AI-mediated commerce is not advancing uniformly. Marketplaces, news and travel currently lead AI referral volume, while beauty and fashion show the fastest growth; finance and consumer electronics are also accelerating. But referral traffic is only one dimension. Sectors differ in research intensity, data structure, transaction complexity, regulation, trust and the ease with which an agent can act. The highest near-term opportunity is where conversational discovery is valuable and transactions are standardised enough to delegate; the slowest categories are high-consequence, regulated or physically ambiguous.
Research method
This report triangulates current primary/public-sector evidence, large-sample research and relevant industry/platform documentation. It prioritises observed behaviour over forecasts, distinguishes vendor claims from neutral statistics, and uses an explicit 8i judgment layer only where source evidence ends. The analysis is current through 15 September 2026.
1. Observed referral traffic is now large enough to segment
Similarweb estimates AI platforms drove an average 770.7 million referral visits per month worldwide between June 2025 and May 2026, up 117.4% year over year. Marketplaces led volume at 46.8 million average monthly visits, with News and Travel close behind. S1
Growth is differently distributed: Beauty grew 312.5% year over year, Fashion 278.2%, Marketplaces 237.3%, Finance 236.5% and Consumer Electronics 215.4% in Similarweb’s category data. S1 This inversion between volume and growth is important for investment timing.
2. Commerce behaviour is shifting beyond referrals
Salesforce’s 2026 State of Commerce work, based on surveys of 4,690 consumers and 3,450 commerce professionals plus behavioural data from more than 1.5 billion shoppers, reports agentic search as the first step in the purchase journey growing 200% year over year. S2
AI influence is broader than clicks because a customer can use an assistant to narrow options and then arrive directly, via search, app or store. Any sector heatmap should therefore combine referral data with decision characteristics.
3. 8i sector heatmap
| Sector | Discovery maturity | Transaction readiness | 2026 8i status | Why |
|---|---|---|---|---|
| Marketplaces | High | High | LEAD | Large inventories, structured data, broad choice, standard checkout. |
| Travel | High | Moderate–high | LEAD | Research-heavy, comparison-rich, bookable inventory; exceptions remain complex. |
| Beauty / fashion | Fast-rising | High | ACCELERATE | Conversational recommendation fits discovery; standard ecommerce transaction. |
| Consumer electronics | High-rising | High | ACCELERATE | Spec comparison, compatibility and pricing are agent-friendly. |
| Financial services | Moderate-rising | Low–moderate | WATCH / BUILD | High research demand but regulated advice, identity and consequence slow delegation. |
| B2B software/services | Moderate | Low–moderate | BUILD | AI research and shortlisting growing; complex consensus and negotiation retain humans. |
| Local services | Moderate | Moderate | EMERGING | Scheduling/price discovery suitable; quality variability and local supply complicate. |
| Healthcare | Moderate | Low | CAUTION | High information demand but clinical risk, privacy and regulation limit autonomous action. |
| Automotive / high-ticket durable goods | Moderate | Low–moderate | EMERGING | Rich research and comparison; financing, trade-in and physical inspection slow full delegation. |
4. A sector-readiness model
- Research intensity: how much comparison/synthesis occurs before purchase?
- Data structure: are product/service attributes machine-readable and current?
- Transaction standardisation: can checkout/booking occur with bounded parameters?
- Consequence of error: can a poor decision be cheaply reversed?
- Regulation and liability: does advice/action trigger professional or legal duties?
- Identity/consent burden: what sensitive data or authority is required?
- Inventory/state volatility: does the agent need real-time price/availability?
- Post-purchase complexity: returns, support, fulfilment, servicing and disputes.
Sources and notes
- Similarweb, AI Referral Traffic by Industry: 2026 Data (3 Sep 2026). [Source](https://aisearch.similarweb.com/blog/ai-referral-traffic-by-industry/)) — Worldwide web referrals June 2025–May 2026; seven industries.
- Salesforce, Agentic Search Grows 200% (28 Jul 2026). [Source](https://www.salesforce.com/news/stories/agentic-search-growth/?bc=OTH)) — Consumer/commerce surveys plus >1.5bn shopper behavioural dataset.
8i TECHNOLOGIES
AI-Mediated Commerce Sector Heatmap 2026
Where AI Discovery, Recommendation and Agentic Transaction Are Becoming Material First
Reference date: 15 September 2026
See what matters next.
| Executive answer AI-mediated commerce is not advancing uniformly. Marketplaces, news and travel currently lead AI referral volume, while beauty and fashion show the fastest growth; finance and consumer electronics are also accelerating. But referral traffic is only one dimension. Sectors differ in research intensity, data structure, transaction complexity, regulation, trust and the ease with which an agent can act. The highest near-term opportunity is where conversational discovery is valuable and transactions are standardised enough to delegate; the slowest categories are high-consequence, regulated or physically ambiguous. |
|---|
Research method
This report triangulates current primary/public-sector evidence, large-sample research and relevant industry/platform documentation. It prioritises observed behaviour over forecasts, distinguishes vendor claims from neutral statistics, and uses an explicit 8i judgment layer only where source evidence ends. The analysis is current through 15 September 2026.
