LIVE Event | Geopost Innovation Talks 2026

Geopost Innovation Talks 2026: Agentic Commerce

The discussion examines how AI is moving from answering questions and recommending products to acting on a shopper’s behalf—searching, comparing, selecting, and potentially purchasing. This shift could compress the customer journey into a single decision moment and make product and delivery information more important in determining which businesses are selected.

Adoption Signals and Examples

• A Geopost survey of more than 31,000 e-shoppers across 22 European countries found that 40% already use AI in their shopping journey.
• Among surveyed SMEs, 15% already use AI to select providers.
• Geopost speakers described the next step as a move from asking AI for recommendations to giving it a mandate to complete a purchase.
• Chronopost’s ChatGPT integration lets consumers track parcels in real time and find pickup points. Shippers can calculate shipping costs and start a pre-filled parcel-label process. Chronopost framed this as an early step toward more personalized and proactive delivery services.
• OpenAI speakers said their near-term focus is helping users understand intent, research options, and compare products—not taking over every stage of checkout and payment. They noted that many merchants have already invested in their own purchasing and checkout experiences.

Business Implications

• Visibility is becoming selectability. Being visible to consumers in traditional search does not guarantee that an AI agent can understand, compare, or recommend a business.
• Delivery moves upstream in the purchase decision. Agents may consider delivery cost, speed, convenience, pickup options, reliability, and sustainability while comparing products—not only after a product is chosen.
• Competition becomes broader and more holistic. AI can assess many products and criteria quickly, increasing the need for brands to distinguish their complete value proposition.
• Retailer economics may be affected. If shoppers increasingly arrive through AI rather than retailer websites, retailers could lose opportunities for upselling, cross-selling, loyalty-building, and retail-media revenue. This raises questions about the profitability of e-commerce models and the roles of retailers, brands, and AI platforms.
• The human experience still matters. As routine shopping becomes more automated, delivery may become a more prominent human touchpoint through home delivery, lockers, and parcel shops.
• Automation is likely to vary by purchase type. Routine, low-involvement purchases may be delegated to agents, while discovery-led or emotionally engaging purchases may remain more directly controlled by consumers.

Frameworks and Practical Playbooks

• Generative Engine Optimization (GEO): The retail expert outlined three pillars:
  • Technical: Make websites accessible to AI systems through clear site and page structures.
  • Content: Organize information so agents can find and interpret specific attributes. Use clear answers to likely customer questions and language that reflects customer prompts.
  • Reputation: Maintain credible external signals, including press coverage, reference resources, social discussions, and customer reviews.
• Build machine-readable product and service data. Structure and keep current information on product features, availability, prices, delivery options, return policies, sustainability credentials, and customer reviews.
• Prioritize data quality and consistency. Align information across channels and ensure it matches the service customers actually receive. Inaccurate or conflicting data can undermine recommendations.
• Make evidence of operational performance accessible. Geopost’s innovation director highlighted three logistics priorities: execute reliably, provide evidence of reliability, and make that evidence accessible to third-party systems.
• Design for intent, not just keywords. Shoppers may describe needs in natural language—for example, wanting work shoes that are comfortable and stylish—so product information should address customer outcomes and attributes directly.
• Keep human judgment and oversight in the decision loop. The academic speaker advised against trusting AI outputs wholesale. Businesses should validate outputs, audit systems, establish governance, and determine when people should review decisions.

Risks and Unresolved Questions

• Trust and explainability: AI can make errors, and businesses need ways to validate recommendations and communicate uncertainty. The academic speaker emphasized that AI’s decision processes are less well understood than many forms of human bias.
• Misleading information: Weak product information can exclude good products from consideration; false reviews or fabricated coverage could distort recommendations.
• Safety and liability: The panel raised an allergy example in which incorrect product information could lead to a harmful purchase, asking whether responsibility would fall to the merchant, customer, or technology provider.
• Payment and fraud: Speakers noted unresolved questions about responsibility if payment details are misused in agent-led transactions.
• Post-purchase fragmentation: Buying several products through one AI interface could still create a poor experience if customers must deal with multiple tracking systems and customer-service teams. More consistent post-purchase experiences may be needed.

Key Metrics and Timelines

• 40% of surveyed European e-shoppers use AI in their shopping journey.
• 15% of surveyed SMEs use AI to select providers.
• Survey sample: more than 31,000 e-shoppers across 22 countries.
• In a comparison of technology adoption, a speaker noted that steam power and electricity took decades for productivity effects to become clear, while internet and IT took roughly 20–30 years. The speaker said AI appears to be unfolding faster, but gave no specific adoption timeline or business targets.
• No revenue, margin, CAC, LTV, churn, or growth targets were provided.

Presenters and Sources

• Julia Seager — Host and technology journalist
• Martin — Geopost Director of Innovation and New Services; surname unclear in subtitles
• Michael Impink — Professor specializing in AI and business strategy
• Eva Bossi — Chronopost Marketing and Digital Director
• William Doge — OpenAI Go-to-Market; name as heard in subtitles
• Vincent Katier — Retail and digital expert; name as heard in subtitles
• Jean-Claude — Geopost Executive Vice President for Marketing, Communications, and Sustainability; surname unclear in subtitles