user panels, they already have the deep context required to ask advanced questions.

3. How to transform those changes into cross-functional outcomes across the org

AI acts as an organizational bridge, breaking down discipline silos and enabling cross-functional partners to interact with design research in new, highly visual environments.

Democratize early concept validation

Brian Rice (Newell Brands): Brian’s team uses generative AI in cross-functional innovation workshops to engage non-creative partners from sales or finance. Designers use AI tools live to instantly turn verbal stakeholder ideas into photorealistic product concepts. This rapid visualization alters stakeholder participation and compresses the time needed to fund or kill an idea.

Takeaways and advice for leaders:

• Visualize ideas live. For stakeholders who aren’t designers, try using generative imaging tools to visually capture and refine ideas during live workshops. It can help paint a clearer picture. 
• Accelerate the kill switch. Use high-fidelity conceptual rendering to evaluate product viability early and compress your testing lifecycle.

Vet your audience before going with synthetic users

Brian Greene (Nationwide): Human researchers and participants working together is the gold standard. However, there are narrow use cases where synthetic users can help with early on tests that are low stakes and focus on the general population where there’s already a lot of data. 

Highly specialized niche panels—like independent insurance agents—should remain human-led due to their complexity.

Takeaways and advice for leaders:

• Segment synthetic persona use. Restrict your use of simulated user models to general-population consumer profiles, where data is more abundant and tests are less complex and lower-stakes. 
• Keep expert studies human-led. Mandate direct human-to-human research whenever design initiatives target highly specialized, technical B2B user segments.

Dissolve silos using fluid studio models

Jennifer Darmour (Oracle): Jennifer uses AI to dissolve sequential department handoffs, shifting Oracle from rigid functional silos to an adaptable "studio model". So while managers still guide career growth, workers form fluid, cross-functional "delivery pods" alongside PMs, engineers, and regulatory specialists to tackle targeted problems in rapid sprint cycles.

Takeaways and advice for leaders:

• Break down sequential handoffs. Think about transitioning your organization away from rigid departments and build multidisciplinary teams instead to allow for more learning and collaboration.
• Pool aptitudes dynamically: Create flexible delivery pods that can be deployed to high-priority business initiatives, without requiring structural reorgs.

Partner with engineering to reduce AI token burn

Anshuman Kumar (Datadog): Scaling AI-automated workflows triggers hidden corporate costs through skyrocketing API token budgets. 

To navigate these constraints, it helps to form tight partnerships between design leadership and engineering squads. Anshuman’s team works with machine learning teams to test and deploy lightweight open-source alternatives, ensuring internal toolchains remain cost-effective.

Takeaways and advice for leaders:

• Partner with engineering early. Align with engineering leadership to track and optimize the team's collective AI token use before it starts to hit limits.
• Explore open-source LLMs. Actively task your technically fluent design and research practitioners with testing open-source AI alternatives to build more cost-effective internal operations.

Wrapping it up

Teams are continuously exploring how AI is changing their workflows, and the results are continually in flux. It’s most important to stay nimble and be a continuous learner, and use your judgment on what should be automated versus what requires a human eye. 

While it’s easy to prioritize speed over everything else, it’s just as important to know when to use that saved time to slow down and take a hard look at complex environments or use cases. 

At the end of the day, all roads should point to improving cross-functional business operations and creating a better end product for your users. 

‍

You may also like…

• 5 Things We Learned Observing 2,000 Verified AI Users
• Scaling a 1,500 Participant Benchmarking Study into Cross-Departmental Insights
• 16 Ways UX Leaders Can Embrace AI in Everyday Workflows