Blog

AI agents, RAG, and integrations — notes from production.

AI Support for Health and Pharmacy E-commerce: Compliance and Escalation

AI support for pharmacy e-commerce can handle order and prescription-status logistics, but every clinical question has to route to a pharmacist.

AI Support for Ticketing and Events: Transfers, Refunds, and Postponements

AI support for ticketing works against event status and time pressure - transfers, access issues, and refunds that are factual once the event state is known.

AI Support for Digital Goods and Software: Licenses, Refunds, and Access

AI support for digital goods runs on entitlement state - activation, seat limits, usage - where refunds are a factual question and abuse is a real risk.

AI Support for Furniture and Big-Ticket Home Goods: Delivery and Assembly

AI support for furniture retail centers on delivery scheduling and lead times, with damage claims and big-ticket refunds routed to a person for assessment.

AI Support for Consumer Electronics: Warranty, Setup, and Troubleshooting

AI support for consumer electronics is half grounded retrieval - warranty and troubleshooting steps - and half transaction, with a clear safety boundary.

AI Support for Travel and OTAs: Changes, Cancellations, and Disruption

AI support for travel booking has to work against live itineraries and supplier rules the agent does not control, with irregular operations going to a person.

AI Support for Food and Grocery Delivery: Live Orders and Refunds

AI support for food delivery runs against a live order with a minutes-long resolution window - real-time status, fast validated refunds, and a safety boundary.

AI Support for Beauty and Cosmetics: Ingredients, Shades, and Allergies

AI support for beauty e-commerce can ground shade and ingredient questions in product data, but reaction and allergy concerns must route to a person.

AI Support for Fashion Retail: Sizing, Returns, and Exchanges

AI support for fashion retail lives on two queues - sizing questions before purchase and returns or exchanges after - each with its own grounding discipline.

Voice AI Fallback to IVR and Human: Designing the Degradation Path

Good voice AI fallback design is a ladder - native speech to cascaded, to a narrow IVR, to a human with context - where every rung has an exit to a person.