The missing protocol in agentic commerce — and the opportunity no one has yet claimed.
Agentic commerce is about to reshape how decisions are made. AI agents will increasingly:
discover products
compare options
optimise baskets
execute purchases
And they'll do it without ever touching a brand's website, app, funnel, or checkout.
This shift exposes a structural gap in the ecosystem:
Loyalty is not yet designed for agents.
Agents can read price, availability, and policies. But loyalty — the most powerful lever in commerce — remains:
fragmented
brand‑specific
UI‑bound
emotionally driven
not machine‑readable
not comparable across merchants
There is no standard way for an agent to understand:
what points are worth
how rewards compare
how tiers and progress should influence decisions
how incentives shape long‑term value
This is why a Loyalty Commerce Protocol (LCP) becomes inevitable.
1. What LCP actually is (and why it matters)
LCP is not a loyalty program. It's not a CRM feature. It's not a platform.
LCP is a protocol — a shared language that allows AI agents, merchants, and ecosystems to understand loyalty value in a consistent, comparable, machine‑readable way.
If UCP is the "grammar" of agentic commerce, LCP is the dictionary of loyalty value.
Where UCP standardises:
product data
pricing
availability
policies
identity
transaction initiation
LCP would standardise:
point value
reward value
accrual rules
redemption rules
tier logic
progress semantics
streaks and milestones
personalised incentives
long‑term value signals
Today, every brand expresses these differently. Agents can't compare them. They can't optimise them. They can't reason about them.
LCP solves this.
2. Why LCP becomes essential in agentic commerce
Agents make decisions based on structured value, not emotion.
Without LCP:
loyalty value is invisible
incentives are incomparable
progress is unreadable
tiers are meaningless
long‑term value is opaque
agents default to price
With LCP:
loyalty becomes a first‑class decision variable
agents can optimise for LTV, not just price
merchants can compete on value, not discounts
incentives become programmable
progress becomes portable and visible
loyalty becomes interoperable across ecosystems
LCP is what allows loyalty to survive — and thrive — in an automated world.
3. The rails are forming — but loyalty is still under‑specified
Google's UCP: the agent–merchant language, not the loyalty brain
UCP defines how agents, merchants, and payment providers communicate. It handles:
structured product and offer data
pricing, availability, policies
identity and trust
secure transaction initiation
UCP includes basic incentive hooks (promos, discounts, points fields). But these are plumbing — not semantics.
UCP does not define:
how loyalty value should be interpreted
how incentives compare across merchants
how progress, tiers, and streaks should be represented
how agents should optimise for long‑term loyalty outcomes
UCP is the commerce + identity layer. LCP would be the loyalty meaning layer.
Visa & Mastercard: agent payments, not loyalty
Their focus is:
agent identity
authorisation
fraud and risk
secure execution
They are building the agent payment rails, not the loyalty layer.
Stripe: programmable checkout, not value exchange
Stripe is preparing for a world where agents execute purchases via APIs, not UI.
Their focus:
programmable checkout
orchestration
developer‑first payments
Again: essential rails, but not loyalty semantics.
4. The emerging agentic commerce stack — and the missing layer
Here's the stack forming around agents:
1
2
3
4
5
1
Identity & trust
2
Commerce structure
3
Incentive plumbing
4
Payment execution
5
Loyalty semantics (missing)
UCP gives agents the fields. LCP would give them the meaning.
But agents don't "check out." They execute transactions programmatically.
The battle shifts from:
"How do we get humans through this funnel?"
to:
"How do we expose the right data and incentives to agents at protocol level?"
Loyalty platforms (Comarch, Antavo, etc.)
Deep loyalty expertise — but:
proprietary
brand‑bound
not neutral
not protocol‑native
They can contribute to LCP, but they are unlikely to own it.
6. Who will try to own LCP?
Most likely:
A protocol‑native startup
Google (if they extend UCP)
Possible:
Visa / Mastercard
Stripe
Less likely:
Bloomreach
Comarch / Antavo
7. Why LCP is a generational opportunity
Loyalty is shifting from brand‑owned UX to ecosystem‑level protocol.
UCP has laid the groundwork by making commerce, identity, and incentives agent‑ready. But the deeper questions — how loyalty is understood, compared, and optimised — are still unanswered.
LCP is the missing layer.
It will define:
how agents reason about loyalty
how merchants compete on value
how ecosystems structure incentives
how LTV is optimised in an automated world
The protocol layer for loyalty is forming. The question is: who will build it?