Bain & Company

Agentic Commerce: Making Loyalty and Promotions Agent-Ready

This sponsored Digiday guide (by Talon.One) explains 'agentic commerce'—autonomous AI agents that research, apply incentives and complete purchases on behalf of consumers—and why brands must make loyalty and promotions machine-readable and agent-visible. The piece documents three emerging protocols: OpenAI’s Agentic Commerce Protocol (ACP, Sept 2025), Google’s Universal Commerce Protocol (UCP, Jan 2026) and Talon.One’s Unified Incentives Protocol (UIP, Jan 2026). It cites Bain and Morgan Stanley forecasts that agentic AI will materially influence U.S. e-commerce by 2030 and warns of risks such as coupon fraud if incentives aren’t centralized. Talon.One argues brands should centralize incentives, enable identity linking, and expose loyalty data in structured formats so AI agents can evaluate membership benefits alongside price. The article includes practical readiness steps for e-commerce, marketing and engineering teams and positions UIP as an extension to standards like UCP to surface loyalty and promotions to agents.

Coca‑Cola CFO Prioritizes Data Matching Ahead of Review

Coca‑Cola's president and CFO John Murphy said the company is focused on matching its proprietary first‑party data with partner and agency-held customer data to create new intelligence. The comments came ahead of a global agency review covering media, data and technology that will be managed by Mediasense, begin in July and conclude with a decision in the fall; the review excludes North America, Japan and Korea and will not touch global creative and PR (which remain with WPP). The piece frames the review as a direct test between Publicis’s strategy of building a proprietary data stack (Epsilon, Lotame, LiveRamp, CoreAI) and WPP’s privacy‑first connective approach (InfoSum, GroupM, WPP Open). Industry experts warned that converting combined first‑party data into decision‑ready outcomes is as much a people and integration problem as a technology one.

Benedict Evans on Where AI Is Headed

Independent analyst Benedict Evans (formerly a partner at Andreessen Horowitz) presented a wide-ranging, pragmatic take on AI’s near-term trajectory and economic impact. He frames the current era as an early, uncertain phase (“1997” for AI), argues AI will be as significant as the internet or mobile (but no larger), and outlines where value is likely to accrue across the AI stack. Evans discusses potential anti-AI backlash, a rise in consulting/professional services around AI deployments, the importance of distribution as a competitive moat, and reframes workforce risk questions from “what percent can AI do?” to “is this a task or a job?”. The piece is published via Lenny’s newsletter and links to Evans’ presentation and related resources. Publication date: 2026-05-31.