Feedvisor on November 4, 2026 extended Agentis, the software it markets to brands selling on Amazon, to Amazon Business, Amazon's purchasing marketplace for organizations. The New York announcement says the platform can now separate business-buyer demand from consumer demand inside a brand's existing Amazon account and set advertising and pricing differently for each group. The only performance evidence Feedvisor offers is one unnamed office supplies brand, where business buyers accounted for 16% of sessions and 45% of revenue.

In Short

Feedvisor has added Amazon Business support to Agentis, a set of software agents that adjusts Amazon ads and prices for brands, so business buyers and ordinary shoppers can be managed as two separate groups. This matters to brands on Amazon because companies often buy the same products through the same listings and ads as consumers, and their larger, repeat orders get lost in the averages. The evidence so far comes from a single unnamed office supplies brand, with no time period, spend figures or control group disclosed.

What Feedvisor changed

According to Feedvisor, Agentis now brings "B2B demand, buying behavior and customer economics into advertising and pricing decisions across Amazon." The extension, in the company's description, lets a brand find where business demand sits in its catalog, see how those customers differ from consumers and act on the difference in both bidding and pricing.

The premise is simple. Business buyers on Amazon usually reach the same product detail pages and click the same sponsored placements as everyone else. They do not, however, behave the same way once they arrive. Order quantities, specification requirements, replenishment cycles, pricing structures and procurement workflows all change what a sale is worth, according to the release, and what it takes to win it.

"B2B isn't another audience segment. It's a different buying context," said Victor Rosenman, CEO and Founder of Feedvisor. "A business buyer may purchase in larger quantities, replenish more often, qualify for different pricing, and be worth considerably more over time. An agentic system should understand that context and fine-tune optimization decisions accordingly. That's what we've extended Agentis to do."

The release states that Agentis for Amazon Business "is now available." It does not give a price, a list of the Amazon Business countries covered, or the number of brands using it.

Four agents, one system

Agentis was built to connect advertising, pricing, demand, competition, inventory and margin in a single system, according to Feedvisor. The Amazon Business extension feeds B2B signals into that system rather than running them through a separate tool. Four agents divide the work.

The Demand Agent weighs B2B performance and customer economics alongside Sponsored Ads, DSP and Amazon Marketing Cloud signals. It decides where to invest to reach business buyers across the funnel. That covers bids, budgets and placements to capture existing demand, and it extends to DSP audience strategy and custom audiences built in Amazon Marketing Cloud to find business buyers who have not yet purchased.

The Margin Agent connects demand to pricing, competition, inventory and margin. Its job, in Feedvisor's description, is to identify where business pricing, quantity discounts and promotions are "supported by the economics." Put differently, it is meant to stop a brand from discounting for bulk orders where the margin cannot carry the cut.

The Intelligence Agent finds where business demand is concentrated and surfaces differences between business and consumer performance. The Orchestrator Agent reasons across the inputs from the other three to set what the release calls "context-aware strategy."

The structure has shifted since launch. When Feedvisor introduced Agentis on May 6, 2026, it described three components, according to the company's launch release: a Demand Engine for Sponsored Ads, DSP and Amazon Marketing Cloud bids, budgets and targeting; a Margin Engine for dynamic pricing; and an Intelligence Engine for competitor and SKU economics. Six months later the engines are called agents, and a fourth, the orchestrator, sits above them. The May release gave no customer counts, performance results or pricing.

Feedvisor's stated aim is a system that can "recognize when the same product represents two very different customer opportunities." The release does not say which Amazon reports, data feeds or account settings Agentis uses to tell a business buyer from a consumer.

The office supplies case

Feedvisor's evidence is an analysis of one office supplies brand's Amazon business. The brand is not named. Business buyers converted at 41%, according to the release, compared with 15% for consumer buyers. They generated 3.31x ROAS against 1.90x, and they bought 40% more units per order.

Those figures shaped the strategy. Agentis then managed B2B and consumer demand at the same time, according to Feedvisor, optimizing against each segment's performance, products and economics. The work was not left to the agents alone. Feedvisor's own team, "working alongside the platform," created dedicated Amazon Business campaigns, adjusted bids and placement multipliers, expanded Sponsored Brands Video and Sponsored Display, factored in B2B pricing and retail readiness, and measured business performance separately from consumer performance.

The outcome, as Feedvisor reports it: B2B orders rose 24% while non-B2B order volume stayed flat. The company calls that "a strong signal that the growth was additive rather than simply shifting existing consumer orders." Business buyers made up 16% of sessions and 45% of revenue.

How the numbers fit together

The figures are broadly consistent with one another. A PPC Land calculation, assuming the 41% and 15% conversion figures are orders per session, puts business buyers at about 34% of orders if they account for 16% of sessions. Applying the 40% higher units per order lifts their share of units to about 42%. A 45% revenue share would then require business buyers to choose slightly higher-priced items, which is plausible but not stated.

The same arithmetic, again PPC Land's own, converts the ROAS figures into advertising cost of sales: 3.31x is roughly 30% of attributed revenue spent on ads, and 1.90x is roughly 53%.

What the case does not show

Several gaps limit what the case can establish. The release gives no time period for the 24% rise in B2B orders, no session or order counts, no ad spend and no revenue in dollars. It does not define conversion rate, a metric whose denominator varies by platform and report. Without a definition, it is unclear whether the 41% counts orders or units per session, a distinction that matters when the same buyers are said to order 40% more units each time.

There is also a framing question. The release presents the 3.31x and 1.90x figures as a comparison between business and consumer buyers in the brand's existing account. Other Feedvisor material describing the same case presents 3.31x as the return after the rollout and 1.90x as the return before it. Those are different claims. One describes two customer groups; the other describes a change over time.

