Two in five American adults would hand at least one everyday task to an AI agent, yet fewer than one in five believe such agents are a good idea, according to The State of AI Adoption in America, a report Channel V Media first published on September 23, 2026 and updated on October 2.

In Short

A survey of 7,675 American adults, analyzed by a New York PR firm, found that 40% would let an AI helper do at least one chore for them, such as booking a restaurant, even though only 8% currently use one. That matters to anyone selling things, because when software picks the restaurant or the groceries, a shopper may see only the one option the software chose rather than a page full of brands and ads. The catch is that the firm sells services to help brands show up in AI answers, and most people still want a human for health care and banking, so the numbers describe interest more than a finished change in behavior.

What the survey measured

Channel V Media, a New York public relations and communications agency founded in 2008, built the report on a Prosper Insights & Analytics survey of 7,675 US adults aged 18 and older, fielded in June 2026. According to the report's methodology, Prosper designed and fielded the survey and collected the data, while Channel V Media selected the questions covered, carried out the analysis and is responsible for the interpretation. Results are shown for all adults and for six segments: Gen Z (1,543 respondents), Millennials (1,924), Gen X (2,039), Boomers (1,853), executives and business owners (1,960) and employees (1,471).

The company describes the document, branded The AI Barometer, as the first edition of an ongoing tracker. Written by Kieran Powell, Executive Vice President of Channel V Media, the report page is dated September 23 and was updated on October 2.

The central split is stark. According to the report, 40.1% of respondents would let an agent take on at least one everyday task, 7.7% use agentic AI at present, and 17.6% say agents are a good idea. Another 41.8% say they are not, and 40.6% are unsure. Current use therefore sits at roughly one fifth of stated willingness, a ratio Channel V Media highlights itself. Awareness is the first hurdle: 67.6% of Americans have never heard of agentic AI.

The report also arrives with a commercial frame that is not hidden. Channel V Media says it helps companies shape how AI engines describe them, a practice it calls AI visibility. Its current and past clients include ad tech companies such as Grapeshot + Oracle, Innovid and Albert Technologies, alongside IBM and Pernod Ricard. The report closes with an offer to audit how ChatGPT, Perplexity, Gemini and Claude describe a company. The findings are vendor-supplied data from a firm that sells a remedy for the problem the data describes. That does not make them wrong, but it shapes how they are read against other research.

Search first, delegation second

Search is the most common use. According to Channel V Media, 53.0% of generative AI users search with it and 48.3% use it for research and summaries, well ahead of writing assistance at 30.4% and content creation at 17.1%. Among Boomers who use generative AI, 65.5% use it to search, the highest of any generation. Gen Z, at 41.6%, is the only generation for which search is not the top use.

That Boomer figure needs a denominator, and the report supplies one. Generative AI use stands at 40.6% of all US adults, with Millennials at 47.2%, Gen Z at 44.3%, Gen X at 42.4% and Boomers at 32.2%. Measured against all adults rather than users, about 21.5% of Americans search with AI, according to the report, and the generational spread narrows: Gen Z 18.4%, Millennials 22.8%, Gen X 22.9% and Boomers 21.1%. Boomers who use AI lean on search hardest, but across the whole generation they search with it at about the national rate.

The rest of the population divides along awareness. According to the report, 30.5% of Americans have heard of generative AI but do not understand what it is, and 12.4% have not heard of it at all. Channel V Media combines those two groups, 42.9% of adults, as the audience still forming its view of the channel. A further 16.5% are excited to try generative AI but do not use it yet. Among Boomers, the aware-but-confused group (46.6%) is larger than the share that uses the technology (32.2%).

The company's argument runs in two stages: AI first becomes the place where consumers form an impression of a brand, then becomes the thing that acts on that impression. The first stage is measured here. The second is mostly intent.

