Yesterday’s announcements divide along a single question: can a marketing organization buy the surface under discussion. Azoma, the agentic commerce optimization platform, reported on August 28, 2026 that Walmart Chief Executive Officer John Furner told the company’s Q2 earnings call that shoppers using Walmart’s Sparky assistant spend 40% more per order than shoppers who do not, and that the number of customers using Sparky rose 70% year over year. Azoma states that Sparky offers no paid placement. A Chief Marketing Officer (CMO) reading those two figures together now owns a channel with a measured revenue premium and no line item that funds it.
The Trade Desk and ActiveCampaign spent August 27, 2026 automating work on inventory their customers already buy. The Trade Desk (NASDAQ: TTD), the advertising technology company behind the Kokai demand-side platform, introduced Kokai Zuma on August 27, 2026, adding a conversational Koa Assistant, specialized agents for audience creation, optimization, troubleshooting and campaign insight, and a rebuilt forecasting engine. The Trade Desk reported an average 32% improvement in cost per acquisition (CPA) in initial results from the new modeling and forecasting systems. ActiveCampaign released Active Intelligence: Wavelength on August 27, 2026, drawing on more than 500 signals from each customer’s own account history to draft campaigns, build segments and flag broken automations, and reported that businesses using Active Intelligence see a 75% higher engagement rate than businesses that do not. Both companies sell speed on channels that already have an auction or a send button attached.
Capital moved the other way across the same two days. Owner, which describes itself as the AI CMO and CTO for local businesses, announced on August 28, 2026 that it raised $240 million at a $2.3 billion valuation in a round led by Growth Equity at Goldman Sachs Alternatives, with Meritech, Redpoint, Headline and Jack Altman participating. Owner stated it has passed $100 million in annual recurring revenue since launching in 2020, and that independent restaurant owners will drive more than $1 billion in sales through the platform during 2026. Levanta, the creator affiliate platform for e-commerce, announced a $22 million Series B on August 27, 2026 led by Volition Capital, bringing total funding above $43 million against 2026 revenue growth of 80% year over year. Advisar, an AI-enabled decisioning layer for retail suppliers, announced a $750,000 pre-seed round on August 27, 2026 led by Clutch VC. None of those three companies sells media. Each sells a position inside a surface where the purchase decision now happens.
NIQ supplied the demand-side number that connects the two groups. NIQ, the consumer intelligence company formerly operating as NielsenIQ, released a cut of its report A Tale of Two Consumers: The Polarized Mindsets Reshaping Global Consumption on August 27, 2026, produced with World Data Lab, reporting that nearly three-quarters of shoppers now use AI for product discovery and that retail media has grown into a $184 billion global market. NIQ also reported that nearly one-third of Western consumers have purchased a product after discovering it on a social platform, that 68% of North American consumers have never purchased through social media, and that nearly 60% of Asia-Pacific consumers shop through social and quick-commerce channels. The retail media figure describes spend a marketing team can direct. The discovery figure describes behavior that runs partly outside it.
The mechanism that decides visibility on those unbought surfaces sits in Azoma’s citation data. Azoma reported on August 28, 2026, from an analysis of millions of shopping agent citations in Q2 2026, that Walmart Sparky draws 36% of its cited sources from earned media, 30% from brand-owned sites, 27% from retailer content, 5% from user-generated content and 2% from affiliate sites. In a companion release dated August 26, 2026, Azoma put the ChatGPT mix at 41% earned media, 37% retailer, 19% user-generated content and 3% brand-owned sites. Read against a retail media plan, both distributions say the same thing: the retailer listing a brand controls and pays for addresses roughly a quarter to a third of what these agents cite. Generative engine optimization (GEO) work on those surfaces routes through public relations and owned content, and the budgets for both usually sit outside the media team.
Work the arithmetic on a mid-size consumer packaged goods brand running $40 million a year through Walmart. Suppose Sparky-attributed orders reach 5% of that volume during 2027, which puts $2 million through a surface carrying no ad inventory, and the 40% basket premium Furner described raises the incremental contribution above what the share alone implies. Against Azoma’s Q2 2026 Sparky mix, the two levers that move visibility are earned media at 36% and brand-owned content at 30%. In most organizations those budgets belong to communications and to the web team. The retail media relationship belongs to the shopper marketing lead. Someone has to fund a channel that appears in no media plan and reports through no existing owner, and that decision lands before the 2027 planning cycle closes rather than after.
The funding rounds sharpen the same point from the supply side. Owner’s pitch replaces the marketing function outright for operators who never staffed one, running websites, ordering, customer relationship management, point of sale and phone ordering through AI agents. Levanta builds inventory inside the creator economy, reporting more than 90,000 vetted creators and adding integrations with ChatGPT and Claude alongside automated product sampling and flat-rate collaborations carrying affiliate-level measurement. Advisar addresses the supplier side of the commerce media network relationship, where brands fund retail media without independent measurement of what the spend returns. Investors priced all three against surfaces that sit outside the auction.