Is the B2B growth new demand, or demand borrowed from elsewhere? The additivity claim rests on flat consumer orders during an unspecified period. No control group or holdout test is described, so seasonality, catalog changes or the expanded Sponsored Brands Video and Sponsored Display activity could all have contributed. And because Feedvisor's team made campaign changes by hand alongside the agents, the release does not separate what the software did from what the account managers did.

Dani Nadel, President and COO of Feedvisor, framed the case as typical rather than exceptional. "For many brands, the B2B opportunity isn't something they have to create. It's already inside the business they have," Nadel said. "Business buyers are often purchasing through the same campaigns and the same listings, but their behavior and value disappear into the averages. Separate that demand, and you start to see a business you couldn't see before: a customer who buys in larger quantities, comes back more often, and returns less. That changes where and how you invest."

The release offers no figures on return rates or repeat purchase frequency to support the "comes back more often, and returns less" part of that statement. It does note that B2B purchasing tends to be concentrated in a subset of a brand's catalog, which Feedvisor presents as a reason to target those products rather than treat the whole catalog as B2B.

The market Feedvisor is pointing at

Feedvisor frames the extension against two third-party figures. Forrester projects that U.S. B2B e-commerce will reach $3 trillion by 2027, according to the release. McKinsey, it adds, found that 73% of B2B decision-makers are willing to spend up to $50,000 in a single online transaction, up from 59% in 2022. The release does not name the McKinsey survey or say when it was conducted.

Amazon's own figures are more directly relevant. Amazon Business reached $60 billion in annualized gross sales in the second quarter of 2026, with more than 11 million organizations buying through it worldwide and more than 1.8 million added in the first half of the year. In December 2025, the figures were $35 billion and more than eight million organizations. The division operates in 11 countries and launched in the United States in 2015.

Amazon has also been building B2B selling tools of its own. On August 31, 2026, it opened Business Deals to more than 11 million registered business organizations across 10 markets, with minimum discounts of 10% for regular deals and 15% for high-velocity events such as Black Friday and Cyber Monday. India, one of the 11 Amazon Business countries, was left out. In November 2025, Amazon Business launched an AI assistant for organizational purchasing for U.S. customers at no extra cost.

Third-party agents in Amazon's own agent era

The extension arrives as Amazon builds agents of its own. The company has spent the past year turning its own advertising stack into an agent product. Amazon introduced Ads Agent at unBoxed in November 2025 to automate campaign work across Amazon Marketing Cloud and its DSP. In February 2026 it opened its advertising APIs to AI agents through an MCP server in open beta. On September 29, 2026, it retired the Amazon DSP name, yet a month later the conversational Ads Agent was still in closed beta for eligible advertisers in 25 countries.

Amazon's push has also narrowed some manual controls. Full-Funnel Campaigns, with a $100 minimum daily budget, leave the split of budget across formats to Amazon. Against that backdrop, the case for a third-party tool rests on seeing a buyer segment that Amazon's own automation may blend together. Whether Amazon's agents will make the same split, and whether Amazon Business signals will remain available to outside platforms on the same terms, are questions the release does not address.

Data access is part of that picture. The Demand Agent depends on Amazon Marketing Cloud for custom audiences. Amazon made first-party paid features in Amazon Marketing Cloud free to query until the end of 2026 and has not said what happens from January 1, 2027.

The money at stake on the advertising side is substantial. Amazon's advertising revenue grew 11% to $19.64 billion in the third quarter of 2026, down from 26% growth in the second quarter, a slowdown Amazon attributed to Prime Day moving to June. Every vendor that promises a better return on Amazon ads is competing for a portion of those budgets.

Feedvisor's record on PPC Land

Feedvisor has made similar pricing-plus-advertising arguments before. In January 2019, it expanded its platform to combine its algorithmic repricer with advertising optimization on Amazon, reporting at the time that clients saw a 35% increase in impressions, a 42% rise in click-through rate and a 46% decrease in advertising cost of sales. Later that year it released real-time price optimization for private labels competing on Amazon, claiming a 30% average uplift in profit contribution and a 27% average increase in sales revenue for clients.

The thread connecting those announcements and this one is the same: pricing and advertising decisions on Amazon are made in separate tools by separate teams, and Feedvisor argues they should be made together. What has changed is the vocabulary, from algorithms to engines to agents. The segment this time is the business buyer.

For brands with a meaningful B2B tail on Amazon, the question the release raises is a practical one. If business buyers convert at more than twice the consumer rate and order more units, a bid tuned to the blended average undervalues one group and overvalues the other, in Feedvisor's framing. The office supplies case suggests the gap can be large. It does not show how often it is.

Timeline

Summary

Who: Feedvisor, an Amazon software company led by CEO and Founder Victor Rosenman and President and COO Dani Nadel, and the brands that sell to both consumers and business buyers on Amazon.

What: Feedvisor extended Agentis, its agentic commerce platform, to Amazon Business. Four agents (Demand, Margin, Intelligence and Orchestrator) now separate business-buyer demand from consumer demand and set advertising and pricing for each. In one unnamed office supplies case, business buyers converted at 41% against 15%, bought 40% more units per order and generated 45% of revenue from 16% of sessions, while B2B orders rose 24%.

When: The announcement was made on November 4, 2026. Agentis itself launched on May 6, 2026.

Where: Amazon, across Sponsored Ads, the DSP and Amazon Marketing Cloud. The announcement was issued from New York and does not say which of Amazon Business's 11 countries are covered.

Why: Business buyers on Amazon often use the same listings and ads as consumers, so their higher order sizes and repeat purchasing are averaged into consumer performance. Feedvisor argues that separating the two lets brands bid and price for each. The evidence rests on one unnamed brand, with no time period, spend or control group disclosed, at a time when Amazon is building its own advertising agents and B2B tools.