Restaurants, groceries and the smart home

Which tasks would people give away? Booking a restaurant leads at 12.1%, followed by buying groceries at 11.0%, managing a smart home at 10.0% and paying bills at 9.8%, according to Channel V Media. Willingness falls as stakes rise. Investing in stocks, bonds or crypto draws 8.1%, healthcare matters 6.9%, teaching children 5.6% and buying a car 5.0%.

A technical point sits inside those figures. The 40.1% headline is an aggregate, derived as 100% minus the 59.9% who chose none of the 13 tasks offered, and no single task drew more than about one respondent in eight. According to the report's methodology, respondents were first asked whether they had heard of agentic AI, then shown a definition - "AI systems designed to act autonomously, making decisions and taking actions to achieve specific goals with minimal human intervention" - and then asked whether it is a good idea and which tasks they would use it for, in a check-all-that-apply question.

Restaurant booking is also the category where agents already work. Google's AI Mode began taking restaurant reservations for US users in November 2025, and the feature reached the United Kingdom on April 10, 2026 with eight partner platforms including TheFork and OpenTable. In August, ChatGPT gained Yelp table booking inside the conversation, and Google added agentic food ordering to Maps through Square and Toast on August 6. The appetite Channel V Media measured for restaurant booking is appetite for something consumers can already do; that is not true of most of the list.

Grocery delegation looks different depending on who asks and how. Criteo research published in late September found 32% of 6,307 respondents across six markets willing to let an AI assistant reorder groceries and staples automatically when supplies run low, and 40% for personal care. The distance between that 32% and Channel V Media's 11.0% suggests framing - automatic replenishment of known items against open-ended delegation - matters as much as underlying attitude. A BCG and Moloco survey covered in January put the share of consumers comfortable letting AI make purchases on their behalf at 45%.

Doubt that does not stop use

Skepticism runs alongside adoption rather than blocking it. The two leading concerns are almost tied, according to the report: 40.4% of Americans say AI needs human oversight and 40.3% worry it gives wrong information, ahead of job losses at 36.2%. Some 29.3% do not trust that AI has their best interests in mind, and only 16.4% report no concerns at all.

The gap between attitude and willingness is widest among Boomers. Just 4.7% call agents a good idea and 54.7% say they are not, yet 18.6% chose at least one task they would hand to an agent, nearly four times as many as approve. Boomer use of agentic AI today stands at 1.6%.

Other research reaches the same shape with different instruments. Yelp and Morning Consult found in April that only 15% of 2,202 US adults trust AI search results a lot, although 65% had used an AI search tool in the preceding six months. Shift Browser's March 2026 survey of 1,448 Americans recorded 81% concerned about AI accessing personal data even as daily use climbed.

Executives ahead of their own staff

The workplace cut is among the report's more distinctive contributions. Executives and business owners use generative AI more than employees, 53.2% against 37.9%, and agentic AI at 17.5% against 6.6%, according to Channel V Media - 2.7 times the employee rate. A third of executives (33.8%) call agentic AI a good idea, against 18.9% of employees, and 48.6% have heard of it, against 35.2%.

Adoption at work does not extend to personal money. According to the report, 76.8% of executives prefer a live person for banking and 77.4% for healthcare. The shares preferring AI chat are close to those of employees: 23.2% against 23.4% for banking, and 22.6% against 20.7% for healthcare. The two groups also worry about different things. Executives are more concerned about the quality of AI-generated responses, 40% against 34%, while employees are slightly more likely to doubt that AI has their interests in mind, 25% against 22%, according to the release.

Both workplace groups sit below the 29.3% all-adult figure on that last measure. The report's generational table shows where distrust concentrates: 44.1% of Boomers do not trust that AI has their best interests in mind, against 28.2% of Gen X, 25.2% of Gen Z and 18.7% of Millennials. The report notes that its generational and workplace breakdowns were supplied by Prosper as separate cross-tabulations and cannot be combined.

Millennials lead, Gen Z hesitates

The generational data cuts against a common assumption about who adopts first. According to Channel V Media, Millennials lead on every adoption metric, at 47.2% for generative AI, and the lead widens on agentic AI, at 13.3% against 8.5% for Gen Z. Some 29.6% of Millennials call agents a good idea, the highest share of any generation, against 22.6% of Gen Z.

