At the Skift Data + AI Summit 2026, a session titled "The Scaling Decision Nobody Warned You About" featured Wei Manfredi, SVP AI & Architecture at IHG Hotels & Resorts, and Arun Nagarajan, Chief Product and Technology Officer at Evolve, moderated by Adriana Lee of Skift. Skift reports Manfredi and Nagarajan argued that fundamentals such as infrastructure, data hygiene, and clean APIs matter more than glamour technologies for making properties visible to AI agents. Skift reports Nagarajan said Evolve ramped a guest-facing AI resolution platform from roughly 30% resolution to 60% resolution in under 120 days. Manfredi was quoted calling culture and people "the bigger roadblock" than technology. Editorial analysis: The session reinforces a recurring practitioner lesson that operational readiness and incentives often determine whether AI pilots scale into customer-facing systems.
What happened
At the Skift Data + AI Summit 2026, a session titled "The Scaling Decision Nobody Warned You About" featured Wei Manfredi, SVP AI & Architecture at IHG Hotels & Resorts, and Arun Nagarajan, Chief Product and Technology Officer at Evolve, moderated by Adriana Lee, Travel Technology and AI Reporter at Skift. Skift reports both panelists emphasized foundational work over product theatrics and rejected separating pilot and production as discrete phases. Skift reports Nagarajan said Evolve ramped a guest-facing AI resolution platform from about 30% to 60% resolution in less than 120 days. Skift quotes Manfredi saying the challenge is "the boring stuff" like infrastructure, data, and clean APIs and that culture and people can be larger barriers than technology.
Editorial analysis - technical context
The messages from the session align with common operational themes in production ML. Organizations that move a capability from pilot to production typically confront data quality, API surface standardization, monitoring, and the need to iterate models and business logic continuously. These are cross-functional problems that often require product, engineering, and operations processes to evolve in parallel.
Context and significance
Industry observers frequently report that labeling an effort a "pilot" lowers performance expectations and governance rigor; the Skift report includes a concrete metric from Evolve showing rapid improvement when standards were raised. For practitioners, that underscores two repeatable patterns: invest in reproducible infrastructure and instrument customer-facing flows so incremental gains can compound into measurable business outcomes.
What to watch
Observers should track whether teams adopt continuous delivery practices for ML, how incentives and roles change as agentic workflows take on customer work, and whether organizations treat data and API hygiene as programmable, monitored assets rather than ad hoc integrations. Skift has not published additional vendor roadmaps or IHG internal plans in this report.
Scoring Rationale #
The session delivers actionable, operational lessons for practitioners about scaling AI into customer-facing systems, reinforced by a concrete metric from Evolve. It is notable for practitioners but not a frontier research or platform release.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.