What these releases do not prove
Every performance figure in this edition carries vendor authorship. The Trade Desk’s 32% average CPA improvement describes initial results from its own modeling systems, with no disclosed sample size, campaign count, test period, or control against the prior Kokai release. ActiveCampaign’s 75% engagement lift compares businesses using Active Intelligence against businesses that do not, which is a self-selected population, and ActiveCampaign discloses no baseline, no matching method, and no definition of engagement. ActiveCampaign’s claim that 84% of marketers run generic campaigns built on industry best practices arrives with no stated survey, sample, or field date.
Azoma sells optimization for the surfaces it measures, so its citation mix analysis carries a commercial interest in showing that retailer listings alone leave a gap. Azoma discloses neither the query set, the categories sampled, nor the method used to attribute a citation to a source type. The Walmart figures Azoma cites are separate and stronger, since Furner stated them on a public earnings call, though Walmart has disclosed no methodology for how it attributes an order to Sparky use, and shoppers who choose an AI assistant may already be higher-spending shoppers.
NIQ’s 74% AI discovery figure appears without a stated sample size, market list, or field date in the August 27, 2026 release. NIQ has now published at least three cuts of the same Tale of Two Consumers research program during August 2026, including a private label finding dated August 19, 2026, so these figures describe one dataset viewed three ways rather than three independent studies. Levanta’s 80% revenue growth and Owner’s $100 million ARR are company-stated and unaudited. Advisar’s $750,000 round funds a product with no disclosed customers or results.
None of these releases establishes that AI-mediated discovery converts at rates that justify moving budget out of paid channels. The Sparky and ChatGPT source distributions describe what the agents cite, which is upstream of what shoppers buy.
Strategic priorities for marketing leaders
- Name an owner for agent-surface visibility before Q4 planning closes. The Azoma source mixes put 36% to 41% of agent citations in earned media, which sits with communications rather than media buying in most structures.
- Ask The Trade Desk for the test design behind the 32% CPA figure. Request campaign count, vertical mix, test window, and the control condition before you model Zuma into 2027 acquisition targets.
- Instrument Sparky and Rufus referrals separately from general AI traffic. Walmart’s 40% basket premium describes orders, and your analytics team can only confirm it against your own order data.
- Audit which agent-cited source types you actually control. Retailer content covers 27% of Sparky citations and 37% of ChatGPT citations, so a brand optimizing only its retail listings addresses roughly a third of the surface.
- Decide whether creator affiliate inventory becomes a measured line or stays in experimentation. Levanta’s flat-rate collaborations with affiliate-level measurement give finance a comparable number, which removes the usual reason to keep creator spend unbudgeted.
- Set an evaluation standard for vendor-stated engagement lifts. ActiveCampaign’s 75% figure compares adopters against non-adopters, and your procurement team can require a matched cohort before that number enters a business case.
Press Release Roundup: August 27–28, 2026
The Trade Desk Introduces Kokai Zuma, the Latest Release of Kokai The Trade Desk (NASDAQ: TTD), a global advertising technology company, introduced Kokai Zuma on August 27, 2026, the newest release of the Kokai platform it launched in 2023 for planning, buying and measuring advertising across the open internet. Zuma adds a conversational Koa Assistant to the Kokai dashboard in open beta, with specialized agents covering audience creation, optimization, troubleshooting and campaign insight, and The Trade Desk plans wide availability of the assistant by the end of 2026. The company rebuilt the platform’s forecasting engine and AI infrastructure to predict available inventory, model campaign outcomes and power Koa in real time, and states that Koa analyzes more than 20 million ad impressions per second. The Trade Desk reported an average 32% improvement in cost-per-acquisition performance in initial results from the new modeling and forecasting capabilities. Zuma also brings Conversion Lift enhancements, a more flexible Report Builder, bulk editing across large numbers of ad groups, and a modeling feature that projects the impact of a change before a trader applies it, and it makes the platform’s periodic-table view optional rather than the default. Founder and Chief Executive Officer Jeff Green stated that AI is helping customers move faster while making measurement more intuitive and actionable. Read the full press release (Published: August 27, 2026 | Source: Business Wire)
74% of Shoppers Use AI for Discovery: NIQ Showcases What That Means for the Consumer Purchase Journey in New Report NIQ, the consumer intelligence company formerly operating as NielsenIQ, released a new cut of A Tale of Two Consumers: The Polarized Mindsets Reshaping Global Consumption on August 27, 2026, produced in collaboration with World Data Lab. NIQ reported that nearly three-quarters of shoppers now use AI for product discovery and that retail media has grown into a $184 billion global market. On regional commerce behavior, NIQ reported that nearly one-third of Western consumers have purchased a product after discovering it on a social platform, that 68% of North American consumers have never purchased through social media, and that nearly 60% of Asia-Pacific consumers shop through social and quick-commerce channels. Troy Treangen, Chief AI and Product Officer at NIQ, stated that consumers are making purchasing decisions earlier and through more channels than before. NIQ positions the findings as building on its earlier report The Commerce Revolution: Where East Meets West. The release does not disclose sample size, market coverage or field dates for the AI discovery figure. Read the full press release (Published: August 27, 2026 | Source: Business Wire)