Gen Z looks more conflicted. Some 22.2% are excited to try generative AI but have not, the highest of any generation, while 21.9% say AI makes them anxious and 14.8% fear losing their own job to it. The fear runs the other way with age: 47.0% of Boomers expect AI to cause job losses in general, against 33.9% of Gen Z, yet only 4.2% of Boomers worry about their own job. RTB House research published on August 10 found 42% of US Millennials willing to delegate a $250 purchase to an AI agent when a seven-day return window applied, which points in the same generational direction.

Where consumers still want a person

A majority chose a live person over an AI chat program in every service category tested: healthcare at 84.0%, banking at 82.6%, travel at 77.5%, telecom and entertainment at 72.6%, and online shopping at 69.1%. Even Millennials, the most open cohort, prefer a person for healthcare (76.3%) and banking (74.3%), although 38.1% of them would take AI chat for online shopping.

Privacy is the most uniform concern in the dataset. According to Channel V Media, 60.5% of Americans are very or extremely concerned about AI violating their privacy, a figure that is exactly 60.5% among both executives and employees and ranges from 56.6% of Gen X to 65.0% of Boomers. The release's subheading presented the 60% as general worry about AI entering banking, careers and shopping, while the report ties it to a single privacy question on a four-point scale. PPC Land uses the narrower reading.

Banks write their own rules

Channel V Media points to one institutional response: six global banks publishing joint principles for AI agents in commerce. The release did not name them, and the report names four - Bank of America, Capital One, ING and NatWest. According to a NatWest Group release dated September 22, 2026, the six are NatWest Group, ASB Bank, Bank of America, Capital One, Commonwealth Bank of Australia and ING Group. Their paper, Building Trust in Agentic Commerce, covers five areas - transparency, safety, privacy and data, choice, and interoperability. The principles are voluntary, and the consortium plans a second paper on implementation.

The advertising relevance lies in the first area. According to Cryptopolitan's summary of the paper, the transparency principle extends to how agents rank options, including sponsored ones, and warns that agents may promote products or payment methods that pay their providers more even when those are not the best value. PPC Land could not open the paper's PDF directly. The banks' concern that consumers do not know whether agents act in their interests is the one Channel V Media cites, and the one its 29.3% figure measures.

Payment networks are building a parallel trust layer. Mastercard's Verifiable Intent, developed in collaboration with Google and aligned with the Agent Payments Protocol and the Universal Commerce Protocol, creates a tamper-resistant record of what a user authorized an agent to do.

Do agents really bypass advertising?

The pitch that accompanied the release argued that a consumer who asks an AI to book a restaurant sees no ads, only what the AI decided was best. The platform record points in a more complicated direction. Google formally brought shopping ads into AI Mode on February 11, 2026, when the surface had more than 75 million daily active users, and by late May paid listings sat mid-answer behind a small Sponsored tag as AI Mode passed a billion monthly users. On August 31 a test placed a sponsored product carousel inside the body of an answer. In ChatGPT, Adthena counted ads on 4.47% of US queries between March and May, while Similarweb recorded sponsored placements in 26% of responses to Free and Go users in August.

Those placements sit at the recommendation step, where a person still reads the output. Whether any paid mechanism reaches the final step of an agent-completed booking is a separate question, and the survey does not address it. The transaction layer has also proved harder than the discovery layer. Walmart disclosed that purchases completed inside ChatGPT converted three times worse than click-out links to its own site, and OpenAI wound Instant Checkout down in March 2026.

A narrower version of the pitch's claim holds up better. When software compresses a page of options into a single choice, fewer brands are seen at all, paid or unpaid, and the ones that are seen depend on what the model retrieved.

