Keynote · 10 slides

When the buyer is an algorithm, the rules of the game change

A blend of the agreed structure and the evidence in the source library. Each slide carries speaker notes. Review and mark up freely.

For decades, brands won by following Byron Sharp's laws: physical availability (be easy to find and buy) and mental availability (be salient, distinctive, top-of-mind). Both assume a human makes the call. This repository asks the question your shelf never had to: what survives when an agent chooses? Does mental availability still matter to a machine that is “immune to nostalgia” — or does it collapse into structured-data availability, feed quality and machine-readable trust?

Slide 1 · Open

When the buyer is an algorithm, the rules of the game change

How will AI agents change buying behaviour, how will retailers and brands win, and what should the CxO do?

Say: Open live — ask ChatGPT to buy something and let Instant Checkout complete it on screen. “A purchase just happened with no shelf, no ad, no human in the aisle. That is the shift we are here to understand.”

Slide 2 · Why now

This is already happening

38→51%consumers using GenAI to shop, 2024→2025

~80%intend to use GenAI to shop in 2026

+304%growth in AI-driven retail referral traffic in 2025

~70%shoppers comfortable letting an agent purchase on their behalf

Sources: Capital One Shopping; Stord; Euromonitor / Adobe Analytics; Ekamoira, 2026

Say: Adoption is not linear, it is exponential, and it has crossed from research into transaction. The question is no longer if, but how fast and in which categories first. Shoppers are not waiting for permission either — roughly seven in ten are already comfortable letting an agent buy on their behalf. Buying journeys move through five stages: human-only, human-plus-AI research, AI-recommends-and-human-approves, agent-driven transactions within limits, and fully autonomous agent buying and replenishment. We are mid-curve: replenishment and commodities are already agent-driven, identity-laden categories lag, and the curve moves left-to-right category by category, not all at once.

Slide 3 · The foundation

Byron Sharp's two jobs

Physical availability — easy to find and buy, everywhere the category is bought.
Mental availability — easy to bring to mind: distinctive assets, emotional reach, salience.
Both laws assume one thing: a human makes the call.

Say: Sharp's laws are the best empirical model marketing has. But they are laws of human memory and attention. Hold that assumption up to the light.

Slide 4 · Old rules don't apply

What works on humans can backfire on agents

The classic e-commerce persuasion toolkit — scarcity messages, countdown timers, vouchers, bundling, strike-through pricing — tested directly against agent decision-making.
Across thousands of simulated shopping rounds, four AI models and four product categories, the tactics built for human psychology did not reliably work, and several backfired outright.
Only authentic star ratings consistently increased an agent's likelihood of choosing a product; a higher price reliably decreased it.

“In thousands of simulated shopping rounds… only star ratings consistently increased choice in the expected direction, while price reliably decreased it.” — Harvard Business Review, “Research: Traditional Marketing Doesn't Work on AI Shopping Agents,” May 2026

Say: This is not a theory, it has been tested. Researchers ran the human persuasion playbook directly against AI shopping agents, and most of it failed or backfired depending on the model. The two things that reliably moved an agent were star ratings and price — exactly what a rational comparison engine should respond to, and exactly what a human swayed by urgency or bundling would not.

Slide 5 · The hinge

How brands grow — the agent is immune to nostalgia

Agents weigh attribute-level data, not jingles or memory structures.
Mental availability does not vanish — it changes owner.
The new salience = clean structured data + verifiable, machine-readable trust.

“AI agents are immune to nostalgia” (Bazaarvoice) · “AI systems don't browse like humans... they require structured, machine-readable content” (Akamai)

3–4xhigher AI visibility for stores with near-complete product data (“Golden Record”)

Say: This is the keynote's hinge. If the agent carries the salience, your distinctive assets must become machine-readable attributes. The jingle never reaches the machine — and this is now measurable: 3-4x visibility for near-complete data. Source: eFulfillment Service, 2026.

Slide 6 · Product businesses

Implications for FMCG & durable goods

FMCG (P&G, Unilever)	Durable goods (auto, electronics, appliances)
Win the recommendation algorithm, not the shelf glance	Win the comparison & configuration agent
Subscription & replenishment becomes the default basket	Ecosystems and lock-in shape agent choice
Brand power → data power (feeds, reviews, claims)	Service revenue becomes the durable margin

Say: For FMCG the battle moves into the replenishment loop and the recommendation engine. For durables, the agent does the spec comparison you used to win with a showroom and a brochure.

Slide 7 · Retailers

The new shelf is an answer surface

Three agent types; third-party “objective” agents threaten disintermediation most.
Whoever owns the answer owns the customer relationship.
Retail media shifts from banner to being in the answer; you now serve two customers — the human and the agent.

Two live bets on openness: Walmart integrates with ChatGPT, Gemini and the Universal Commerce Protocol while building its own “Sparky” agent — reach over control. Amazon keeps agent interactions inside its own walled garden, led by “Buy for Me” — control over reach. Neither bet has won yet.

Source: Bain & Company, 2030 agentic retail forecast; ALM Corp, 2026

Say: The shelf war becomes an answer war. Decide where you meet the third-party agent and where you fight to own the relationship yourself. Walmart vs. Amazon is the test case running live right now.

Slide 8 · Service businesses

Implications for services: agent-to-agent

Travel	Insurance	Banking & wealth
AI trip planners; dynamic packaging; agent-to-agent booking	AI brokers; product comparison; claims automation	Personal finance agents; automated switching; advice agents

When the customer's agent negotiates with your agent, price, terms and clarity beat brand familiarity.

Say: Services feel safe because they are complex. They are not. The complexity is exactly what agents are built to handle — comparison, switching, packaging. Gartner sees $15T of B2B purchasing intermediated by agents by 2028.

Slide 9 · The C-suite

What every CxO should do

CEO	Own the strategic question and exposure; decide where to own vs. meet the agent.
CIO	APIs, MCP, agent identity & security, agent-readiness, architecture.
CFO	New economics, shifting revenue pools, agent-era metrics, investment priorities.
CMO	Generative Engine Optimisation, brand discoverability, zero-click marketing, content architecture.
CCO	Commercial model, pricing, channel conflict, the agent-era sales model.

Say: This is not a marketing project. Every C-suite seat has a different lever, and they have to pull together.

Slide 10 · Act & close

30 / 60 / 90 — win the human's trust, win the agent's logic

First 30 days — Understand	31–60 — Experiment	61–90 — Operationalise
Educate leadership; map buying journeys; assess exposure → Assessment	Test AI discovery; build APIs; create GEO content → Pilots	Define roadmap; governance; budget; capabilities → 12-month roadmap

Say: Close on the title flip: the future belongs to brands that know how to sell to an agent, not just a human. Point the room to howtoselltoanagent.com as the living playbook.