AWS Marketplace and the Composed Shelf: What Agentic Procurement Changes for ISVs and the Channel

Depending on geography, between 7% and 12% of SMB and midmarket buyers use a cloud marketplace to discover software. The rest arrive at AWS Marketplace, or at any of its competitors, already decided. A partner or an ISV brings them, and they transact there for contract consolidation, committed-spend drawdown, and procurement governance rather than for anything resembling search.

Call it the Discovery Deficit. Cloud marketplaces have functioned as procurement rails, not demand engines. They close deals that were originated somewhere else, by someone else, usually a partner.

That gap is why the AWS Marketplace agentic procurement announcements matter, and it is also why most coverage is aimed at the wrong question. Whether AI improves marketplace search is not interesting. Whether a marketplace that has never originated demand in the smaller segments can begin to do so, once the buyer stops being a person typing keywords, is a different question entirely, with different consequences for everyone downstream.

Three changes, and what each one is actually buying

AWS Marketplace has made three structural changes that are easy to read as feature releases. Read against the Discovery Deficit, each is doing something more specific.

The first is the replacement of lexical search with conversational discovery. Agent Mode, launched at re:Invent 2025, lets a buyer describe a requirement in natural language, upload an RFP or a requirements document, and receive ranked recommendations with side-by-side comparisons. Conversational search converts better than keyword search, which is unsurprising. The more important change is in what the interface is for. A keyword catalog fulfills a decision the buyer already made, and works only for someone who knows what to type. A conversational one helps make the decision, and deciding is the step SMB and midmarket buyers have always outsourced, because they have no procurement function to run comparative analysis internally. That is also why so few of them discover software in a marketplace: a catalog that cannot help you decide is little use to someone who cannot decide alone.

The second is building for machines to read rather than people. Most web pages assemble themselves in the browser, so a crawler or an agent that arrives sees almost nothing. AWS builds Marketplace pages to arrive complete, which means an agent reading one gets the whole listing. It has also opened the catalog to direct queries through an MCP server, so a buyer's own AI assistant can ask it questions without visiting a page at all. Most platforms building AI discovery are building a destination and trying to keep the buyer inside it. AWS is doing close to the opposite, and that choice says more about the strategy than anything else in the set. Making the catalog legible to agents AWS does not own is a distribution choice rather than an experience choice, and it concedes that the buyer's first conversation about software will happen somewhere else. The competitive unit shifts accordingly, from whose marketplace interface is best to whose catalog is most readable by someone else's agent.

The third is the automation of the transaction, which arrives from two directions at once. Express Private Offers let a seller define rate cards, discount tiers, volume breaks, and qualification criteria in advance, so an offer can be generated and accepted without a human negotiating it, which lowers the cost of serving a small software deal. In June 2026, AWS cut its professional services listing fee from 2.5% to 0.5%, and takes it to zero where those services are bundled into a multi-product solution or attached to a Migration Acceleration Program engagement, which lowers the cost of attaching services to that same deal. Multi-product solutions then let software, third-party components, and professional services sit in one listing under one procurement flow. A fee that vanishes at the exact moment services are attached to software is a routing incentive rather than price relief, and it routes toward the bundled listing as the default unit of sale. All of this aims at deals neither AWS nor its partners could previously work economically, which are the same deals where the Discovery Deficit lives.

Individually these read as product announcements; together they describe a platform trying to convert itself from a procurement rail into a demand engine, which is a considerably harder thing to be.

The Composed Shelf

Catalogs have always been curated at ingestion. Someone assigns the category, sets the tags, fixes the ranking, and every buyer who arrives afterward navigates decisions made before they showed up. That model breaks down as a catalog grows, because one fixed shelf has to serve buyers who have nothing in common.

Agentic discovery does none of that in advance. A buyer describes a problem, and the system builds a ranked set of products for that one request, drawing on a knowledge graph that holds what each listing does, what it has proven, and what it connects to. A different question produces a different set, and no master ranking sits behind them.

That is the Composed Shelf, and it changes who reaches the top. A fixed catalog rewards the biggest sellers through a loop: volume earns placement, and placement produces more volume. When the shelf is built for one question at a time, that loop has nothing to grip. What decides the answer is which product fits what this buyer asked for, and on a narrow request a small vendor can fit better than a large one.

Matching happens at the listing, which is where this gets awkward for most sellers. A listing built to cover an entire product line gives an agent nothing precise to match against. Sellers built them that way for a good reason: listings were expensive to create and maintain, so one broad listing beat five narrow ones. That arithmetic changed in June 2026, when AWS's AI-assisted listing tooling cut creation from weeks to minutes by generating structured listings from a seller's own site, documentation, and case studies. Several precise listings now cost about what one broad listing used to, and each becomes its own entry point. None of that shifted liability. It writes from material the seller provides, and the seller stays responsible for whether those product and compliance claims are true.

