The 90-Minute Account Health Review (Powered by AI)

What you’ll have at the end

• A 6-slide QBR you can send the client the same day
• One-page account snapshot (revenue, margin, volume, health score)
• AI-generated summaries: top ticket themes, NPS verbatims, risk signals
• A ranked “Next 3 Fixes” list with owners, deadlines, and expected impact

What you need (15 minutes prep)

Data (last 90 days unless noted):

• Revenue & margin by month and by service/SKU
• Tickets/cases (created, closed, reopen rate, SLA breach, tags/labels)
• NPS/CSAT scores + verbatim comments (last 6–12 months)
• Product/usage (logins, active users, feature adoption, uptime)
• Contract (term, auto-renew date, expansion opportunities)

AI setup (tool-agnostic):

• One workspace/chat with access to your CSVs or a live connection
• These 5 analysis prompts (save them as snippets)

P1 – Ticket Themes
“You are a support analyst. Cluster these tickets by issue theme. For each cluster give: name, % of total, trend vs prior 30 days, avg time-to-resolve, SLA breach rate, representative example.”

P2 – NPS Themes
“You are a voice-of-customer researcher. Extract top positive/negative themes from NPS verbatims. Quote 1–2 short lines per theme and map each to product area.”

P3 – Churn Risk
“You are a retention analyst. Using usage drop, SLA breaches, unresolved P1s, and negative sentiment, score churn risk 0–100 and explain why in two bullets.”

P4 – Revenue Snapshot
“You are a CFO. Summarize monthly revenue, gross margin, and variance vs plan. Flag expansion or contraction risk. One paragraph plus a 4-row table.”

P5 – Next 3 Fixes (RICE)
“You are an operations leader. Propose 5 fixes. Score with RICE (Reach, Impact, Confidence, Effort). Return the top 3 with owners, ETA, and expected outcome metric.”

The 90-Minute Runbook

0–10 min: Load & frame

• Export CSVs (tickets, NPS, revenue, usage).
• Tell the AI the account context (industry, footprint, contract dates, VIP stakeholders).
• Paste quick account goals (e.g., “Renew Q2, expand seats +15%”).

10–30 min: Support signals

• Run P1 – Ticket Themes on the ticket export.
• Ask for trend deltas (last 30 days vs prior 30).
• Follow-up: “Show the 3 fastest rising themes with sample tickets and root-cause guesses.”

Decision: pick 1–2 themes to turn into fixes.

30–45 min: Customer sentiment

• Run P2 – NPS Themes on verbatims.
• Follow-up: “Which negative themes overlap ticket clusters? Rank by frequency × recency.”

Decision: choose the one VOC theme to address this quarter.

45–60 min: Commercials & usage

• Run P4 – Revenue Snapshot on revenue/margin export.
• Feed usage data (logins, adoption) and ask:
  “Correlate usage with NPS and renewals. Are we expanding or shrinking users per site?”

Decision: flag expansion levers (features with highest stickiness) or contraction risks.

60–70 min: Churn risk

• Run P3 – Churn Risk using all summaries.
• Quick sanity check: adjust the score ±10 if you know context AI can’t see (e.g., exec sponsor change).

70–85 min: Next 3 fixes

• Run P5 – Next 3 Fixes (RICE) across the problem set you’ve identified.
• Confirm each fix has: owner, ETA, metric (e.g., “Reduce SLA breaches on ‘Access Issues’ from 14% → <5% by Feb 15”).

85–90 min: Assemble QBR deck

Use this 6-slide skeleton (copy/paste into your deck tool):

1. Account Snapshot – Revenue (L3 months), margin, usage, NPS, health score
2. What’s Working – 3 bullets with proof charts
3. What’s Not – top ticket/NPS themes with trend arrows
4. Churn/Expansion Outlook – risk score, why, and expansion levers
5. Next 3 Fixes – RICE table with owners & dates
6. Asks & Dates – decisions needed, renewal/PO timeline, next QBR date

Templates you can reuse

A. One-page Account Snapshot (paste into a doc)

• Revenue (Last 3 mo): $___ → $___ → $___ (Δ vs plan: ___)
• Gross Margin: ___% (Δ MoM: ___pp)
• Usage: DAU/MAU ___ / ___ (Δ: ___), Feature X adoption ___%
• Support: __ tickets opened / __ closed, SLA breach ___%, Reopen ___%
• NPS: ___ (Promoters ___%, Detractors ___%) • Top themes: [A], [B]
• Health Score (0–100): ___ • Risk: Low/Med/High
• Key Dates: Renewal //__, Expansion target: ___ seats

B. RICE Table (for “Next 3 Fixes”)

(RICE = Reach × Impact × Confidence ÷ Effort)

Quick charts to drop into your deck

• Tickets by Theme (stacked bar): Top 5 clusters, last 90 days by week
• NPS Over Time (line): With call-outs for feature releases or outages
• Revenue & Margin (combo): Bars for revenue, line for margin %
• Usage vs Renewals (scatter): Accounts or sites plotted, trendline and target band

Owner ops tips (so this sticks)

• Standardize inputs: Same 4 CSV exports every QBR; name them predictably.
• Version control your prompts: Keep P1–P5 as shared snippets; refine as you learn.
• Metric contracts: Every fix must tie to one metric that will appear on slide 1 next QBR.
• Close the loop: Send a 2-paragraph recap and the 3 commitments within 24 hours.
• Cadence: Run this 90-minute review monthly; do a deeper dive quarterly.

Email template to send with the deck

Subject: [Client] – 90-Day Health Review & Next 3 Fixes

Hi [Name],
Attached is your 6-slide health review. Highlights:

• Support volume is stable; Access Issues trending up (+22%); SLA breaches at 9.4%.
• NPS is 41; detractor comments center on onboarding clarity.
• Usage dipped 7% at two sites after the admin change.

Next 3 Fixes (with dates):

1. Tighten access SOPs & alerts (breaches <5% by Feb 15 – Maria)
2. Revamp onboarding email sequence (DAU +10% by Jan 30 – Dev)
3. Quarterly admin training cadence (Reopen −30% by Mar 1 – CS)

Can we confirm these owners/timelines and book the next QBR for [date]?
—Alex, BoostMyAI

FAQ

How do we compute the health score fast?
Start simple:
Health = 40% Usage Trend + 25% NPS + 20% SLA Compliance + 15% Revenue Variance
Normalize each 0–100; adjust weights over time.

What if we don’t have NPS?
Use CSAT and ticket sentiment from verbatims. Flag “NPS data gap” as a fix.

Can we automate the whole thing?
Yes—schedule the exports, run a notebook or workflow that feeds the same prompts, and push the 6 slides to a template. Keep a human in the loop for the “why” and the “asks.”

Bottom line: With tight inputs and reusable prompts, any owner can turn raw ops data into a crisp, client-ready QBR in 90 minutes—and, more importantly, leave with three concrete fixes that move the metrics that matter.

Contact Us
Questions or want this for your business? Book a 15-min consult. Contact us!