Challenge

After early enthusiasm for GenAI in his department, the CFO noticed adoption began to stall. Some employees engaged actively, while others were uncertain or resistant. Leadership supported experimentation but lacked clarity on where AI could drive value or why uptake varied across the team.

Strategic approach
Over four weeks, we led a diagnostic engagement across finance, HR, legal, and operations. Through interviews, workflow mapping, and readiness assessment, we identified adoption barriers, surfaced automation opportunities, and developed a 6-month roadmap with targeted pilots and change strategies tailored to the team’s mindset and culture

Outcome

• Team engagement increased to 85%

• Roadmap delivered with clear pilots, success metrics, and communication plan

• Identified peer champions and use cases across functions

• Sparked interest in broader rollout across departments​

 

Key insight

Adoption stalled not due to technical barriers, but because teams lacked shared mental models about where AI adds value. Successful adopters had clearer boundaries around what to automate and why.

Challenge

Guest feedback and booking data were captured in separate tools, making it difficult to understand how sentiment influenced behavior. Insights remained fragmented, anecdotal, and disconnected from business outcomes.

Strategic approach

I led a cross-functional pilot to unify 12 months of qualitative and behavioral data. Working with internal teams and an external ML partner, we applied NLP techniques—topic modeling, sentiment analysis, and clustering—and developed a journey-based feedback taxonomy. The output was delivered through a custom dashboard and cross-functional workshops to support decision-making across product, brand, and operations.

Outcome

The analysis revealed actionable patterns—like sleep quality as a strong loyalty driver, and room controls as a low-impact investment area. The initiative led to a 10% lift in NPS and conversion, while introducing new standards for insight integration across UX, data, and strategy.
 

Key insight

The highest-impact insights came from negative correlation analysis—understanding what guests care least about proved as valuable as understanding what they value most.

Challenge

The client’s teams were manually reviewing 10–25 political newsletters daily to create client briefings, a time-consuming, inconsistent process prone to missed insights and duplicated effort. There was no standardized approach, and summaries varied by team and individual interpretation.

Strategic approach

We led the strategic design and guided implementation of a GenAI-powered insight system that combined summarization, clustering, and conflict detection to generate structured daily digests. Working closely with internal stakeholders and technical partners, we shaped a modular pipeline using Claude 4, OpenAI embeddings, and LangChain orchestration, tailored to team workflows in Slack and email.

Given the company’s high reputational stakes, we prioritized risk mitigation through source transparency, human-in-the-loop review, and gold-standard benchmarking to minimize factual errors and build internal trust. Adoption was supported through onboarding sessions, live editorial reviews, and shared evaluation rituals that helped teams build confidence in working alongside AI—without compromising quality.

Outcome

• 80% reduction in time spent on newsletter synthesis

• <5% editorial review time by week 8

• Summaries featured in 30%+ of client briefings

• Reused architecture in two additional AI pilots

• Created a repeatable model for AI governance, feedback, and adoption

​​

Key insight

Success required treating AI as a junior editor, not a replacement. "Confidence scoring"—where AI flagged its own uncertainty—proved critical for maintaining trust.

Challenge

The research team faced a large volume of unstructured inputs—interviews, workshop notes, and field observations—creating a backlog and slowing decision-making. This manual synthesis workflow was a clear candidate for a generative AI proof of concept.

Strategic approach

We scoped and led the pilot using GPT-based tools to support thematic clustering and summary generation. We developed modular prompt templates, introduced “review flags” to guide human QA, and tested the system across three live research projects. Team onboarding, internal demos, and feedback loops supported understanding and adoption.

Outcome

Reduced synthesis time by ~40%. The pilot improved consistency, accelerated turnaround, and now serves as a reference project across the organization—sparking broader interest in AI literacy and insight automation.

​

Key insight

The most valuable AI contribution wasn't speed—it was consistency. AI provided a reliable foundation that researchers could enhance with domain expertise and strategic thinking.

Challenge

Five internal analytics teams were working in parallel, using different tools and workflows without shared governance. This fragmentation created inefficiencies, unclear ownership, and inconsistent adoption of data tools—slowing down consultant workflows and limiting the value of analytics services.

Strategic approach

As a small cross-functional team, we led an in-depth discovery effort across needs, roles, and departments. We mapped the end-to-end analytics journey, identified friction points, and co-created a new organizational model grounded in user insight. I partnered with leadership to redesign team structures, streamline tooling, and embed UX into long-term operational strategy.

Outcome

The five legacy teams were unified into a single analytics function with clear roles and roadmap ownership. This large-scale transformation delivered over $50M in annual savings, improved tool discoverability and internal alignment, and positioned UX as a strategic partner in org-wide change.

​

Key insight

Biggest efficiency gains came from eliminating inefficient workflows that had evolved organically.