Evolve Deflects 60% of Guest Inquiries With AI
Skift reports that Evolve has rebuilt its technology stack over two years to make AI investments traceable directly to the P&L, according to the company's Chief Product and Technology Officer, Arun Nagarajan. Nagarajan told Skift the company began using LLMs in late 2024 for listing builds and QA, deployed AI for routine guest support in late 2025, and is now deflecting over 60% of those conversations without human intervention. Skift also reports Evolve owners earn 18% more revenue and book 9% more nights than the market average. Industry context: Companies converting pilots into infrastructure often tie AI outcomes to financial KPIs, where metrics such as deflection rate and owner revenue lift become the primary levers for further investment and operational rollout.
What happened
Skift reports that Evolve, a U.S. hybrid vacation rental management company, spent two years rebuilding its stack to make AI investments traceable directly to the P&L, per reporting and quotes from Arun Nagarajan, Chief Product and Technology Officer. Nagarajan told Skift the company started using LLMs in late 2024 to accelerate listing builds and QA, deployed AI for routine guest support in late 2025, and currently deflects over 60% of those guest conversations without human intervention. Skift reports that Evolve owners now earn 18% more revenue and book 9% more nights than the market average.
Technical details
Arun Nagarajan is quoted in Skift saying the early uses were focused on speeding listing creation and quality assurance, and that guest-support automation followed. The article frames the work as moving from pilot projects to platform-level tooling so outcomes can be connected to financial metrics on the P&L. Skift provides the specific timelines: late 2024 for LLMs in listings and QA, late 2025 for guest-support deployment, and broader use across engineering, revenue management, and sales in 2026 (as quoted by Nagarajan).
Editorial analysis - technical context
Companies operationalizing LLMs into customer-facing workflows commonly follow a staged path: automate high-volume, low-risk tasks first, instrument end-to-end metrics, and then expand into adjacent functions. Instrumenting AI so it maps to owner-level KPIs, such as revenue lift and nights booked, reduces ambiguity in ROI calculations and helps justify cross-functional adoption. Implementations that report high deflection rates typically pair automation with clear escalation paths and monitoring to catch quality regressions.
Industry context
For travel and property-management operators, measurable revenue and booking lift are persuasive evidence that AI moves beyond cost-cutting experiments into product value. Skift's reporting on Evolve illustrates how sector-specific workflows, like listing creation and guest messaging, are natural targets for LLMs because they scale repetitive language tasks and directly influence conversion and guest experience.
For practitioners
Watch for the same indicators Skift highlights: deflection rate for routine support (percent automated), owner- or customer-level revenue lift, booking conversion changes, and whether teams have tied those metrics into P&L reporting. Observers should also track quality-monitoring practices, fallback/escalation design, and the pace at which automation expands from narrow tasks to revenue-side functions.
Scoring Rationale
This is a notable deployment showing measurable business impact from operationalizing `LLMs` in a travel vertical. It is not frontier research or a major platform release, but the reported **60%** deflection and **18%** revenue lift are directly relevant to practitioners deciding whether to instrument AI against financial KPIs.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
