Agentic commerce is the point where AI agents stop assisting a purchase and start completing it — discovering options, verifying terms, negotiating, transacting, and handing over the outcome — on behalf of both the buyer and the seller. Soon a meaningful B2B deal will close in your industry through exactly that loop: a buyer’s agent carrying a mandate and budget meets a seller’s agent carrying a catalogue and pricing authority, and the entire journey — first signal to signed order — completes with no human walking a single step.

The customer journey has been the central artifact of our discipline for two decades. We mapped it, measured it, removed its friction, designed its emotional peaks. Now comes the strangest turn in its history: the journey still exists, every stage of it — but for the first time, nobody experiences it. This article maps that journey: the evidence it is already built, what it becomes when the halves connect, and how to prepare.

1. Both Halves Are Already Built

Start with the seller’s side, because it is further along than most executives realize. McKinsey’s July 2026 research on rewiring customer experience for the agentic era starts from familiar ground — CX leaders outperform peers on growth and total shareholder returns by as much as two times — then delivers the agentic finding: 41% of AI deployments in customer-facing functions have fully scaled, making them 3.5 times more likely to reach scale than deployments in any other business domain. The cases behind that statistic are audited and specific. A UK energy retailer with six million customers deployed a natural-voice agent that within five months understood roughly 75% of needs at 85% intent accuracy — lifting satisfaction 6% with annual run-rate impact above $10 million. A US premium automotive OEM handling 50,000+ cases a year deployed an agent synthesizing structured and unstructured data, driving a 24% increase in first-contact resolution and 30% higher productivity among its human agents. A €40 billion European telco connected decisioning across channels for a €40 million margin impact. The selling machine is in production — and producing.

Now the buyer’s side — and the data should recalibrate anyone who thinks this is a 2030 conversation. McKinsey’s own survey work found that 44% of European consumers who have tried AI search now use it as their primary search method, and about half of consumers use AI when searching for products online. The B2B picture is at least as advanced: a 2026 multi-source analysis spanning 680 million AI citations found that 73% of B2B buyers now use AI tools in their purchase research. TrustRadius’s 2026 B2B Buying Disconnect Report, surveying 1,862 technology buyers, found 63% used AI during their purchase journey — one of its fastest-growing research resources. Buyers still click the final button themselves; delegating the click is what the payment industry has spent two years building.

The third piece: the rails. Visa’s own newsroom declares it without hedging: in 2026, AI agents “won’t just assist your shopping — they will complete your purchases.” Visa’s single merchant integration now supports agent-initiated payments across four protocols — 100+ partners globally, hundreds of real agent transactions already completed in controlled settings. Google, OpenAI, Stripe, and Mastercard have each shipped their own agentic transaction standards, and McKinsey projects AI agents could be responsible for $1 trillion in US transactions alone by 2030. Seller agents in production. Buyer-side AI at majority adoption. Payment rails live. And the milestone has already been crossed: in March 2026, Mastercard announced its first live, authenticated agentic transaction in Hong Kong, conducted with HSBC and DBS Hong Kong: an AI agent booked and paid for a ride to Hong Kong International Airport, secured by tokenized credentials authenticated with Mastercard Payment Passkeys — with the same capability confirmed across cards from Citi, Hang Seng, Standard Chartered, and Mox. The two halves of the autonomous journey are built; what remains is the moment they routinely connect.

One more number makes the stakes unignorable: the same buyer analysis found traffic from AI search converts at 14.2%, against 2.8% for traditional organic — a 5.1x advantage — yet only 22% of marketers track their AI visibility. The highest-converting channel in digital commerce is forming right now, and four out of five companies cannot see whether they appear in it.

2. The Meeting Point: Every Stage Survives, Every Stage Transforms

McKinsey gives the right lens for that connection: agentic AI shifts CX from designing journeys to governing “decisions in motion” — live choices made continuously by goal-driven agents — with agentic commerce as the early signal of its second horizon: an agent that knows preferences, narrows options, verifies terms, completes the purchase, and carries context forward so nothing resets at the handoffs.

Extend that logic to both sides of the table and the classic journey redraws itself, stage by stage:

Discovery becomes retrieval. No awareness campaigns reach a buying agent. It queries, retrieves, and synthesizes; you are discovered through the structured information AI systems can access, not the impressions you buy.

Evaluation becomes verification. A buyer agent does not read testimonials; it cross-checks claims against specifications, delivery records, and pricing consistency. Reputation, for an algorithm, is the audit trail.

Negotiation becomes protocol. Terms, discounts, and configurations are exchanged agent-to-agent within mandates humans defined in advance — negotiation in milliseconds, inside boundaries set upstream.

Transaction becomes tokenized authorization. The four live payment protocols exist precisely for this step: verified agent identity, scoped spending authority, revocable consent, auditable intent.

Onboarding becomes context transfer. The purchase completes and the buyer’s agent already holds the configuration, the entitlements, the setup steps — “nothing resets at the handoffs,” realized between companies rather than within one.

Loyalty becomes standing preference. Retention is no longer an emotional bond renewed at touchpoints; it is a default written into the buying agent’s logic — and contested at every repurchase by every competitor’s cleaner data.

3. What Changes When Nobody Walks the Journey

Here is the conceptual shift leaders must absorb: the journey doesn’t disappear — it loses its witnesses in the middle and concentrates all human experience at its two ends.

