Developing
Opinion column: survey samples are mostly not given, the Amazon ad revenue figure is uncited and the apparel scenario is hypothetical.
In 60 seconds
- Sensor Tower data (via CBS News) is cited for Meta's Muse reaching 2.5 million downloads in two weeks.
- Checkout.com 2026: 72% of UK and US merchants expect consumers to adopt agent-led shopping faster than merchants are ready for.
- Deloitte (13,500 European consumers): 31% would let AI make the final purchase and 47% would not.
- Accenture: 56% would have an agent recommend brands; 37% of behaviourally loyal consumers would accept an agent switching them.
Impact Layer
- What changed
- Marketers are being told to plan for consumer agents that choose, compare and buy.
- Why it matters
- If an agent picks the product, brand building and product data must work for machines as well as people.
- Who is affected
- Brand, e-commerce and loyalty teams.
- Martech stack area
- SEO & Search, Commerce
- Stack impact
- Make policies such as returns, pricing and availability consistent and machine-readable across your site and feeds.
- Business impact
- Low Contributed opinion on MarTech (owned by Semrush). Only Deloitte's sample size (13,500) is given; the author admits he does not know Target's decision-making, and forecasts are unquantified.
- Related technologies
- What to watch
- Independent data on agent-led purchases and platform liability rules.
Greg Kihlstrom, principal at The Agile Brand, argues in a MarTech column published October 13, 2026 that consumer AI agents are shifting control of the customer relationship. The page says the views are the author’s own.
Data cited
- Sensor Tower (via CBS News): Meta’s Muse agent reached 2.5 million downloads in two weeks.
- Checkout.com 2026: 72% of UK and US merchants expect consumers to adopt agent-led shopping faster than merchants are ready for.
- Deloitte (13,500 European consumers): 31% are open to letting AI make the final purchase; 47% are not.
- Riskified (US and UK shoppers): 50.8% say the AI platform should be liable for unauthorised purchases; 23.2% say the retailer; 18.7% say the consumer.
- Accenture: 56% would have an agent recommend brands; 37% of behaviourally loyal consumers would accept an agent switching them.
The argument
Brands can still compete by making policies consistent, machine-readable and portable through loyalty programmes.
Caveats
Sample sizes and methods are missing for most surveys. The Amazon ad-revenue figure ($68 billion) is uncited, and the Target example rests on one reporter’s test. The apparel scenario is hypothetical and “coherence at velocity” is the author’s own framing.
Our read
The survey numbers are soft, but the practical request is solid: be consistent and readable for a machine.
Sources
- MarTech: AI agents are coming between you and your customers (opens in a new tab)
Source for the column's claims and the survey figures it cites.