Every festival sale has the same shape: marketing spends weeks preparing the offers, ops preps inventory - and support walks in on day one to find 300 unread messages. Query volume during Diwali week routinely runs 5–10x normal, and it's the one time of year you can least afford slow answers, because every question has a cart behind it.
Why surge queries are the easiest to automate
Festival-week questions are the most repetitive of the year: "Is the offer valid on this product?" "Will it arrive before Diwali?" "Can I combine coupons?" The same five questions, hundreds of times. That's exactly the pattern an AI agent handles best - if the answers exist in its knowledge base before the surge starts.
The pre-sale checklist
One week before the sale, upload a single "festival FAQ" document: offer terms and validity dates, coupon stacking rules, extended delivery timelines by region, the return window for sale items, and COD limits. Write dates explicitly - "valid until 12 November" beats "valid till the festival" - because your customers will ask on the 11th at 11pm.
Then test like a suspicious customer: ask about combining offers, ask about a city with delayed delivery, ask what happens if the gift arrives late. Every gap you find now is a hundred escalations you won't get during the sale.
A knowledge-base hour in October is worth a support-team week in November.
During the surge: watch the gaps, not the volume
Once the sale is live, the dashboard number that matters isn't total conversations - the bot absorbs those. Watch the knowledge gaps: questions the bot couldn't answer. Each one is usually a missing sentence, not a missing team member. Add the answer, and the next hundred customers asking the same thing get it instantly.
After the festival, the same document becomes next year's template. The surge stops being a crisis and becomes a checklist item - which is exactly what it should have been all along.