Consumer intelligence company NIQ and digital data provider Similarweb today announced a collaboration to build a measurement solution connecting AI-driven product discovery to verified sales outcomes, with an initial version due in the fourth quarter of 2026.

In short

NIQ and Similarweb are building a joint measurement product that tracks how AI assistants such as ChatGPT and Gemini influence what people buy, from the first question a shopper asks to the sale that follows. It matters because brands currently have almost no way to tell whether an AI platform mentioned their product, sent traffic their way, or helped close a sale, even as more shopping journeys start inside a chat window rather than a search box. The practical change arrives in the fourth quarter of 2026, when an initial version launches in a limited set of categories and markets, expanding from there.

A gap opens between discovery and the checkout

Shopping used to leave a trail. A person searched, clicked, landed on a product page, and either bought something or did not, and most of that trail could be reconstructed afterward with a reasonable degree of confidence. Agentic commerce breaks the trail in a new place. According to NIQ, AI agents are now moving beyond simple product discovery to become part of the purchase itself, with new infrastructure enabling consumers to complete an entire buying journey, from item discovery to evaluation and recommendation to purchase, inside a single AI experience.

Two pieces of infrastructure make that possible. Google's Universal Commerce Protocol and OpenAI's Agentic Commerce Protocol both let an AI system carry a shopper from a question through to a completed transaction without the usual detour through a browser tab or a retailer's own checkout flow. When that happens, the referral data, the click, and the session that advertisers have relied on for two decades of digital measurement simply do not get generated in the familiar form.

That absence is the problem NIQ and Similarweb say they are trying to solve. According to NIQ, the work adds a measurement component to its Commerce Intelligence strategy, focused on deploying the company's AI-ready content directly into agentic commerce workflows and equipping brands and retailers with the signals needed to react effectively. Troy Treangen, Chief AI and Product Officer at NIQ, framed the commercial stakes plainly: "AI is becoming a new commerce channel, and NIQ intends to make it measurable. Our clients want to know where AI is already influencing their business, how quickly that influence is growing and what they should do about it."

What each company brings to the table

The division of labor in the announcement is fairly clean. NIQ supplies product intelligence, consumer behavior data and retail sales measurement, the kind of information that historically comes from point-of-sale records, retail panels and syndicated market research. Similarweb supplies what the release calls foundational digital data, meaning visibility into what happens on the open web and inside AI platforms themselves, before and after a purchase decision gets made.

Susan Dunn, Chief Revenue Officer at Similarweb, described the handoff between the two data sets: "AI's influence does not stop when a consumer leaves an AI platform. Similarweb provides the foundational digital signals that reveal what happens across the entire AI-driven consumer journey. Combined with NIQ's product and sales intelligence, those signals can give brands and retailers a clearer view of how AI-driven discovery translates into commercial outcomes."

Treangen made a similar point from the other direction: "NIQ has the AI-ready data foundation and the product, consumer and sales intelligence needed to answer those questions. With Similarweb, we can extend that intelligence across the AI-influenced purchase journey: from discovery to traffic, conversion and sales."

Put together, the pitch is that NIQ can tell a client what actually sold, while Similarweb can tell that same client what happened on the path to the sale, including the parts of the path that now run through a chatbot rather than a search engine results page.

Five areas of measurement, one missing number

According to NIQ, the solution will focus initially on five areas. The first is consumer intent, meaning what people are asking AI assistants and which needs, questions or prompts shape their agentic commerce decisions. The second is what the company calls agentic shelf visibility, covering where products appear when an AI assistant recommends options to a shopper and how a brand's presence compares against competitors in that recommendation.

The third area is product content readiness, which asks whether a brand's product content is complete, structured and optimized enough for an AI assistant to find, understand and accurately recommend it. The fourth is AI-driven traffic, meaning how much of the traffic reaching a brand or retailer's product pages is driven by AI platforms, counting both direct clicks and the subsequent visits that agentic discovery influences indirectly. The fifth is AI-driven conversion, which connects that AI-influenced engagement to verified purchasing behavior across channels.

None of the five areas comes with a published methodology, a sample size, or a data-collection window in the material NIQ and Similarweb released today. That is not unusual for an early-stage product announcement, but it does mean the areas function, for now, as a stated intention rather than a measurable capability. NIQ said more information, including initial categories, markets and availability, will be announced as development progresses.

The platforms in scope and the timeline that is not

The companies named five AI platforms the new solution is expected to cover: ChatGPT, Gemini, Google AI Mode, Perplexity and Claude. That list covers the platforms most commonly cited in industry discussion of AI-driven shopping, though it does not include every conversational commerce surface currently operating, and NIQ did not specify whether coverage across the five would be uniform at launch or added incrementally.