At the front end sits the mandate moment: the conversation where a human briefs their agent — goals, budget, constraints, values, tolerance for trade-offs. At the back end sits the outcome moment: where a human receives what was bought and judges whether it delivered. Everything between them is executed, not experienced. That makes those two moments the most valuable real estate in the new journey, and almost no company designs for either of them yet. Mandate design matters commercially: the criteria a customer writes into their agent’s brief — “prioritize delivery reliability,” “prefer certified suppliers” — become the selection algorithm your company must win inside. Influencing how mandates get written may become the highest-leverage marketing activity of the next decade: persuasion moving upstream, from purchase to brief.

The middle, meanwhile, changes its currency from persuasion to verification. And the buyer data carries a warning: TrustRadius found 94% of buyers who used AI fact-check its responses at least some of the time. Humans do not yet extend blind trust to their own agents — so mandates will be conservative, oversight real, and the winners will be sellers whose claims survive review by a skeptical principal. In the journey nobody walks, credibility is not what you say about yourself; it is what an algorithm can confirm and defend to its human afterwards.

4. How to Prepare: Five Moves for the Journey’s New Shape

Preparation here is journey work, not just technology work — and it can begin this quarter.

1. Re-map your journey for two travellers. Draw the agent’s version beside your existing map: what does a buying agent need at each stage? Every blank stage is revenue already conceded.
2. Become retrievable, not just visible. Since discovery now happens inside AI engines, structure your product, pricing, and capability information so machines can retrieve and cite it accurately. Marketing built for human eyes is invisible at the new front door.
3. Pilot one agent-accessible transaction path. The protocols are live; make one repeatable, lower-risk offering purchasable end to end by an authorized agent, and learn from real agent behavior before competitors define the norms.
4. Design the two human moments deliberately. Build the mandate experience — help customers brief their agents well — and elevate the outcome experience, now your only emotional touchpoint. These are the journey’s new peaks, and they are wide open.
5. Instrument the agent funnel separately. Agent traffic, verification failures, quote-to-completion rates, silent deselections: none appear in today’s CX dashboards. An agent never complains on the way out; readers of From Systems to Signals will recognize why real-time visibility becomes existential here.

5. The Last Word

Line up the verified facts of 2026 and the direction stops being a forecast; it becomes arithmetic. On the selling side, 41% of customer-facing AI deployments have fully scaled — 3.5 times the rate of any other business domain — with audited results behind the statistic: a 6% satisfaction lift and eight-figure run-rate impact at a six-million-customer energy retailer, a 24% jump in first-contact resolution at a premium automotive manufacturer, a €40 million margin impact at a European telco. On the buying side, 73% of B2B buyers already research through AI, 44% of European consumers who tried AI search have made it their primary method — and every one of them is a single mandate away from delegating the transaction itself. Between them, the rails are no longer promises: four payment protocols in production, 100+ partners, a first fully agentic payment completed in Hong Kong this March, and McKinsey projecting $1 trillion in US agent-driven transactions by 2030. Meanwhile the channel where this journey begins already converts at 5.1 times the rate of traditional search — and four out of five companies are not even measuring their presence in it.

Those are the facts, and they permit one reading: the journey nobody walks is not arriving — it is assembling, piece by verified piece, around companies that are still designing exclusively for human footsteps. For twenty years we asked one question of every journey: how does it feel? Now the question splits in two: does it verify, in the middle — and does it matter, at the ends? The companies that answer both will discover something unexpected: removing humans from the middle of the journey doesn’t make commerce less human. It concentrates the humanity where it always counted most — in the intent that starts the journey and the outcome that ends it. Design brilliantly for those two moments, build verifiable truth for everything in between, and the age of agent-to-agent commerce will not disintermediate you. It will distill you.

👉 Stay ahead of CX, AI, and innovation — Subscribe to my weekly LinkedIn Newsletter “CX Insights by Ricardo S. Gulko.

Share it if it’s useful — and let’s connect for more. Ricardo Saltz Gulko

My columns in several respected CX publications.

• My recent articles on Eglobalis: https://www.eglobalis.com/blog/
• My recent articles on the European CX Organisations ECXO.org : https://ecxo.org/blog/
• My recent articles on CMSWire: https://www.cmswire.com/author/ricardo-saltz-gulko/
• My German articles on CMM360: https://www.cmm360.ch/author/ricardo/
• My old articles on CustomerThink: https://customerthink.com/author/rgulko/

Primary Sources

Every statistic in this article was verified directly against the live pages below on July 21, 2026.

• McKinsey & Company — Rewiring Customer Experience for the Agentic Era (decisions in motion; three horizons; 41% of customer-facing AI deployments fully scaled, 3.5x other domains; agentic commerce description; UK energy retailer and European telco cases)
• Eglobalis — From Systems to Signals: How Real-Time Data Are Rewriting B2B CX

• Loganix / PRNewswire — 73% of B2B Buyers Use AI Tools in Purchase Research (multi-source analysis, 680 million AI citations, March 2026)
• TrustRadius / PRNewswire — 2026 B2B Buying Disconnect Report (1,862 technology buyers; 63% used AI in purchase journey; 94% fact-check AI responses)
• Visa — Visa and Partners Complete Secure AI Transactions (official newsroom: agents will complete purchases in 2026; Trusted Agent Protocol)
• TechInformed — Visa Opens One Integration for AI Agent Payments (four agent protocols supported; 100+ partners; hundreds of real agent-initiated transactions)
• Digital Commerce 360 — How Visa and Mastercard Are Approaching Agentic Commerce (McKinsey $1 trillion US agent-transaction projection by 2030; Stripe, Google UCP, OpenAI ACP ecosystem)
• Mastercard — Mastercard Completes Its First Live Agentic Transaction in Hong Kong (official newsroom, March 2026: authenticated agentic transaction with HSBC and DBS