White paper
Agentic AI and the new dynamics of retail
For 30 years, digital commerce was engineered for human eyes and manual clicks. By 2030, autonomous AI agents are projected to drive up to 25% of global e-commerce. Here is how enterprise leaders can prepare their digital core for delegated machine buying.
E-commerce is undergoing its sharpest structural shift since the start of the commercial web. For decades, organizations optimized for human emotion – visual merchandising, promotional banners, and guided cart funnels.
That model is breaking down. Shoppers are delegating product discovery, price comparisons, and purchasing decisions directly to autonomous software agents.
These AI buyers do not care about sleek storefronts. They evaluate structured utility: verified inventory, real-time pricing APIs, fulfillment velocity, and strict return policies. If your product catalog sits behind unstructured web pages and client-side scripts, autonomous agents will bypass your brand for competitors they can read.
AI does not fix poor infrastructure, it amplifies it. To win in an agent-led economy, enterprises must modernize their digital cores into machine-readable, protocol-agnostic platforms built to serve both human shoppers and AI proxies.
What you'll learn in this briefing
The shift from clicks to delegated buying.
Understand how machine proxies make purchasing decisions and why brands risk being demoted to invisible back-end suppliers without machine-first integration.
The dual discovery engine.
Balance external reach across third-party platforms like ChatGPT, Google Gemini, and Microsoft Copilot with high-trust, brand-owned conversational interfaces.
Protocol-agnostic architecture.
Connect your systems to emerging standards — including Model Context Protocol (MCP), Agentic Commerce Protocol (ACP), and Universal Commerce Protocol (UCP) — without building fragile point-to-point integrations.
Delegated authority and payments.
Implement cryptographic spending certificates, network tokenization, and tiered autonomy models that authenticate purchases without manual checkout forms.
Governance and dispute resolution.
Establish immutable evidentiary trails to resolve out-of-bounds agent actions, data poisoning risks, and multi-party liability disputes.
The four-phase enterprise roadmap.
Move from experimental pilots to full production readiness across data unification, multi-protocol scaling, and automated post-purchase operations.
Modern architecture in action
See how leading brands are already engineering the foundations for agentic scale:
| JD Sports: | Integrated inventory across multiple retail banners using composable architecture and modular APIs, preparing for real-time checkout inside external AI conversational platforms. |
| Screwfix: | Unified cart management and stock logic across digital and store touchpoints to serve lightning-fast automated requests alongside trade customers. |
| Pet Valu: | Eliminated regional inventory silos and automated promotional logic through an API-first engine, giving machines immediate visibility into accurate discounts. |
Meet the author
Rav Hayer
Managing Director, UK and Ireland; Head of BFSI Europe, Thoughtworks
Rav advises global enterprises across retail, marketplaces, and financial services on engineering modern, resilient architectures that turn emerging technologies into reliable commercial advantage.