On timing, NIQ was specific about the quarter and vague about nearly everything else. An initial version of the solution is due in the fourth quarter of 2026, beginning with what the company called a focused set of categories and markets and expanding over time. Asked directly in the accompanying FAQ material whether the measurement is available today, NIQ said no, and reiterated that specific launch timing and market availability were not being announced at this stage.

That degree of reticence extends to commercial terms. No pricing model was disclosed. No named launch clients appear in the announcement. No sample categories or countries were specified beyond the general reference to a focused initial set. For a marketing team trying to decide whether to wait for this measurement layer or build an interim workaround with existing tools, the announcement supplies a quarter and a direction of travel, and not much else.

How this fits NIQ's wider AI portfolio

NIQ has been building toward this kind of product for some time, and the announcement is explicit that Agentic Commerce Measurement sits alongside, rather than inside, two other initiatives already underway. According to NIQ, NIQ Optiq and its Optiq Bridge component are intended to bring NIQ intelligence into AI-powered applications and enterprise environments, while ConnectAI brings NIQ's own technology, engineering and data science teams together with clients to build AI-enabled workflows. Agentic Commerce Measurement, in this framing, extends that broader strategy into Commerce Intelligence specifically, by helping organizations understand how AI is influencing consumer discovery, traffic, conversion and sales.

That framing lines up with reporting on NIQ's other AI moves. Five days before today's announcement, NIQ said it was working with The OpenAI Deployment Company to extend the Optiq product line into enterprise systems, while stating in the same document that none of its proprietary data becomes available to the public through ChatGPT. That earlier arrangement covers Optiq Chat, Optiq Mobile and Optiq Bridge, with expanded versions of Chat and Bridge due in early September 2026, and it left several governance questions open, including what an authorized AI agent may retrieve from NIQ's data and how usage would be metered or priced.

The scale of the data foundation behind both efforts is the same. NIQ's proprietary data runs on what the company has described as a 160-petabyte engine spanning roughly 260 million product items and 10.5 billion product attributes, drawing on more than 7.4 trillion dollars in global consumer spend tracked across more than 90 countries. Agentic Commerce Measurement would sit on top of that same foundation, adding the discovery-to-conversion layer that the OpenAI enterprise arrangement does not itself provide.

NIQ's chief executive statement in the release frames the new measurement work as a natural extension of that portfolio: Agentic Commerce Measurement adds a new capability to NIQ's growing AI portfolio, intended to help clients see and act on AI's commercial influence in the marketplace, according to NIQ.

The broader measurement gap this is meant to close

The problem NIQ and Similarweb describe is not unique to their announcement. Retail media networks and platform vendors across the industry have spent much of 2026 acknowledging that standard measurement, built for a world of clicks and sessions, does not map cleanly onto a world where an agent completes a transaction on a shopper's behalf. NIQ itself has been narrowing this gap from the retail measurement side for over two years. The company's Discover platform, launched in July 2024, consolidated consumer panel and retail measurement data onto a single cloud platform, and NIQ later added a global data clean room on Snowflake in October 2025 to support first-party data enrichment and outcome measurement inside a governed environment.

More recently, NIQ reported 34 percent growth in revenue from AI-native solutions in its second quarter of 2026, alongside quarterly revenue of 1,124.2 million dollars, with roughly 51 percent of the company's top 100 clients using at least one AI-native solution. That growth figure applies to a defined product category, the BASES AI suite and Retailer Analytics offerings, and does not cover Agentic Commerce Measurement, which is a separate and newer initiative not yet contributing revenue.

The wider industry context helps explain why NIQ and Similarweb are moving now. Retail media has grown into what NIQ has separately described as a 184 billion dollar global market, and AI-based product discovery already accounts for a meaningful share of how shoppers begin their research, according to consumer research NIQ has published elsewhere. Other measurement vendors have been building comparable capabilities from adjacent angles: LiveRamp began piloting agentic AI in June 2026 to close what it called the workflow gap between commerce data and campaign outcomes, while a separate startup, Minerva, raised 20 million dollars the same month alongside its own OpenAI collaboration aimed at restructuring marketers' first-party data. Experian has taken a related approach, pairing identity resolution with activation inside governed client environments rather than moving data between systems.

Practitioners in the region have already flagged the shortfall this new product line is meant to address. At an IAB summit in Australia in June 2026, industry speakers said the retail media sector's measurement debt was being incurred at the same moment that agentic commerce began to arrive, leaving buyers with standardization gaps in a market already worth an estimated 2 billion dollars in Australia alone. European buyers have voiced a similar concern through IAB Europe, which continues to list measurement and attribution as the two areas of retail media most in need of standardization, and which already tracks NIQ among the attribution and measurement suppliers active in that market, alongside firms such as dunnhumby, Criteo, Epsilon and LiveRamp, according to the organization's retail and commerce media landscape map.

