← Back to joewilson.ai
Case File / Director of CS
2025 — 2026 / Rapsodo

Joe Wilson.

Director of Customer Success, Golf & Diamond Sports

AI-NativeOperator2026
THESIS

Ran customer success at a hardware-to-SaaS sports tech company as a one-person CS Ops, data engineering, and analytics function — because AI was the multiplier. The work below is what one Director normally hands to a five-person team. Here, it shipped solo, faster, with Claude as the engineering partner.

Without AI

Director of CS hires 1 data engineer, 1 analyst, 1 CS Ops manager, 1 support engineer, and 1 BI partner. Six headcount, 12 months to stand up the same system.

AI-Native

Director of CS pairs with Claude across code, data, and strategy. Ships the same system solo. Compresses the org. Owns the outcome end-to-end.

SCOPE

Full lifecycle ownership across two product lines — MLM2Pro (golf) and Diamond Sports (baseball). NRR / GRR accountability. CS and Support team leadership. Direct executive reporting to the CEO. Every section below is work that AI made possible at a one-person scale.

01 / 05
Strategy & Program Design
Director-level
AI as thinking partner
The Four-Program CS Framework

LaunchPad (activation) → KeepSwinging (engagement) → NextRound (renewal) → Champions (advocacy). A retention architecture pressure-tested against behavioral data, not vibes.

Behavioral Retention Model

Reframed the company’s assumption that fast users retain. Mixpanel segmentation showed the opposite: converting users engage slowly and deeply, while speed-runners churn. Lifecycle journeys redesigned around depth.

Executive QBRs

Quarterly reviews to the CEO and leadership team on retention performance, reliability risk, and roadmap priorities — sourced from support, usage, and revenue data Claude helped synthesize into narrative.

02 / 05
Data Engineering
Normally a team
Shipped solo with Claude Code
Mixpanel → Iterable Real-Time Pipeline AI-Built

Stood up a Google Cloud Run integration that bypassed MuleSoft’s 24-hour batch delay. Behavioral events triggered lifecycle messaging the same day. Built with Claude as the engineering co-pilot — no data engineering team required.

Zendesk + Jira Unified Signal Layer AI-Built

Pulled support tickets and engineering escalations into a single customer signal pipeline feeding weekly issue reviews and closed-loop product feedback. Replaced manual cross-tool reporting that previously consumed half a CS Ops headcount.

Zendesk BigQuery Analysis Pipelines AI-Built

Built ticket classification and engineering escalation reporting for Diamond Sports — surfacing P0/P1 patterns, recurring root causes, and time-to-resolution drift. Direct SQL work with Claude pairing on schema and query design.

Zendesk WebView / Unity Integration

In-app support layer for the MLM2Pro mobile experience — closed the gap between product moment and support response.

03 / 05
Analytics & Modeling
Normally a data science team
Shipped solo with Claude
Churn Signal Model — Baseball AI-Built

Identified leading indicators of churn for Diamond Sports accounts. Built without a data scientist — Claude handled the feature engineering and statistical scaffolding.

Sales Call Predictive Analytics AI-Built

99% churn prediction and 96% success prediction accuracy on sales call conversation data — correlating spoken intent to downstream account health.

NPS / VoC Framework

Stood up the full system from scratch — instrumentation, reporting cadence, and qualitative analysis loop. Surfaced onboarding friction, activation gaps, and recurring failure modes for Product and Marketing.

04 / 05
Support Operations
AI in the response loop
Not just behind it
AI-First Tier 1 / Tier 2 Tooling

Deployed Claude-assisted support workflows across triage, response drafting, and knowledge base generation. Cut response times and accelerated KB coverage without expanding the support team.

Iterable Lifecycle Journey Redesign

Rebuilt MLM2Pro trial-user journeys around depth-of-engagement triggers — fewer messages, better timing, higher conversion. Designed against real behavioral cohorts from the Mixpanel pipeline above.

05 / 05
Executive Business Cases
Strategic deliverables
Modeled, written, defended
Legacy Device EOL Proposal

Made the business case for free Pro 2.0 hardware upgrades across 253 legacy-device accounts — balancing ethical renewal practice with ~$184K ARR protection on a sunsetting product line.

Diamond Sports Product Engineering Report

Comprehensive read on engineering escalation patterns, time-in-stage drift, and the product reliability story Engineering and Product leadership needed in one place.

Kelton AI Operations Lead Transition Doc

Authored the role positioning, scope, and capability stack for an AI Operations Lead role — translating CS lifecycle work into a cross-functional AI ops mandate.

STACK
Claude / Claude CodeMixpanelIterableGoogle Cloud RunBigQueryZendeskJiraSalesforceElasticsearchMongoDBUnity (WebView)SQL / Python
CLOSE

This is what AI-native actually looks like. Not the buzzword. The structural reality. The difference between a CS director who manages a team — and one who is the team, amplified across data, code, and strategy at the same time.