Seems like not a day goes by when there isn't sensational news in the world of AI and perhaps every other day in the world of Agentic Commerce.
Agentic Commerce is a broad term and in the B2C context is basically your personal AI agent who has context around your buying habits and needs and also has all your details in terms of addresses, payments etc and can do your shopping for you.
In the B2B context, Agentic Commerce enters the realm of checking out on a surface that is not the merchant website or an app. This has been described well in Scot Wingo’s post here
Having Co-founded 2 startups and taken them to exits in the Saas/Logistics/Ecommerce space, I have had the good fortune to work with a lot of enterprise retailers in various verticals like Fashion, Apparel, Beauty, Hardware, Grocery, Alcohol etc. When I started building CartAI , I wanted to use all my insights in working with these top retailers and analyze some potential pitfalls that shopping agentically “off property” may pose.
It is inevitable that we will have checkout within the AI search experience. We have already seen some of that take shape at Perplexity(buy with Pro, Paypal partnership etc) and some news coming out of OpenAI that covers this. These new AI based discovery engines will cross the chasm from being affiliates that send utm_source web traffic to channels that will generate orders for the merchants/Retailers. With this new channel, the complexity associated with a new channel for a merchant/retailer doesn’t just go away. Below are a few traditional pain points
Integration : Integrating and maintaining that integration takes dedicated resources. This is on top of maintaining the retailers own e-commerce stack, existing channels and marketplaces and integrating inventory across stores, DC’S etc
Order management: Syncing of yet another channel with existing Order management system or managing these orders outside of that is an additive pain point.
Financial reconciliation with the channel if channel is merchant of Record.
Post purchase, Customer service , Returns for these orders are the merchant’s headache or a parallel stack has to be setup by the channel where the orders are being sourced from to deal with this.
Customer data and Loyalty: This is an age old debate between channels, marketplaces and merchant/retailers. Merchants/Retailers want first party data period. Loyal customers who can be marketed to are worth way more than a one time order.
We are seeing that an increasing number of searches are now starting on AI platforms. These eyeballs are valuable and inevitably will be monetized in various ways with E-commerce conversion being one of them.
So then, given a clean sheet of paper, how would you design an ideal solution that harnesses the power of Gen-AI discovery but also incorporates all the E-commerce lessons learned till date. Below are a few key areas to consider
Friction Free consumer experience : The “click out” has been a necessary evil in E-commerce. Even before this Gen-AI revolution, product discovery and conversion were increasingly happening in separate properties(unless you are a marketplace like Amazon). This divergence has been accelerated with the advent of AI based discovery. The ability to Checkout at the point of discovery/inspiration certainly reduces the friction associated with clicking out of the present experience to go on to a merchant/retailer web/app property.
Payments : It is imperative that retailers/merchants are “Merchant of Record”. This means that the payments are being cleared directly by the merchant payment processors. All this ofcourse needs to happen securely with PCI compliance and other security features that are table stakes in any Ecommerce payment scenario
Integration : The ideal solution would not have any merchant/retailer side api integration yet should be able to inject orders taken into retailer OMS systems so they can be fulfilled like all other native retailer orders without special gymnastics on the retailer end for orders coming from this new agentic channel. To really be able to scale this paradigm globally, retailer specific integrations become a barrier to scale. Platform specific integrations are interesting and get you more merchants with one integration but nailed up api integrations seem archaic in the brave new world of AI. MCP servers?…OK very interesting and definitely something to consider as they get adopted more by platforms and individual retailers. See nekuda post on Shopify MCP rollout
Loyalty : The ability to let consumers harness the loyalty programs of their favorite merchants while having the convenience of checking out at the point of discovery would be a great value add to not only the consumers but also to the merchants/retailers.
Customer service, post purchase and returns : If the orders coming in are““merchant native” as described above it will be easy for the merchants to use their existing processes for customer service, post purchase and returns.
We truly believe that incorporating all the above points in a solution would be the perfect balance between harnessing the power of AI based discovery and making the solution scalable and retailer friendly. This is what we at CartAI have set out to do and this post describes some of our founding principles for our product. More to come!