What the announcement does not say

Several details that a marketing or measurement team would need before committing budget are absent from today's release. There is no confirmation of which product categories will be included when the initial version launches in the fourth quarter, nor which countries or regions will see the first rollout. There is no pricing structure, whether that means a subscription tier, a consumption-based model, or a bundled add-on to existing contracts. There is no mention of a beta program, an early-access group, or named launch clients tested ahead of general availability.

The FAQ material accompanying the release does answer a few structural questions clearly. Asked how the new measurement relates to NIQ Optiq and ConnectAI, NIQ described the three as addressing different but connected parts of its AI strategy, with Optiq and Optiq Bridge focused on bringing NIQ intelligence into AI-powered applications, ConnectAI focused on embedding NIQ's own engineering and data science talent with clients, and Agentic Commerce Measurement focused specifically on connecting AI-driven discovery to commercial outcomes. Together, according to NIQ, these three efforts are meant to connect AI-ready data to decision, action and measurement.

That structural clarity does not extend to the technical mechanics of how the five stated measurement areas will actually be calculated. It is not stated how agentic shelf visibility will be defined or scored across platforms with very different interfaces, since a Perplexity answer card and a ChatGPT product recommendation do not present information in comparable formats. Nor is it stated how AI-driven traffic will be distinguished from ordinary referral traffic when a platform strips tracking parameters or routes a shopper through an anonymized session.

Why this matters for the marketing community

For advertisers, retailers and agencies, the significance of today's announcement lies less in what NIQ and Similarweb have built so far, which by their own account is not yet available, and more in what the announcement signals about where measurement budgets and attention are likely to move over the next several quarters.

The immediate practical implication is a widening gap between brands that already have some visibility into AI-driven shopping behavior and those that do not. NIQ's own reporting from earlier in the year found that a majority of its largest clients were already using at least one AI-native solution, and that adoption among clients grew sharply year over year. A brand or retailer waiting for a fully formed measurement product before engaging with agentic commerce at all risks falling behind competitors who are already instrumenting their own interim tracking, however imperfect that tracking might currently be.

There is also a governance question that echoes NIQ's separate arrangement with OpenAI's deployment arm. That earlier collaboration left open questions about what an authorized AI agent can retrieve from a supplier's retail measurement data, how that access gets audited, and what happens if licensed data resurfaces inside a downstream AI output. Agentic Commerce Measurement raises a parallel set of questions from the opposite direction: rather than asking what an AI system may retrieve from NIQ, it asks what NIQ and Similarweb will be permitted to observe about consumer interactions with AI systems that are not their own, and how that observation is disclosed to the platforms and consumers involved. Neither today's release nor the earlier OpenAI arrangement addresses that question directly.

Finally, the concentration of capabilities inside a small number of measurement vendors is itself worth tracking. NIQ already holds retail measurement, purchase-based audiences, a data clean room, programmatic distribution partnerships and an outcome-measurement contract with at least one major European retail media joint venture. Adding an agentic commerce measurement layer built with a major digital data provider extends that footprint further. That is not, on its own, a reason for concern, but questions about independence and potential conflicts of interest, already being asked of measurement suppliers in adjacent categories such as television and attention measurement, are likely to follow this product as it moves from announcement to availability.

Timeline

Summary

Who: NIQ, the consumer intelligence company listed on the New York Stock Exchange under the ticker NIQ, and Similarweb, the digital data company listed on the New York Stock Exchange under the ticker SMWB. NIQ's Chief AI and Product Officer, Troy Treangen, and Similarweb's Chief Revenue Officer, Susan Dunn, both provided statements in the announcement.

What: A collaboration to build an Agentic Commerce Measurement solution that connects AI-driven product discovery to consumer intent, product content readiness, AI-driven traffic and verified sales conversion. The solution is intended to cover five AI platforms, ChatGPT, Gemini, Google AI Mode, Perplexity and Claude, and to measure five areas: consumer intent, agentic shelf visibility, product content readiness, AI-driven traffic and AI-driven conversion. No pricing, launch markets, launch categories or named early clients were disclosed.

When: Announced on September 2, 2026. An initial version of the solution is scheduled for release in the fourth quarter of 2026, beginning with a focused set of categories and markets before expanding further.

Where: NIQ is headquartered in Chicago, with operations in more than 90 countries. Similarweb provides digital data used globally. The measurement solution itself is not tied to a single geography, though its initial fourth-quarter release will begin in an unspecified limited set of markets.

Why: As AI agents increasingly complete purchases without generating the clicks, referrals and sessions that conventional analytics tools rely on, brands and retailers have limited visibility into whether AI platforms are representing their products accurately or driving measurable sales. Today's announcement signals that two established measurement vendors intend to build a joint solution addressing that gap, though the product itself remains under development, with material details, including price and precise rollout timing, still unannounced.