Analyst: AI Agents to Bolster, Not Bypass, Platforms
Industry analyst Eric Seufert pushed back against a report suggesting AI agents will disintermediate commerce platforms like DoorDash and Uber. He argued these platforms are well-positioned to benefit by embedding AI agents themselves, leveraging their aggregated user attention. This view was echoed by Dan Hockenmaier, who noted that delivery apps maintain their commissions by ensuring service reliability.
- The concept of "agentic commerce" suggests AI agents will autonomously handle shopping for consumers, from product discovery to purchase. This could shift the consumer-brand relationship, with AI agents, not branded apps or websites, becoming the primary interface for commerce. - A recent note from Citrini Research projected that AI agents could transact directly with restaurants at a significantly lower take rate (around 7%) compared to DoorDash's current blended cut of about 25%, and then separately auction delivery jobs to drivers, which would compress platforms' margins. This analysis contributed to a 7% drop in DoorDash's stock price on the day it was released. - [Both DoorDash and Uber](https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEOYtlbMQo_a1VsDtuE_zo7stc3SPGodfUGVuyNhMNT_vo8LtekPvabfFTYC_Bov8hHP1BHaPIlGp7qBQZqpBuZ2HhnL33puoqDUqkfpheKgDUQ123HayT_aMFjGFBTKa-slqoCn7pkFbsvORporkV7DqYS7Mkx5G7zDyb_TjAMdvx0UeLbORZz18I7e2r_qQ57eweFsJLv-gl33kku) are actively integrating AI to improve their existing services. DoorDash has implemented AI-powered voice ordering to handle the 50% of customer calls that restaurants were previously missing, while Uber's CEO has confirmed the development of an in-app AI chatbot. - Platforms like DoorDash and Uber utilize AI for more than just customer-facing features. AI algorithms are used for order assignments, route suggestions, and fraud detection, such as identifying fake delivery photos or unusual GPS patterns. - The potential for disintermediation extends beyond food delivery to ride-sharing. Agentic AI could compare options across Uber, Lyft, and other services to select the cheapest or fastest ride, diminishing the value of any single platform's app. - A viral Reddit post in early 2026, which was later debunked as a likely AI-generated hoax, alleged that a food delivery app used a "desperation score" to offer lower-paying orders to drivers who frequently accepted cheap runs. Both DoorDash and Uber Eats denied the claims. - Some analysts argue that the "DoorDash problem" is overstated for food delivery and ride-sharing marketplaces. They contend that the complexity of managing service delivery, logistics, and risk provides a significant moat that AI agents are unlikely to replicate themselves. - Major retail and e-commerce platforms are unlikely to allow independent AI agents to disintermediate them and will instead build their own proprietary agents to maintain control over the user experience and their valuable advertising businesses.