ActiveCampaign Launches Active Intelligence: Wavelength, Marketing AI Tuned to Your Business ActiveCampaign, a marketing automation platform for growing businesses, launched Active Intelligence: Wavelength on August 27, 2026, the first named release of its Active Intelligence layer. Wavelength draws on more than 500 signals from each customer’s own account, including account history, performance patterns and customer data, and then drafts campaigns, builds segments and identifies broken automation journeys before a marketer begins planning. ActiveCampaign stated that 84% of marketers still run generic campaigns built on industry best practices rather than their own data, and reported that businesses using Active Intelligence see a 75% higher engagement rate than businesses that do not. ActiveCampaign’s partner documentation describes the release as account-aware intelligence for agency-managed accounts, surfacing recommendations specific to each client rather than to a category benchmark. The company discloses no sample size, matching method or engagement definition behind the 75% figure. Read the full press release (Published: August 27, 2026 | Source: Business Wire)
Levanta Raises $22M Series B from Volition Capital to Deploy the Infrastructure for Creator Commerce Media Levanta, the creator affiliate platform for modern e-commerce, announced a $22 million Series B investment on August 27, 2026 led by Volition Capital, which also led the company’s $20 million Series A in 2024. The round brings Levanta’s total funding above $43 million, and the company reported 2026 revenue growth of 80% year over year. Levanta has expanded from Amazon affiliate marketing into a platform spanning Amazon, Walmart, Shopify and other retail marketplaces, and states it gives brands access to more than 90,000 vetted creators plus millions more through AI-powered Social Discovery. Recent platform additions include automated product sampling, flat-rate collaborations carrying affiliate-level measurement, and integrations with ChatGPT and Claude. Levanta plans to use the capital to bring dozens of additional retail marketplaces onto the platform and to expand its go-to-market organization. Larry Cheng, founding partner at Volition Capital, stated that brands and marketplaces increasingly treat creators and affiliates as measurable drivers of demand beyond owned channels. Read the full press release (Published: August 27, 2026 | Source: GlobeNewswire)
Advisar Lands $750K to Bring Clarity to Retail Media Spend Advisar, an independent AI-enabled decisioning layer for retail suppliers, announced on August 27, 2026 that it raised a $750,000 pre-seed round led by Clutch VC, an Austin-based seed-focused venture firm. Participants include Venture Center Arkansas Fund, The Validation Fund at Winrock International, Rekall Capital and White River Capital Advisors, alongside angel investors Michael Paladino, co-founder of RevUnit, and Jeff Amerine, general partner at Cadron Capital. Advisar operates from Bentonville, Arkansas, and targets the measurement gap suppliers face when funding retail media programs without independent visibility into what the spend returns. The release discloses no customer count, product availability date or performance results. Read the full press release (Published: August 27, 2026 | Source: Business Wire)
Owner Raises $240M Led by Goldman Sachs Alternatives to Build the AI-Native Platform for Every Local Business Owner, an AI-native platform for local businesses starting with restaurants, announced on August 28, 2026 that it raised $240 million at a $2.3 billion valuation in a round led by Growth Equity at Goldman Sachs Alternatives, with existing investors Meritech, Redpoint, Headline and Jack Altman participating. Owner describes itself as the AI CMO and CTO for local businesses, building and running customer websites, online ordering, mobile apps, customer relationship management, customer support, point of sale and AI phone ordering, with AI agents managing and improving each component automatically. The company stated it launched in 2020 and has surpassed $100 million in annual recurring revenue, that independent restaurant owners will drive more than $1 billion in sales through the platform during 2026, and that thousands of local businesses and more than 100 million American consumers use it. Growth Equity at Goldman Sachs Alternatives has invested more than $17 billion since 2003, and Goldman Sachs reported approximately $4.0 trillion in assets under supervision as of June 30, 2026. Read the full press release (Published: August 28, 2026 | Source: PR Newswire)
Walmart Sparky Optimisation: What Brands Should Do to Get Recommended as Sparky Users Spend 40% More Azoma, the agentic commerce optimisation platform serving AI shopping agents including Walmart Sparky, Amazon Rufus, ChatGPT and Gemini, published guidance on August 28, 2026 for consumer brands and retail media agencies seeking to improve how Sparky surfaces their products. Azoma anchored the guidance to figures Walmart Chief Executive Officer John Furner gave on the company’s Q2 earnings call, where Furner stated that the number of customers using Sparky rose 70% year over year and that customers who use Sparky for shopping spend 40% more per order than those who do not, with Walmart worldwide e-commerce up 23% in the quarter. From its own analysis of millions of shopping agent citations during Q2 2026, Azoma reported that Sparky’s cited sources break down as 36% earned media, 30% brand-owned sites, 27% retailer content, 5% user-generated content and 2% affiliate sites, and stated that Sparky carries no paid placement a brand can buy onto. Azoma lists L’Oréal, Unilever, Mars, Beiersdorf and Reckitt among its clients. The release discloses no query set, category coverage or citation attribution method. Read the full press release (Published: August 28, 2026 | Source: GlobeNewswire)
Frequently asked questions
What marketing technology and AI news was announced on August 27 and 28, 2026?
