Kayla Salvucci

Proving AI ROI to executive leadership through adoption and operational impact measurement. Built a multi-layer adoption framework, distinguishing genuine workflow impact from passive usage by tracking real downstream actions. Defined helpfulness, accuracy (including hallucination) metrics. Anchored baseline on ticket lifecycle analytics. Average support time spend per ticket dropped 35% post-launch.

A single source of truth for the Chief Customer Officer's entire book of business. Worked with the CCO and her leadership team to define top-level KPIs that broke down into KPIs for each sub-organization in her purview. Locked in the exact calculation logic, built the underlying data model with forecasts and pacing calculations, combining statistical projections with run-rate tracking so leadership could see not just where they stood against targets, but where they were headed.

Building a system to measure a customer support outcome that, by definition, leaves no record. How do you quantify a case that was successfully prevented? I broke deflection into two measurable categories: implicit (session behavior showing a customer engaged with AI-suggested content and never filed) and explicit (cases created but fully self-resolved with zero agent touch). Defined the behavioral rules for every ambiguous edge case, turning an abstract AI investment into a concrete, trustworthy operational metric — and surfacing which product areas were under-documented or under-supported along the way.

Replacing a manual spreadsheet heuristic with a time-series ML model that adapts as the business changes. The old model was obscure, complex, outdated, and laborious. I made the case to leadership, rebuilt it with SARIMA, and surfaced forecasts in leadership dashboards. The model organically picked up on the structural shift: customer growth rising while ticket volume declined due to AI deflection — something the bookings correlation-based model had no way to detect. Refresh time dropped from ~2 weeks of cross-team coordination to about a day. Persistent accuracy at the skill and location level directly advised hiring and budget.

Modernizing territory management and rep workflows through targeted AI integration. Started with user interviews across reps and leadership to discover what they actually needed. Then drove a multi-part rebuild: scoping and writing technical requirements for a white space analysis built onto our internal AI platform; scoping for and collaborating with the Salesloft systems team to build a territory management view; configuring AI-enriched account fields (competitors, tech stack, signals); standing up a Gemini-powered research and outreach agent; and deeply scoping executive meeting prep summaries built with engineering. Back-of-envelope: ~$2.5M in annual operational cost savings across the sales org from research and prep time alone.

Measuring whether automated case routing actually improved support outcomes. The support team rolled out skills-based routing — automatically directing cases to agents with the right expertise. Leadership needed to validate the effort and have ongoing visibility into skill group performance. I built an operational dashboard tracking core OKRs, such as escalation rate, time-to-assignment, and handoffs across each skill group. Worked with data engineering to build the underlying models. Clear directional result: lower escalation rates, faster assignment, faster resolution.