A lot of business advice is about finding new customers. But many businesses struggle because existing customers slip away quietly. This lesson covers two things. First, how the AI can help you spot a customer who is drifting away. Second, how to test a business plan against the costs you really pay.
The mistake most people make
Asking the AI to write a business plan and trusting the big forecasts it makes up.
“Write a business plan for an organic spice distribution company in Kerala with 5-year financial projections.”
“Year 1 revenue: ₹2 crore. Year 2 revenue: ₹12 crore. Year 3 revenue: ₹50 crore. With growing health awareness, your business will see rapid growth across South India...”
How it works: the kirana credit book
A kirana owner goes through his credit book on a Sunday. Sharma ji has bought about ₹4,000 of groceries every month for years. This month it was ₹800. The owner doesn't wait for Sharma ji to say he's leaving. He sends a WhatsApp: "Sab theek hai? A fresh batch of ghee has come in." The AI can help you spot the same drop across hundreds of customers.
Three ideas for keeping customers and planning
1. Early warning signs
Customers usually drift before they leave. They take longer to reply. The gap between orders grows. Their messages change from product questions to billing complaints. Paste your order history and ask the AI to list who shows these signs.
2. The profit on one sale
Start with one sale. Take the price, then subtract what you pay for the product, packing, delivery, platform or payment fees, and returns. What's left is what that sale really earns.
3. When the cash arrives
A business can make a profit on paper and still run out of cash. It happens when you pay suppliers quickly but customers pay you slowly.
Rules for a plan you can trust
Here are my figures. Use only these: """ - Selling price per unit, including GST: [₹ amount] - What I pay for one unit: [₹ amount] - Packing and delivery per unit: [₹ amount] - Payment gateway or distributor fee: [% or ₹ amount] - When customers pay me: [e.g. 45 days after delivery] - When I pay suppliers: [e.g. 15 days in advance] - Fixed costs each month (rent, salaries, software): [₹ amount] - Units I expect to sell each month: [number] """ Your task: 1. Work out what each unit earns after its own costs. Then work out how many units I must sell each month to cover my fixed costs. Show every sum step by step. 2. Work out roughly how much cash is tied up at the monthly volume I gave, because customers pay later than I pay suppliers. Show the sum. 3. Name the 2 costs above that would hurt most if they rose. Rules: - Do not invent any figure. If you need one I haven't given, write [CHECK: what's needed]. - Do not work out GST or any tax, and do not state tax, legal or lending rules. Where they matter, write [CHECK: ask my CA]. - End with this line: "Have a CA check these figures before you act on them."
The AI works only from your figures and shows every sum, so you can check each step. It marks gaps instead of filling them. Your CA confirms the result before you act.
Working File Templates
Copy or save these plain text files to use directly in your workspace.
An operational indicator table categorising customer health into Green, Amber, and Red tiers.
# Customer Churn Risk Matrix: B2B Accounts Review Cycle: Bi-weekly · Owner: Account Management Team ## Tier Definitions & Action Triggers | Risk Level | Trigger Indicators | Commercial Impact | Mandatory Intervention Protocol | | :--- | :--- | :--- | :--- | | Green (Healthy) | Re-orders placed within standard cycle (±5 days); payment received within agreed credit terms; WhatsApp responses within 4 hours. | Stable recurring revenue. | Standard quarterly satisfaction check-in. | | Amber (At Risk) | Re-order delayed by >15 days beyond usual cycle; order volume drops by >25%; payment delayed past credit limit. | Potential early warning of competitor trial. | Account manager calls within 48 hours to investigate stock levels or operational changes. No discounting. | | Red (Critical) | Zero orders for 60+ days; key contact leaves company; billing disputes unresolved >14 days. | Imminent account loss. | Operations lead visits facility or schedules director-level meeting to resolve outstanding grievances. | ## Notes Never offer automatic discounts to Amber accounts. In 80% of B2B cases, delays are driven by transporter issues or customer inventory backlogs rather than pricing.
Additional Prompts for This Lesson
Tailored prompts for specific workflows and departments.
Drafts a tailored re-engagement plan for accounts showing signs of cooling off before they churn.
You are a customer success operations assistant. Here is the account profile and usage history: """ [Paste customer name, order frequency history, average order value, and recent communications] """ Your task: 1. Identify the primary behavioural signal indicating churn risk (e.g. gap between re-orders, drop in order volume, unanswered check-ins). 2. Formulate a 30-day intervention calendar with three specific touches: - Day 1: A low-friction WhatsApp or email query focused on an operational problem (not a sales pitch). - Day 10: A high-value update or technical bulletin relevant to their past purchases. - Day 25: A leadership check-in from our operations head offering an account review. 3. Keep all communication short, respectful, and free of desperate discounts.
Systematises retention efforts so struggling accounts receive thoughtful intervention before canceling.
How can a business with a good profit on each sale still run short of cash?
Show the answer
Profit on paper ignores timing. If you pay suppliers in 15 days but customers pay you in 60, every new sale ties up more cash. Growing fast can make the gap worse.