Seven wire releases carried the marketing technology and AI news for August 27 and 28, 2026. On August 27, The Trade Desk introduced Kokai Zuma with agentic AI and a simplified measurement framework, NIQ released a cut of A Tale of Two Consumers reporting that nearly three-quarters of shoppers use AI for product discovery, ActiveCampaign launched Active Intelligence: Wavelength, Levanta raised a $22 million Series B, and Advisar raised a $750,000 pre-seed for retail media decisioning. On August 28, Owner raised $240 million at a $2.3 billion valuation, and Azoma published Walmart Sparky optimisation guidance built on Walmart’s Q2 earnings figures.
What did The Trade Desk announce with Kokai Zuma on August 27, 2026?
The Trade Desk (NASDAQ: TTD) introduced Kokai Zuma on August 27, 2026, the newest release of the Kokai platform it launched in 2023. Zuma adds a conversational Koa Assistant in open beta plus specialized agents for audience creation, optimization, troubleshooting and campaign insight, and rebuilds the platform’s forecasting engine to predict inventory and model campaign outcomes. The Trade Desk reported an average 32% improvement in cost-per-acquisition performance in initial results, without disclosing sample size, test window or control condition.
What did NIQ’s August 27, 2026 report find about AI and product discovery?
NIQ released a cut of A Tale of Two Consumers: The Polarized Mindsets Reshaping Global Consumption on August 27, 2026, produced with World Data Lab, reporting that nearly three-quarters of shoppers now use AI for product discovery and that retail media has grown into a $184 billion global market. NIQ also reported that nearly one-third of Western consumers have bought a product after discovering it on a social platform, that 68% of North American consumers have never purchased through social media, and that nearly 60% of Asia-Pacific consumers shop through social and quick-commerce channels. The release discloses no sample size, market list or field dates.
What did Azoma report about Walmart Sparky on August 28, 2026?
Azoma reported on August 28, 2026 that Walmart Chief Executive Officer John Furner stated on the company’s Q2 earnings call that customers using Walmart’s Sparky assistant spend 40% more per order than those who do not, and that Sparky user numbers rose 70% year over year. From its own analysis of millions of shopping agent citations in Q2 2026, Azoma found Sparky’s cited sources break down as 36% earned media, 30% brand-owned sites, 27% retailer content, 5% user-generated content and 2% affiliate sites. Azoma states that Sparky carries no paid placement a brand can buy onto.
What did ActiveCampaign launch on August 27, 2026?
ActiveCampaign launched Active Intelligence: Wavelength on August 27, 2026, the first named release of its Active Intelligence layer. Wavelength uses more than 500 signals from each customer’s own account history and performance patterns to draft campaigns, build segments and flag broken automation journeys. ActiveCampaign stated that 84% of marketers still run generic campaigns built on industry best practices rather than their own data, and reported a 75% higher engagement rate among businesses using Active Intelligence, a comparison of adopters against non-adopters with no disclosed baseline or matching method.
How should marketing leaders respond to AI shopping surfaces that carry no paid placement?
Assign an owner for agent-surface visibility before Q4 2026 planning closes, because Azoma’s Q2 2026 analysis puts 36% of Walmart Sparky citations and 41% of ChatGPT citations in earned media, which sits with communications rather than media buying in most organizations. Instrument Sparky and Rufus referrals separately from general AI traffic so your own order data can test Walmart’s 40% basket premium. Audit which cited source types you control, since retailer content covers 27% of Sparky citations and 37% of ChatGPT citations, leaving a brand that optimizes only its retail listings addressing roughly a third of the surface.
