Zac Perkins

The Problem

Coaching every CSM after every call wasn't scalable — I could only manually listen to a fraction of conversations, so most calls happened with zero structured feedback, and the coaching I gave depended entirely on whichever calls I happened to catch. No common rubric existed for comparing CSMs consistently.

The Solution

I built an AI-powered coaching scorecard combining Claude (rubric design) and Gong AI (automatic call scoring), evaluating every CSM-customer interaction against 10 weighted dimensions — rapport, active listening, problem resolution, communication clarity, product expertise, customer advocacy, follow-through, strategic guidance, and objection handling — with explicit 0–4 rating criteria written to reward real skill over performative habits.

Impact

The Problem

Two bottlenecks limited how proactively my team engaged customers: knowing which accounts were ready for a VP-level touchpoint meant manually digging through Salesforce every two weeks, and building a quality cold-outreach list meant manually cross-referencing CRM data, LinkedIn, and news for every contact.

The Solution

• CS Executive Connect — a fortnightly workflow that scans Salesforce for the right moment to reach out, ranks accounts by tier, identifies the most senior reachable contact, and drafts a personalized executive-voice email for review.
• AI Contact Mapping — an on-demand workflow that builds a 10-contact outreach map for any account, enriched with LinkedIn research and persona scoring, with personalized emails ready to send.

Both are built around hard quality gates: every email has to pass a "noise test" — would this be indistinguishable from the 100 other emails in this person's inbox? If yes, it gets rewritten.

Impact

The Problem

Static reports go stale the moment they're exported. Pipeline reviews, renewal health, and daily priorities lived in decks and spreadsheets that were already out of date by the time anyone opened them.

The Solution

• Renewals & Expansion Dashboard — live renewal and expansion pipeline, cross-referenced with support cases and call activity.
• Renewal Win/Loss Analysis — automatically classifies every closed renewal, separating real downsell from a quiet price drop.
• Weekly Pipeline Review — the full open pipeline categorized by urgency, with inbox signals layered in.
• Morning Dashboard — a single daily view of calendar, inbox, pipeline, and renewals, with one-click prompts into deeper AI workflows.

One architecture, four use cases — each dashboard calls Salesforce, Gmail, or Calendar directly every time it's opened. Nothing to rebuild by hand.

Impact

The Problem

Onboarding programs usually discover a customer is stuck at the QBR, or at renewal. Usage-based health scores make it worse: the two most common stall causes — a single-threaded champion and an implementation with no defined business outcome — both read as green. By the time a human notices, weeks of the time-to-value clock are gone.

The Solution

A prototype agent that reads usage, ticket, engagement, and stakeholder signals daily, then does three things a health score doesn't: it checks for blockers before classifying, so a security review doesn't get misread as disengagement; it names a specific stall pattern rather than emitting a risk number; and it assigns decision rights by reversibility — the agent owns observation, humans own consequence. Every draft carries an owner, a date, and the evidence behind it, and nothing customer-facing sends without a person.

Why it matters

The Problem

Risk was being caught inconsistently, defined differently by every team that touched an account, and reported up in whatever format someone had time to put together that week. A CSM's read on a shaky relationship and a hard financial risk nine months from renewal were getting treated the same way, and leadership found out about the largest losses close to when the customer did.

The Solution

I designed a two-tier risk framework, Risk Mitigation versus a confirmed Red Account, with entry and exit criteria set the moment a record was created, and a RACI model that gave every risk reason (competition, financial, organizational, product) an explicit owner across CSM, Renewals, Sales, and Solutions Engineering. Detection combined CSM judgment with automated engagement and risk-event triggers, unified into one dashboard that fed a weekly theatre-level review, a monthly Executive Red Account Summary, and Top 20 deal reviews at the CEO and Board level.

Impact