TL;DR

When an AI agent shops on a consumer’s behalf, it does not start with an open field of every brand in a category. It starts by deciding which brands are worth considering at all, and it makes that decision before it discovers a single product. The signal it reads first is loyalty state: identity, status, history, and ownership signals that tell the agent this consumer already has a relationship with this brand. That is the identity handshake, and it does the work paid acquisition used to do. Most brands never make the consideration set because their relationship signals are not machine-readable. We build those signals into the lifecycle, with the NCTR ecosystem as the proof we run at consumer scale.

The brands that get surfaced are not always the ones with the best product. They are the ones the agent can confirm the consumer already belongs to.

The agent’s moment of decision

Picture the agent at the start of a shopping task. A consumer has handed it a goal: find a running shoe under a budget, restock a pantry, compare two espresso machines. The agent does not begin by reading every brand in the category. It begins by narrowing, the same way a person does when they already trust a few names and ignore the rest. The agent assembles a consideration set, a short list of brands worth evaluating in depth, and only then does it discover and compare products within that set.

This is the step most brands never see, and the one that decides everything downstream. A brand that does not make the consideration set is not compared, not cited, and not bought. It is not that the agent judged the product and passed. The agent never evaluated the product at all, because the brand never made the list.

What the agent reads before discovery runs

So what does the agent read to build that list? Not the homepage. Not the ad. Before discovery runs, the agent reads the signals that tell it whether this consumer already has a relationship with a brand: identity, status, purchase history, and ownership signals. If the agent can confirm the consumer is a known customer of a brand, a member, a holder of status, someone with a standing relationship, that brand earns a place in the set ahead of brands the consumer has no tie to.

This is the identity handshake between brand, customer, and agent. In an authenticated exchange, where the agent knows who the customer is and what they are owed, that handshake is doing the consideration-set work that paid acquisition used to do. A brand used to buy its way into consideration with media spend. In the agent era, an existing, machine-readable relationship buys the same place, and it compounds instead of resetting every campaign.

The layer underneath the frame

The commerce lifecycle has five stages a brand can name in five seconds: Discovery, Search, Commerce, Loyalty, Payments. On that frame, loyalty looks like the fourth stage, the thing that happens after a sale. The deeper truth is that loyalty is not a stage in sequence at all. It is the layer underneath the frame, the one the agent reads first.

That reframing matters because it changes what loyalty is for. A points program designed to be redeemed after purchase is invisible at the moment that decides everything, because the agent reads the consideration set before any purchase is in view. A relationship signal that is structured, current, and machine-readable is the opposite: it is exactly what the agent reaches for first. The job is not to reward the customer after the fact. It is to be legible as a relationship before discovery runs.

Why this rewires acquisition

For a decade, the growth playbook was acquisition: spend to put a brand in front of enough buyers, convert a slice, repeat. That playbook assumed a human who could be reached with a message. When the agent is the one assembling the consideration set, the message has no reader. The agent is not moved by an impression; it is moved by a signal it can verify.

So the spend that used to buy attention now has to buy something more durable: a relationship the agent can read. A brand with a deep, machine-readable base of known customers walks into every agent-run shopping task already on the list. A brand without one starts each task from zero, hoping discovery surfaces it on product merit alone, against competitors the agent has already decided to trust. The relationship is the moat, and it is the one thing a competitor cannot buy back overnight.

What a brand makes readable

The work, then, is to make the relationship legible to a machine. That means the signals an agent uses to confirm a customer belongs, identity, status, history, ownership, are present, current, and structured for retrieval rather than locked inside a loyalty app a human logs into. It means treating those signals as discovery infrastructure, not post-purchase decoration.

NCTR is the version of this running in production. Wingman reads NCTR loyalty state to decide which brands surface in the Bounty Hunter app: the agent checks the relationship layer first, then surfaces accordingly. That is the identity handshake operating at consumer scale, today, not a slide about a future. If you want to see whether an agent can read your brand’s relationships at the moment it builds a consideration set, talk to the studio, or read the Human-to-Agent Shopping POV underneath this shift.

Key Takeaways

  • An AI agent builds a consideration set before it discovers a single product; a brand that does not make that set is never compared, cited, or bought.
  • The signal the agent reads first is loyalty state, identity, status, history, and ownership, not price or copy.
  • This identity handshake does the work paid acquisition used to do: an existing, machine-readable relationship earns a brand its place ahead of brands the consumer has no tie to.
  • Loyalty is not the fourth stage in sequence; it is the layer the agent reads first, which is why a redeem-after-purchase points program is invisible at the deciding moment.
  • The work is making relationship signals current, structured, and machine-readable, so the brand makes the consideration set before discovery runs.

FAQ

How does an AI agent decide which brands to consider?

An AI agent builds a consideration set before it discovers products, the same way a person narrows to a few trusted names. It reads relationship signals, identity, status, history, and ownership, to decide which brands the consumer already has a tie to, and gives those brands a place in the set ahead of brands with no relationship.

What is the identity handshake in agentic commerce?

The identity handshake is the exchange between brand, customer, and agent in which the agent confirms a consumer’s relationship with a brand, membership, status, ownership, before discovery runs. In an authenticated exchange, that handshake does the consideration-set work paid acquisition used to do.

Why does a traditional loyalty program not help here?

A loyalty program built to be redeemed after a purchase is invisible at the moment that matters, because the agent reads the consideration set before any purchase is in view. What helps is a relationship signal that is current, structured, and machine-readable, so the agent can read it first.

What signals does an agent read at the consideration step?

Identity, status, purchase history, and ownership signals, the data that tells the agent whether this consumer already belongs to a brand. The brands whose relationship signals are machine-readable get surfaced; the brands whose signals are locked inside a human-facing app get skipped.

How does NCTR demonstrate this today?

In the NCTR ecosystem, Wingman reads NCTR loyalty state to decide which brands surface in the Bounty Hunter consumer app. The agent checks the relationship layer first, then surfaces brands accordingly, which is the identity handshake operating at consumer scale in production.