ChannelX Logo
Search
Close this search box.

Why it’s time to optimise for Amazon Rufus AI search

Category: Data & Insights
Why it's time to optimise for Amazon Rufus AI search

Following on from our Leaders Interview with Max Sinclair, Founder and CEO at Azoma, we can report on some interesting findings on the uptake of Amazon Rufus, the marketplace’s new generative AI shopping assistant.

We know that shoppers are starting to use AI for shopping, Perplexity launched their product search and then ChatGPT launched AI Shopping in May this year, but Amazon is already replacing search engines such as Google and Bing as many consumer’s default shopping engine… but are shoppers opting to use Rufus instead of the default Amazon text, barcode or image search?

Amazon’s only comment is that “hundreds of millions of purchases” have been powered by Rufus, and although that sounds like a big number, what does it really mean in terms of total purchases on Amazon? Well Max and Azoma have run the numbers for us in this blog post and the results are that:

  • Amazon has revealed that during Prime Day they had to allocate 80,000 AWS Inferentia and AWS Training chips, collectively processing 3 million tokens per minute.
  • Assuming the average query length on Rufus is similar to that on Perplexity, 10-11 words, and each word comprising approximately 1.5 tokens (common in natural language processing), this results in an average of 15.75 tokens per query
  • Given Rufus’s infrastructure was built to process 3 million tokens per minute, Rufus has the a capacity for 274.3 million queries per day
  • Amazon receives approximately 2 billion searches per day, so Rufus accounts for about 13.7% of searches

If almost 14% of searches on Amazon were powered by Rufus back in October, that’s likely to be an even bigger number today and strongly suggests that brands need to be thinking about AI optimisation… and not just on Amazon but for all your products everywhere they are sold. AI wants rich contextual information and this can be text but it can also be images – wave bye bye to product shots on a white background, AI wants lifestyle shots to infer product usage whether this be in the home (which room?) the garden, or the office.

If you want to know more, watch our ChannelX Leaders Interview with Max and if you want to know even more he’ll be happy to chat!

RELATED POSTS..

Amazon get strict on Reference Pricing

Amazon get strict on Reference Pricing

AI-caution prompts rising demand for authentic, human-first content in social buying journeys

AI-caution prompts rising demand for authentic, human-led content in social buying journeys

Per item charges coming for Amazon FBA removal and disposals

Per item charges coming for Amazon FBA removal and disposals

AI lessons US retailers can teach the UK

Across the Atlantic: AI lessons US retailers can teach the UK

1.5% Amazon Fuel surcharge for FBA/MCF from 17th April

1.5% Amazon Fuel surcharge for FBA/MCF from 17th April

Latest

Take a look through a selection of the latest articles on ChannelX

Register for Newsletter

Receive 5 newsletters per week

Gain access to all research

Be notified of upcoming events and webinars