URBANO·CX/AI

Customer Experience · AI Systems

Gerard Urbano

Director, Support, Education & Enablement

AI Deployment & Productivity — CX

Thirteen-plus years in CX, starting on the frontline and moving through documentation, education, and enablement leadership. That arc came together: I unified Support, Customer Education, and Enablement into one 15-person operating model. Having all three under one roof meant customer programming, day-to-day support, and even how we sold could draw on the same learnings — including running Support itself at ~40,000 tickets a year, 4.7 CSAT, and 95% SLA — and I got there by building the AI systems that made the team faster, not smaller.

7AI systems built across Support, Education & Enablement
4.7CSAT, up from 3.5
95%SLA attainment, up from 85%
50%agent productivity lift from AI tooling
3×headcount growth, zero voluntary attrition
$60Krevenue generated from one AI-enabled headcount

Field-deployed AI systems

Deployed AI Usecases

Five systems I designed and built myself, in production — using AI to turn a pile of disconnected tools and processes into something a team could actually run on.

Here's how I think about AI: it complements us, it doesn't replace us. We steer it, correct it, teach it — and it works for us. Done right, it's not taking away our ability to think or get creative. It's more like a superpower: it just helps us think bigger and do more.

SYS.01 · Unified Knowledge Hub
Gemini · internal help site · Slack · Drive · GitHub · Notion
Problem

The company had too many sources of truth — three help-site subspaces, a dev site, plus Slack, Drive, Docs, GitHub, and Notion internally. Nobody agreed on where to look first.

Build

A single interface, powered by Gemini, connecting every one of those resources into one central hub. Slack, the help site, and the rest became targeted lookup tools instead of competing sources of truth.

+20%user efficiency across platforms via federated search
SYS.02 · CS Knowledge Assistant
Claude · MCP → shared Claude Project · JIRA · GitHub · Slack
Problem

CSMs are supposed to focus on business value, but kept getting pulled into product questions they weren't equipped to answer fast. Technical Support needed the same speed on inbound tickets.

Build

A custom MCP — later converted into a shared Claude Project for the whole CS org — pulling from JIRA, every help doc, GitHub product code, and Slack. It didn't just search for answers; it read the code when needed and drafted the reply.

Org-wideany CSM or Support agent could now answer a product question
SYS.03–05 · Technical Support Ops Suite
Claude Code · Shortcut · PagerDuty

Three tools, one queue, built off the same underlying data — one view for the manager, one for the agent, one for the exec.

Watchtower — manager, daily

Queue-wide health check

Reads every open request end to end — not just the metadata — stripping quoted trails and signatures to score three real signals: customer sentiment from their actual language, days since the team's last reply, and how long a request has sat open. Segments the queue by type (AI/product escalations, platform issues, engineering-owned bugs, active implementations), flags which accounts and owners are carrying the load, and closes with a short, ranked list of the calls to make that week.

Buddy — agent, real-time

Agents' primary workspace

Replaced the platform's native queue as agents' main way of working — prioritized by sentiment, priority, last response, days open, CSAT. Caught duplicate tickets, flagged incident-level themes, linked open engineering tickets, added team broadcasts and PagerDuty scheduling.

5× efficiency
Manager Buddy — leadership, real-time

Command center for the queue

An operations dashboard built to run the day, not just report on it: live SLA, CSAT, and first-response time with date-range comparisons from today out to last year, tickets auto-grouped by theme so managers can see where the volume actually sits, and an incident detector that clusters related tickets into one flagged event — confidence score, affected accounts, likely symptom — before it turns into an escalation. One click deep into team performance, capacity, schedule, QA, and a rollup built for execs.

Auto-detects incidents
SYS.06 · Enablement Operating System
Claude Code · custom AI-scored intake · automated reporting
Problem

I inherited an Enablement team with a reputation — disorganized, missed dates. Managers said yes to every ask with no real way to prioritize the work.

Build

Started with an Enablement Charter — what the function does, what it doesn't, how it operates — to anchor everything after it. Then an AI-scored intake form that checks every request against the charter and current company priorities, and generates the task list, due dates, and materials needed. Added retention analysis and automated weekly/monthly reports for Sales leaders.

80%skill retention · 100% on-time enablement delivery
SYS.07 · Beta Program Microsite
Custom microsite template · PM/PMM intake form · Slack feedback routing
Problem

The company shipped fast, which meant running open and closed beta programs — but customers felt out of the loop as features changed mid-beta. Feedback flow was inconsistent between product teams and participants, and product information lived in disparate, disorganized places.

Build

Built a microsite template that only needs a PM or PMM to fill in the essentials per beta feature — description, feature value and why to adopt, how it improves workflows or CX, screenshots, and the owning product team — to generate a single point of truth customers could be pointed to. Routed inbound feedback straight to Slack instead of Support, so engineers and product leaders heard from customers directly, with no support bottleneck in between.

+50%inbound beta feedback, with the Support bottleneck removed

Career log

Experience

  1. 05/2025 – 09/2026

    Director, Customer Support, Education & Enablement

    GLADLY

    Took on Technical Support and Enablement, unifying it with Customer Education under one function.