Inside each listing, what the graph actually reads is the agent card rather than the marketing narrative: structured metadata covering function, pricing mechanics, deployment model, integrations, and third-party validation. How that metadata is formatted matters less than which fields carry weight, and the fields that carry the most weight are the ones a seller cannot write for itself. Validated specializations, confirmed case studies, and demonstrated deployment patterns rank ahead of product copy.

That same weighting cuts against something the channel has long relied on. AWS states that internal partnership tier carries no weight in buyer-facing ranking, that tier governs how field teams engage a partner and nothing beyond that, and that position is earned on relevance and validated proof rather than on relationship classification. Partner tier has been meaningful currency in every major vendor program for two decades. It still buys field engagement, but it does not buy discovery, and as discovery shifts toward agents the share of demand that tier can influence keeps shrinking.

The channel economics, and the questions still in motion

Partners are hearing two things about their economics, and the two only make sense together.

The first is that automating high-volume, low-dollar renewals costs partners nothing. When an ISV moves tens of thousands of no-touch renewals to automated offers, it is describing contracts where a manual partner touch costs more than the margin on the renewal. That volume historically either churned or was serviced at a loss. Automating it captures revenue that was uneconomic to work by hand, and it frees partner capacity for expansion motions where partner margin actually lives. Where a partner owns the relationship, a Channel Partner Private Offer keeps that partner as seller of record, carrying its own integration, customization, and managed services at its own margin under its own contract.

That leaves the question partners actually care about, which is, who decides whether they are in the deal at all. Express Private Offers do not make that call. The ISV does, the way it always has, based on the relationship. A partner who gets cut out of a deal has been cut out by the software vendor, not by the marketplace, and that is a conversation to have with the vendor.

The second thing partners are hearing concerns matching. As AWS routes partner recommendations into opportunities algorithmically, that reach extends into exactly the long-tail segments where field coverage is thinnest.

Put side by side, the two claims say something neither says alone. If the long tail were purely uneconomic for the channel, it would not warrant an attach mechanism at all. The mechanism exists because the volume has value. What was uneconomic was originating those deals by hand, and automated matching lowers origination cost rather than servicing cost. That points partners at a different investment than the one they may be defending. The opportunity is not holding onto renewal administration but being findable at the moment a match is made, which rewards listing precision and validated proof over field relationships.

Several things partners will ask about are not settled yet.

Express offer mechanics do not yet extend to Channel Partner Private Offers. Once they do, a partner will be able to generate and close automated offers from its own storefront rather than routing every one through the marketplace console.

The ranking signals also repay a close reading. Asked directly whether the graph weighs deployment, actual consumption, or customer retention, AWS answered with a different list: match to the buyer's described need, AWS specializations, validated case studies, service coverage, and buyer engagement. None of the three outcome measures appears in it. That may reflect telemetry not yet wired in rather than a decision against it, but partners should plan against what is weighted today rather than what could be.

And moving traditional consulting partners into a marketplace motion is a multi-year journey rather than a switch, comparable to the arc SaaS took before it became a normal marketplace category. That is AWS's own characterization, and it tells partners they are early rather than late.

What this asks of vendors and partners

For ISVs, metadata is now the customer acquisition surface. That means decomposing a broad offering into precise, use-case-specific listings, structuring each for machine parsing, and investing in the validated signals a knowledge graph weighs most heavily and a competitor cannot fabricate. Free-trial and product-led motions matter more than they did, because they generate usage evidence quickly and without an enterprise sales history.

For the channel, margin is migrating out of renewal administration and into the work of assembling and running the solution. Zero-fee treatment for bundled professional services exists to make attached services the default rather than the exception, and partners who continue to sell integration expertise outside the transaction flow are working harder for the same money. Recurring revenue in this model comes from custom integration, from offerings a partner packages once and sells many times, and from managing the financial dimensions of consumption-based AI pricing, which most SMB and midmarket buyers cannot model on their own and will not attempt to.

Co-sell is not being displaced. Discovery is tipping quickly, while transaction and delivery stay human-led because that is where trust, customization, and services live, so co-sell gets fed better-qualified demand that an agent originated. AWS frames the open question for the next 18 months as who curates the shelf when the buyer is an agent, and says it now treats curation as a product rather than a background function.

And that returns to the Discovery Deficit, which is the number to watch. If the share of SMB and midmarket buyers discovering software through a marketplace moves meaningfully off 7% to 12% over the next four quarters, the agentic thesis is validated and the origination economics of the channel change with it. If it does not move, then the marketplace has automated its transaction layer without ever having solved the demand problem, and partners remain the front door regardless of how sophisticated the catalog behind it becomes.

AWS did not have to do any of this. The transaction rail was already working and profitable, and going after buyers who were never there is the harder bet.