“You left something in your cart.”
It repeats the event. The customer may still have exactly the same unanswered question.
Kahol Labs / Concept case study
Meet Kahol Saathi. A sales recovery concept for Indian D2C brands, designed to help interested shoppers move from “not sure” to a more confident purchase decision.
A working, scripted prototype. Shopify and WooCommerce integrations are proposed; no customer results are claimed.
01 / The problem
A shopper sees your ad, likes the product, chooses a colour and starts checkout. Then the journey stops. The ad did its job. The buying decision is still unfinished.
For a fashion brand, that hesitation might be a sizing doubt. For another customer, it might be fabric, delivery timing or the return policy. Sometimes the interruption is simpler: a phone call, a failed payment or a moment of distraction.
Buying more traffic does not directly resolve those questions. A recovery experience should understand where the shopper paused, offer relevant help and make the next step easy—while recognising that some people simply do not want to buy.
That is a different problem from “Remind me this product exists.”
unfinished checkouts
in our example
Imagine 1,000 shoppers starting checkout, with only 100 completing an order. The remaining 900 are unfinished journeys. It is tempting to call this “90% of ads wasted”, but ad spend and checkout abandonment measure different things.
Some visitors were browsing. Some will return by themselves. Some cannot be contacted. The practical question is: which eligible shoppers could benefit from a useful answer now?
For context, Baymard’s cart-abandonment research reports an average around 70% across its collected studies. That is cart abandonment, not an Indian D2C visitor conversion benchmark. Our 90/10 split is an example.
02 / The product thesis
Kahol Saathi is built around a simple hypothesis: timely, relevant answers may remove some of the friction between product interest and a completed checkout.
“You left something in your cart.”
It repeats the event. The customer may still have exactly the same unanswered question.
“Need help choosing the right size?”
It opens a path to understanding the hesitation. Product facts, policies and a clear checkout step support the decision.
Help the shopper interpret a verified size chart. Ask for the missing measurement instead of inventing a fit guarantee.
Answer fabric, delivery and return questions from store-approved information. If a detail is missing, say so.
Resolve the actual doubt first. Any price adjustment would need explicit merchant approval; the demo never invents a discount.
03 / The proposed journey
Store events or checkout records identify an incomplete purchase. A production integration would verify incoming events, handle repeats safely and check contact eligibility.
The example waits 15 minutes. That timing is a testable starting point, not an established optimum. Payment and contact status must be checked again before outreach.
A follow-up invites a question. The conversation can then focus on size, material, delivery or returns using the relevant product and approved policies.
If the shopper is ready, offer the appropriate secure checkout. If they want a person or ask to stop, respect that choice and end automated recovery.
A reply, click or promise to buy is not recovered revenue. A live implementation would wait for the store’s authoritative order/payment event.
04 / Explore the business case
Move the assumptions to explore the opportunity. The 40–50% recovery range is an ambitious scenario to test—not a measured Kahol Saathi result, forecast or guarantee.
900 abandoned × 60% eligible × 45% assumed recovery = 243 orders.
Order value is before discounts, returns, taxes, product costs and operating costs. Some shoppers would return without help. A comparison group is needed to measure additional orders and profit attributable to recovery. The ratio shown is a scenario ratio, not proven advertising performance.
The model uses checkout starters consistently. It does not equate checkout completion with visitor conversion. Read Shopify’s guide to recovery-rate measurement for the underlying distinction.
05 / Experience the product
Meet Neer Studio, a fictional apparel store. Act as the merchant and the shopper: start recovery, fast-forward the waiting period, resolve a question and simulate the purchase.
Try the full journey: start recovery, advance the waiting period, ask a product question, then simulate a purchase.
Neer Studio · Fictional apparel store
Sage · Size M · Qty 1
₹1,499Saathi checks payment again before starting the conversation.
Sample product facts: 100% cotton; M chest 38–40 inches; L chest 40–42 inches; ₹1,499 including standard delivery; estimated delivery 4–6 business days; return request within 7 days for unworn items with tags. These are fictional demo policies, not Kahol Labs offers.
AI shopping assistant concept · Neer Studio
Start recovery in the store view to see the customer experience here.
No payment details needed. This simulates a confirmed order.
Use sample questions only. Replies follow predefined product facts; they are not generated by a live model.
Designed around trust
Try “paid elsewhere” before outreach. The demo cancels recovery instead of chasing a completed purchase.
Size, fabric and policies come from the sample catalogue. Unknown questions receive an honest fallback; a human request ends the automated flow.
A conversation is not a sale. Session recovery value changes only after a simulated payment confirmation.
06 / A familiar buying moment
The sample shopper has a 41-inch chest and a size M kurta in the cart. Another reminder alone does not tell them whether it is the right choice.
In the demo, Saathi uses the sample chart: M covers 38–40 inches and L covers 40–42 inches. It can explain the range, suggest L for that measurement and update the sample cart when the shopper agrees. The shopper can also ask about the cotton fabric and the store’s return window.
This is the experience we want to test: relevant information that makes a decision easier. If the question goes beyond the available facts, the assistant should offer a person rather than invent an answer.
Even a good conversation can end without a purchase. The user might decide against the product, choose another store or ask to stop. A credible recovery product respects all of those outcomes.
07 / From prototype to product
Proposed production architecture for Shopify and WooCommerce. The current interactive page demonstrates the experience; it is not connected to either platform.
Signed event ingestion, a checkout/order state model and duplicate-event handling. A payment arriving during the wait should cancel the scheduled recovery.
Product variants, stock, size charts, shipping and return policies retrieved from approved sources. Price changes and other actions stay behind merchant-defined permissions.
In a production build, AI could help interpret questions and draft grounded replies. Missing facts, special requests and complaints need an accessible human handoff.
Order confirmation, opt-outs, message caps, conversation review and access controls. Production readiness requires integration testing and operational monitoring.
Already running a Shopify or WooCommerce pilot? Access live abandoned checkouts, Meta WhatsApp conversational threads, and the autonomous Gemini 2.0 sales brain directly.
08 / Prove the value
The next milestone is evidence from a controlled pilot. A large recovery number in a calculator is not evidence that the product works.
Measure eligible abandoned checkouts, current return-to-purchase behaviour, order value and the most common customer questions. Separate first-time visitors from checkout starters.
Randomly reserve a comparable group that does not receive the new recovery experience. This helps estimate how many orders would have happened anyway.
Compare additional completed orders and contribution after messaging, model costs, discounts, refunds and returns. Watch opt-outs, complaints and handoff quality alongside sales.
More profitable completed purchases than the comparison group, with acceptable customer experience and operating costs. The result needs to be measured—not assumed from a 40–50% scenario.
09 / Good questions
No. This is a concept case study with a working scripted demo. The percentage is an adjustable, ambitious modelling assumption. No live recovery experiment or client result is presented here.
No. Ad spend, visitor conversion and checkout abandonment have different denominators. The case study uses a hypothetical group of checkout starters to illustrate the unfinished journey.
The demo uses predefined answers from a fictional product catalogue. It shows the proposed experience, state transitions and safeguards without sending messages, accessing a real store or taking payments.
The demo does not offer discounts. In the proposed product, any discount or order-changing action would need explicit merchant rules and authorisation.
The demo lets you simulate payment elsewhere, which cancels recovery. A human handoff ends automated conversation, and an opt-out stops further follow-up. These behaviours would also need verification against real store events in a production integration.
Kahol Saathi / A Kahol Labs concept
A thoughtful recovery experience starts with the shopper’s question. The opportunity is to answer it well, respect the decision and measure what changes.