1. Observed referral traffic is now large enough to segment
Similarweb estimates AI platforms drove an average 770.7 million referral visits per month worldwide between June 2025 and May 2026, up 117.4% year over year. Marketplaces led volume at 46.8 million average monthly visits, with News and Travel close behind. [S1]
Growth is differently distributed: Beauty grew 312.5% year over year, Fashion 278.2%, Marketplaces 237.3%, Finance 236.5% and Consumer Electronics 215.4% in Similarweb’s category data. [S1] This inversion between volume and growth is important for investment timing.
2. Commerce behaviour is shifting beyond referrals
Salesforce’s 2026 State of Commerce work, based on surveys of 4,690 consumers and 3,450 commerce professionals plus behavioural data from more than 1.5 billion shoppers, reports agentic search as the first step in the purchase journey growing 200% year over year. [S2]
AI influence is broader than clicks because a customer can use an assistant to narrow options and then arrive directly, via search, app or store. Any sector heatmap should therefore combine referral data with decision characteristics.
3. 8i sector heatmap
| Sector | Discovery maturity | Transaction readiness | 2026 8i status | Why |
|---|---|---|---|---|
| Marketplaces | High | High | LEAD | Large inventories, structured data, broad choice, standard checkout. |
| Travel | High | Moderate–high | LEAD | Research-heavy, comparison-rich, bookable inventory; exceptions remain complex. |
| Beauty / fashion | Fast-rising | High | ACCELERATE | Conversational recommendation fits discovery; standard ecommerce transaction. |
| Consumer electronics | High-rising | High | ACCELERATE | Spec comparison, compatibility and pricing are agent-friendly. |
| Financial services | Moderate-rising | Low–moderate | WATCH / BUILD | High research demand but regulated advice, identity and consequence slow delegation. |
| B2B software/services | Moderate | Low–moderate | BUILD | AI research and shortlisting growing; complex consensus and negotiation retain humans. |
| Local services | Moderate | Moderate | EMERGING | Scheduling/price discovery suitable; quality variability and local supply complicate. |
| Healthcare | Moderate | Low | CAUTION | High information demand but clinical risk, privacy and regulation limit autonomous action. |
| Automotive / high-ticket durable goods | Moderate | Low–moderate | EMERGING | Rich research and comparison; financing, trade-in and physical inspection slow full delegation. |
4. A sector-readiness model
- Research intensity: how much comparison/synthesis occurs before purchase?
- Data structure: are product/service attributes machine-readable and current?
- Transaction standardisation: can checkout/booking occur with bounded parameters?
- Consequence of error: can a poor decision be cheaply reversed?
- Regulation and liability: does advice/action trigger professional or legal duties?
- Identity/consent burden: what sensitive data or authority is required?
- Inventory/state volatility: does the agent need real-time price/availability?
- Post-purchase complexity: returns, support, fulfilment, servicing and disputes.
5. Why marketplaces gain an early structural advantage
Similarweb observes AI assistants disproportionately sending users to large multi-category marketplaces. [S1] The reason is structural: a marketplace gives an assistant broad inventory, price and availability coverage with one destination. Smaller brands therefore need a dual strategy—direct machine discoverability and high-quality presence in aggregators/platforms that assistants already use.
6. Regulated sectors will be assistive before autonomous
Finance and healthcare have strong informational use cases but high consequence. AI can research, explain, compare and prepare; full recommendation or transaction authority introduces suitability, privacy, professional responsibility and identity controls. Visa’s consumer research also shows trust falling sharply when AI moves from shopping help to payment authority. [S3]
The likely adoption sequence is therefore information → recommendation → workflow assistance → bounded transaction, with different ceilings by category.
Source register
Primary and high-value sources used. Time-sensitive claims were checked against the latest available release as of the reference date. Commercial/vendor research is identified as such and is not treated as neutral market truth without corroboration.
[S1] Similarweb, AI Referral Traffic by Industry: 2026 Data (3 Sep 2026). Source — Worldwide web referrals June 2025–May 2026; seven industries.
[S2] Salesforce, Agentic Search Grows 200% (28 Jul 2026). Source — Consumer/commerce surveys plus >1.5bn shopper behavioural dataset.
[S3] Visa, Trust Index for agentic commerce (9 Sep 2026). Source — US consumer trust/delegation evidence.
[S4] Adobe, 2026 AI and Digital Trends — global hub. Source — Global consumer/CX research on AI-assisted journeys.
[S5] McKinsey, 2026 Global B2B Pulse. Source — B2B channel complexity and digital purchasing.
[S6] Google Search Central, AI features and your website. Source — Platform mechanics and visibility requirements.