    • Scaled Support from 3 to 11 with a BPO program at flat budget; 24/7 coverage, zero attrition, ~40,000 tickets/year across L1/L2/DSE tiers.
    • Improved CSAT from ~3.5 to ~4.7 and SLA from 85% to 95%, holding zero voluntary attrition through mentorship programs.
    • Built a custom AI agent through Claude Code, lifting agent productivity 50% with real-time guidance for research and response.
    • Designed AI-powered federated search, boosting user efficiency 20% across multiple platforms.
    • Led enablement across six functions — 80% skill retention, 100% on-time delivery. Grew Enablement site usage 50% via a custom content hub.
    • Built AI via Claude Code to automate Enablement capacity planning and task management.
    • Created an AI Optimization Specialist role, converting Support from a cost center to revenue-generating — ~$60,000 generated from a single headcount.
  2. 09/2022 – 05/2025

    Director, Customer Education

    GLADLY

    • Defined and executed Gladly's global Customer Education strategy — built the academy infrastructure that became the foundation for company expansion.
    • Expanded Academy footprint via an elearning partner: usage up 300%, from 5,000 to 20,000 learning hours, at a 75% completion rate.
    • Cut implementation time-to-value 50% (6 weeks → 3) through structured onboarding paths and role-based enablement.
    • Drove 100% adoption of the train-the-trainer framework across enterprise customers (Nordstrom, Ulta, Sephora).
    • Built a self-serve onboarding curriculum enabling SMB expansion and doubling implementation capacity from 10 to 20.
  3. 06/2021 – 09/2022

    Senior Manager, Customer Education

    GLADLY

    • Launched DITA-structured, enterprise-grade help documentation — the primary source of truth across CX, Product, and Sales; cited by customers (Manscaped, Movado) as a differentiator over Zendesk and Salesforce.
    • Architected Gladly Academy end to end — ADDIE-based learning paths, certifications, video tutorials, live sessions, release-aligned education — delivering 43,000+ training hours at a 74% completion rate.
  4. 01/2020 – 06/2021

    Manager, Technical Writing

    GLADLY

    • Owned product documentation and customer training through hyper-growth. Led the redesign of help.gladly.com (~500,000 visits/year) to cut support overhead.
    • Led an LMS rollout for SMB and Enterprise growth; automated customer training saved CSMs 300+ hours/year.
    • Established the company's release-note process and knowledge management program.
  5. 05/2014 – 06/2019

    Sr. Content Manager / Content Manager

    TAULIA INC. (SAP)

    Content strategy for a supplier/buyer base of 1.5M+ companies and $1T+ in invoice volume.

    • Drove a 300% improvement in CSAT engagement; cut administrative escalations 56% and "how-to" ticket volume 15% via self-service redesign.
    • Built a real-time Network Operations Center (NOC) to monitor service-center SLA performance — an 8% support-visit-to-case ratio.
    • Implemented KCS v6 and a real-time live-chat translation service across a 4-language help center.
    • Led the company's LMS launch and the Salesforce Knowledge Lightning transition.
  6. 2009 – 2014

    Information Analyst / Service Desk (Level I → II)

    TACO BELL CORPORATE

    • SME and trainer for a service desk handling 600K–1M annual escalations across hardware, back-of-house apps, and Linux/Windows systems.
    • Revamped Remedy knowledge management with decision trees and a keyword system — the format was later adopted by KFC and Pizza Hut.
    • Built interactive LMS training and led rollouts including the e*Restaurant cloud application.

Capabilities

Skills & Platforms

AI & Automation
ClaudeGeminiMCPAgentic agentsFederated searchAI-first workflowsVercelGoogle Apps Script
Platforms
GladlyZendeskSalesforce Service CloudRemedyJiraWorkramp LMSPagerDutyLivestormSlackGoogle Workspace
Metrics & Voice of Customer
NPSCSATSLA managementAdoption trend analysisQAContent hub development
Operations & Leadership
Omnichannel supportBPO & outsourcingEscalationsCapacity planningKCS v6ADDIE designOrg design & hiringBudget ownershipMentorship
Award

CIO "Breakthrough Result" Award

Taco Bell Corporate — for the e*Restaurant pilot training implementation.

Background

Jurupa Valley HS — GED

Riverside Community College — Psychology & Science coursework.

AI deployments

The actual interfaces, not slides about them

Six systems I designed, built, and ran in production — shown here exactly as they work.

manager-buddy · operationsLive file
Manager Buddy — the actual file, branding swapped and every name/account fictional. 11 views: Dashboard, Signals & Incidents, Team Performance, Queue, Customer Health, Capacity, Team Schedule, QA Program, Shortcuts, Implementations, Training, Executive View — sidebar and period tabs are all live. See SYS.03–05.
watchtower · daily digestRecreated — sample data
Watchtower / Escalation Watch — reads every open thread, scores three real signals, ranks the calls to make. Scroll the frame — signal cards expand on click. See SYS.03–05.
enablement-portal · homeLive file
Enablement Portal — entry point into the intake wizard and tracker. See SYS.06.
enablement-tracker · programsLive file
Enablement Tracker — ranks every program by priority band, due date, RICE. See SYS.06.
buddy · assign-to-meLive file
Buddy — assign flow — the one-click claim-and-route interaction agents use instead of the native ticket queue. See SYS.03–05.
channel-finder · internal searchLive file
Channel Finder — describe the question, get pointed to the right internal source. Part of SYS.01.