How EDT turned WhatsApp into a kitchen OS and resolved 82% of support
Nia handles EDT's LUMA and FLOW frontline on WhatsApp and web chat at an 82% AI resolution rate and a 94% AI deflection rate.
82%
AI Resolution rateAI resolution rate is the share of all conversations Sagepilot closed end to end without a human handoff.
94%
AI Deflection rateAn AI deflection rate is a customer support metric that measures the percentage of inbound inquiries handled by an AI system or self-service tool without needing a human agent.
Challenge
Most conversations arrive on WhatsApp, peaking at dinner time and running late into nights and weekends. The human team works 10:00 to 19:00, Monday to Friday. India cooks dinner after the team goes home.
Solution
Nia on the frontline across WhatsApp and edtworld.com, trained on the products, warranty policy, and EDT's communication standards, becoming Recipe by Ramsay the moment someone asks what to cook.
Results
- 82% AI resolution rate in a recent 30-day window
- 94% AI deflection rate
- Peak cooking-hour support that no longer waits for morning
Rethought from the ground up
EDT builds new-generation appliances for modern Indian kitchens: LUMA, a PureGlass air fryer whose transparent borosilicate bowl lets you watch food cook, and FLOW, a PureGlass precision kettle. The Mumbai company, founded by Naiyya Saggi and Vyasateja Rao, designs everyday devices to be rethought from the ground up, not merely upgraded.
That kind of product attracts a high-intent customer: someone who reads the spec sheet, asks about materials, registers the warranty, tracks the parcel, and then wants to know what to cook tonight.
India cooks dinner after the team goes home
EDT's research found the real friction in cooking is not execution, it is deciding: working out what to cook with the ingredients, time, dietary preferences, and people in front of them. That created a support surface most appliance brands have never staffed for.
Alongside warranties, product specs, order tracking, and service requests, most conversations arrive through WhatsApp, with demand peaking at dinner time, staying active late into the night, and continuing across weekends. The human team works 10:00 to 19:00, Monday to Friday.
Nia, and Recipe by Ramsay
Sagepilot put Nia on the frontline across WhatsApp and the edtworld.com widget. She is trained on LUMA, FLOW, warranty policy, order and shipment workflows, and EDT's communication standards: complete words, complete sentences, warm explanations, and calm issue handling.
Nia checks Shopify for order context, Shiprocket for live shipment status, registers marketplace warranties, and routes judgment-heavy cases with the full story attached. When someone asks what to cook, she becomes Recipe by Ramsay, the LUMA recipe specialist EDT announced as India's first AI kitchen OS on WhatsApp: no app download, no onboarding, no context switch, just a message to the interface customers already use.
The kitchen OS numbers
In a recent 30-day window, Sagepilot helped EDT reach an 82% AI resolution rate and a 94% AI deflection rate. Nia carried the frontline on WhatsApp and web chat: warranties, tracking, product questions, safety triage, and Recipe by Ramsay, with Shopify and Shiprocket connected so she could finish the work instead of only answering it.
- Dinner-time, late-night, and weekend demand no longer waits for morning
- The human team keeps the judgment cases: color changes, address changes, service escalations
- 82% resolution and 94% deflection in a recent 30-day window
“The next frontier of AI is Physical AI, intelligence embedded directly into the products people use every day. Recipe Ramsay is not a generic AI layered onto an appliance, but a tightly integrated intelligence layer that enhances a user's experience with LUMA's core performance while reducing everyday cognitive load. By delivering it through a familiar interface like WhatsApp, we're moving beyond novelty toward infrastructure, where technology works quietly and meaningfully inside the home.”
Dinner decided before LUMA preheats
Recipe by Ramsay is the single largest reason people message EDT. Nia asks the questions a chef would, then returns a LUMA-calibrated recipe with mode, temperature, timing, and the preheat step already built in.
LUMA · Roast mode
180°C · 15 minutes · Jain paneer tikka
The missing ingredient becomes a next step
Most cooking apps stop when a customer lacks an ingredient. Nia keeps the meal moving with a realistic substitute, then offers a live comparison link so discovery, decision-making, and sourcing stay inside one WhatsApp thread.
- Substitute suggested without breaking the recipe
- Ingredient comparison link prepared
- Decision, sourcing, and cooking stay in one thread
Warranty, tracking, and handoff in one exchange
Nia resolves the repetitive post-purchase load herself: marketplace warranty registration, website order tracking, Shiprocket shipment checks, and the careful cases she should not decide alone, routed with full context attached.
Marketplace warranty
Channel identified, serial and purchase details collected, warranty recorded.
Order tracking
Shopify order found, Shiprocket AWB read, live shipment status sent back.
Color change
Origin Orange request routed to the specialist team with the context attached.
When tone is a policy
Safety questions are where a generic assistant can do the most damage. Nia is trained to stay calm, avoid alarmist wording, collect only the right diagnostic details, and transfer physical issues without making the customer repeat themselves.
- Calm-tone issue playbook applied
- Only relevant troubleshooting questions asked
- Physical fault escalated with context
The busiest hour is covered
EDT's peak is 6 PM, activity stays high late into the night, and demand continues across weekends. Nia keeps order, recipe, and support answers moving when the human team is offline.
Ordered Origin Orange last week. Any update? Also can I still change the color?
- Shipment status checked from the live AWB
- Color change routed to EDT specialist team
- Customer answered while the team was offline
6 PM
busiest hour
Nights
and weekends covered
Going live
Phase 1 · Nia on WhatsApp
EDT put one AI teammate on the busiest channel, trained on LUMA, warranty policy, order flows, shipment status, and the brand's written communication standard.
Phase 2 · Recipe by Ramsay
Customers text Hi or Recipe and get LUMA-calibrated recipes with mode, temperature, timing, preheat steps, and dietary context.
Phase 3 · Commerce in the conversation
Missing ingredients became substitutes and buying links; product curiosity became live demos, pre-booking, and The Edit community invites.
Today · Kitchen OS
In a recent 30-day window, an 82% AI resolution rate, a 94% AI deflection rate, and peak cooking-hour support that no longer waits for morning.
More stories
Join the brands that run their customer experience on Sagepilot
Live in about 48 hours, across every channel your customers use.