Quick answer: Not for the content that matters most. Agentic commerce is real and growing — ChatGPT now offers shopping research built on GPT-5 mini, and Google has turned on agentic checkout with a "Buy for me" button inside AI Mode and Gemini, live at select US retailers (commercetools). But these agents pull structured product data — price, availability, specs, schema — not editorial prose, for the narrow slice of queries that are already transactional ("buy X," "best price on Y"). Research, comparison, and expertise content that helps someone decide what to buy, or which brand to trust, in the first place is a different job entirely — and it's a job AI shopping agents still can't do without content like it existing somewhere for them, and the models behind them, to find.
It's an easy headline to write: AI agents will do the shopping, so nobody needs your blog anymore. It's also premature. Agentic commerce is genuinely emerging in 2026 — not hype — but it's concentrated at one narrow point in the funnel, and conflating "agents can complete a purchase" with "agents have replaced the reason someone chose to purchase" misses where the real risk sits.
What is agentic commerce, actually, in 2026?
Agentic commerce is AI systems doing more than answering a question — they browse, compare, and in a growing number of cases, complete the purchase itself. Per commercetools, three platforms moved this from concept to product in the past year:
- Perplexity launched a shopping experience with conversational product discovery, personalized product cards, and instant checkout powered by PayPal.
- OpenAI introduced shopping research in ChatGPT, built on GPT-5 mini, for comparative product guides and deep research across retail sites.
- Google launched agentic checkout across Search's AI Mode and Gemini, with a "Buy for me" button now live at select US retailers, letting an agent execute the purchase directly on a merchant's site.
That last point matters: it's autonomous execution, not just recommendation. The agent isn't sending a shopper to a page anymore — it's completing the transaction itself.
How many shoppers are already using this?
More than most marketers assume. commercetools reports 73% of shoppers already use AI somewhere in their shopping journey — 45% for product ideas, 37% to summarize reviews, and 32% for price comparison — and 70% say they're at least somewhat comfortable with an AI agent making a purchase on their behalf. Morgan Stanley's projection, cited in the same piece, is more striking still: nearly half of online shoppers will use AI shopping agents by 2030, accounting for roughly 25% of their spending.
| Metric | Figure |
|---|---|
| Shoppers already using AI somewhere in their shopping journey | 73% |
| — using it for product ideas | 45% |
| — using it to summarize reviews | 37% |
| — using it for price comparison | 32% |
| Comfortable with an AI agent purchasing on their behalf | 70% |
| Online shoppers projected to use AI shopping agents by 2030 (Morgan Stanley) | ~50%, ~25% of spend |
Source: commercetools, "AI Trends Shaping Agentic Commerce".
Which content is actually at risk
The exposed layer is narrow and specific: transactional, bottom-of-funnel content whose entire job is answering "which one, at what price, buy now." Best-price roundups, spec-comparison tables built purely to drive a click-to-buy, and thin "top 10 products" posts are exactly what a shopping agent now shortcuts — commercetools is explicit that this depends on "structured data, enriched metadata and clean catalogs" so an agent can recommend a specific SKU. An agent doesn't need your paragraph explaining why a product is great; it needs a clean feed with price, stock status, and specs. If that's the entire value your content offered, an agent will route around it.
Which content isn't — and why
Everything upstream of "buy now" is a different problem, and it's one agents are nowhere close to solving. Per commercetools, "most agentic capabilities cluster at the start of the sales funnel" — browsing, discovery, matching — which still depends on a shopper (or an agent acting for them) forming an opinion about which brand to even consider. That opinion doesn't come from a product feed; it comes from research, comparisons, and first-hand reviews — the kind of content that gets discovered and cited by AI search engines in the first place.
This is also where brand consideration gets decided before an agent is ever asked to check out. If an AI system has never encountered your brand described anywhere credible — a comparison post, a Reddit thread, an independent review — it has nothing to recommend. Reddit's dominance as an AI-cited source is the same phenomenon from a different angle: agents lean hardest on content showing real experience and independent consensus, not a product feed. Editorial, expertise-driven content is how a brand earns a place in that consideration set — a job structured data cannot do alone.
The honest hedge: this is early, not mainstream
None of this is fully arrived. "Buy for me" is live at select US retailers, not universally; commercetools's own framing is that agentic capability today clusters at browsing and discovery, with checkout automation still expanding retailer by retailer. The more aggressive figure in the space — 90% of B2B buying being AI-agent-intermediated by 2028, over $15 trillion in spend — is a forward projection, not a 2026 reality. Treat agentic commerce like any emerging channel: real enough to prepare for, not mature enough to bet a whole content strategy on either direction.
What to actually do about it
- Clean up your structured data now. Product schema, accurate pricing feeds, and clean catalogs are the baseline an agent needs to recommend you at all in a transactional query — not optional, and not a content-writing problem.
- Keep investing in research and comparison content. This is the layer agentic commerce doesn't touch — still where consideration gets built, still what gets a brand cited by name before a shopping agent ever gets involved.
- Build topical authority, not one-off posts. A single article is easy to route around; a cluster of connected content is what builds topical authority a model actually recognizes.
- Measure past last-click. If an agent completes checkout on a retailer's site, last-click will undercount the post that built the consideration weeks earlier — the blind spot covered in how to measure content ROI in 2026.
- Distribute the same expertise everywhere. Research behind a strong comparison post doesn't have to live only as a blog post — turning it into short-form video keeps it visible on surfaces agentic commerce hasn't touched.
Frequently Asked Questions
Will AI shopping agents replace blog traffic entirely?
No — they replace the narrowest slice of it: pure transactional, price-comparison content whose only job was routing a click to checkout. Research and expertise content that shapes which brand gets considered remains largely untouched, and still needs to exist somewhere for an agent's model to find it.
Is agentic commerce actually mainstream yet in 2026?
No. commercetools' own reporting describes agentic checkout as live at select US retailers and most agentic capability as still clustered at browsing and discovery, not full-funnel autonomous buying — an emerging pattern worth preparing for, not a completed shift.
What should I do first if I'm worried about agentic commerce?
Fix your structured data before you touch your content calendar. Clean product schema and accurate feeds are the prerequisite for an agent to recommend you in a transactional query — no editorial content compensates for a missing or broken feed.
Does this mean I should stop writing comparison or "best of" content?
Only the thin, purely transactional version. Comparison content built on real usage and first-hand detail is exactly the research layer agentic commerce doesn't replace — it's what still earns a brand its place in an agent's consideration set.
How does being cited by AI search engines connect to agentic commerce?
An agent still has to learn a brand exists and is worth recommending before checkout — and that discovery runs through the same mechanics covered in how AI search engines discover and cite content. Skip that layer, and clean product data alone won't get you considered.
Prepare for both halves of the shift
Agentic commerce is going to keep eating the purely transactional end of search — that part is close to settled. What it doesn't touch is the research and trust-building work that decides which brand gets considered before a checkout button appears, and that work still runs through content built to be found, cited, and trusted by AI systems. See how ButterBlogs structures content for AI search optimization — schema, structure, and cited evidence that keep your brand in the consideration set agents draw from, not just the product feed they check out from. Check pricing to start building that layer now, while it's still early.