Artificial intelligence is transforming commerce, shifting the focus from assistance to direct execution through digital agents. Imagine telling an agent, "Organize a trip for the family to Italy, respecting the budget, choosing hotels we liked, and managing all the details." The agent won't just provide links, but will create a complete itinerary and finalize the purchase.

The Centrality of Data

This transition implies that "good enough" data is no longer sufficient. In an agent-driven economy, the limit is not speed, but the reliability on which machines make decisions. Master data management (MDM) becomes fundamental, tracking who an agent represents, what it can do, and who is responsible in case of problems. MDM transforms autonomous action into legitimate and scalable trust.

The Challenges of Agentic Commerce

The introduction of agents creates new challenges:

  • Identity: Who is the individual, across channels and devices, with enough certainty for automation?
  • Permissions: Who is the agent, and what permissions and limits define their actions?
  • Suppliers: Who is the supplier, and are we sure it's the right one?
  • Responsibility: Who is responsible if the agent acts with permission, but against the user's intent?

Without deterministic signals, agents could make mistakes, compromising trust or requiring continuous human verification.

Preparing for the Future

To prepare for this transformation, companies should:

  1. Treat agents as governed identities, defining onboarding, authentication, authorization, monitoring, and retirement processes.
  2. Prioritize entity resolution where the cost of being wrong is highest (payees, suppliers, employee-versus-personal identity, product categories).
  3. Create a reusable context service that every workflow and agent can call.
  4. Precompute and compress signals, curating the context upstream to speed up decisions.
  5. Expand autonomy only when trust is earned, with a governance framework for disputes and high-risk actions.