Why are we doing this?

E-commerce started in 1994.

32 years have since been spent on the art of driving e-commerce traffic and maximizing revenue per visit.

Buying clicks for e-commerce costs $350bn+ per year.

A further ensemble industry of creative professionals, media buyers, CRO specialists etc. has perfected the art of maximizing the return-on-ad-spend. Success often meant perfecting ROAS, not necessarily offering the best value.

CX has been perfected.

Sites are optimized for speed, with 0.1s page load improvements increasing conversion and spend per visit by 8.4% and 9.2% respectively. Products are ordered and priced to maximize the cross-section between consumer preferences and merchants’ business objectives.

Check-out is optimized.

Amazon perfected the 1-click check-out. Customers can buy now and pay later (16% do so in the US, and 8% in the EU). Late fee disclosure i.e. drip pricing added an extra 15-21% of spend per customer (controversially so).

It’s time to optimize for agents.

By 2030, 10-20% of e-commerce GMV will be AI agents, driving $190-385bn of revenue in the US alone (Morgan Stanley). 15% of US consumers already completed an online purchase directly through AI. At 23% the same figure in China is even higher.

Agents and humans shop distinctly.

Optimizing pricing, product descriptions and visuals, and UX is fundamentally different when selling to AI agents. This is not a general statement about an industry in flux. We derived this conclusion after running 10,000+ agent shopping simulations. Read more about our experiments here.