A PR agency's data, and a PR agency's conclusion

The release's central inference is that because AI draws on what has been published about a brand, including news coverage, earned media now shapes what AI does on a consumer's behalf. That conclusion is not derived from the survey, which measured consumer attitudes rather than what AI systems cite. Gretel Going, President of Channel V Media, set out the position in the release.

"Consumers are letting AI act for them before they fully trust it, so brands can't wait for that trust to arrive before they show up in AI's answers," Going said. "The brands that earn a role in bigger decisions like banking and healthcare will be the ones AI engines can describe credibly, and that credibility comes from what trusted sources, like media, have published about them."

The report's author is more direct about where the commercial opening lies. Powell wrote that "the four in ten Americans who are unsure whether agentic AI is a good idea are the group that matters most for adoption," and that beyond early adopters, "the rest need to be convinced, and that is a communications job."

The agency has made the argument before. Its April 2026 report with Dynata, based on 250 US marketing executives, positioned earned media as the channel most likely to feed AI-generated answers. The independent evidence is less settled. A Meltwater analysis covering August 2026 found YouTube the most-cited platform across eight AI systems, displacing Reddit - neither of them a news publisher. A GetCited study of 106,758 AI citations found 86% of best-product answers citing a commercial source, and was itself produced by an agency that sells article placements. A critical survey of 45 generative engine optimization studies concluded that no reviewed technique produced a stable cross-platform effect, and that body-only rewrites could cut a page's top-ten presence by 16%.

The report itself is more transparent than the release about how its numbers were built. It publishes the wording of each survey question, the sample size of every segment and the derivation of its aggregate figures, and states that multipliers and point gaps are calculated from the percentages shown. It does not state a margin of error. Prosper's consumer work for the National Retail Federation, covered by PPC Land in July 2025, carried a stated margin of 1.1 percentage points on a sample of 7,581. Channel V Media told PPC Land on October 2 that the same margin applies to the 7,675-adult June 2026 survey.

That margin covers the full sample. The six segments, each between 1,471 and 2,039 respondents, carry wider margins of roughly 2.2 to 2.6 points at the conventional 95% confidence level, on PPC Land's arithmetic. Gaps of a few points between segments, such as the three-point difference between executives and employees on distrust, sit close to that range.

Why this matters for the marketing community

For media buyers, the most telling number in the report is not the 40.1%. It is the 7.7%, together with the fact that no single task drew more than 12.1%. Budgets follow behavior rather than stated intent, and delegated purchasing remains a minority activity concentrated in low-stakes errands that platforms have already built into their products.

The discovery layer is a different matter. With about 21.5% of all US adults already searching through generative AI, the answer box is where brand consideration begins for a meaningful audience, and Google and OpenAI are selling inventory inside those answers at rising density. What lags is measurement. An IAB framework published on August 3 found that only 16% of brands track AI visibility systematically, with more than 20 vendors applying methods that return different answers for the same brand.

The banks' principles add a question advertisers have not yet had to price. If disclosure of sponsored ranking inside agents became a condition that issuers attach to agent payments, the commercial terms of paid placement in agent-led commerce would be set partly by the institutions that move the money, not only by the platforms that sell the ads. Who writes those terms first?

Correction

This article was updated on October 2, 2026. The original version was written from Channel V Media's press release and pitch email. The pitch linked to the full report, which PPC Land has since reviewed in its October 2 form. The 30.5% figure, originally described as Americans who understand generative AI but have not used it, refers to those who have heard of it but do not understand it. The article now includes the report's 40.6% generative AI usage rate for all adults and its estimate that about 21.5% of Americans search with AI. It also states that respondents chose from 13 tasks and were shown a definition of agentic AI, which the report discloses. The 7% and 8% figures for current agentic AI use in the pitch and release both derive from the report's 7.7%. The 2.7x executive multiple is calculated from published figures of 17.5% and 6.6%. Employee figures for banking and healthcare preferences have been added. The report names four of the six banks, and Channel V Media says a 1.1-point margin of error applies to the full sample. The report's publication date is September 23, 2026, not October 8. Figures throughout now use the report's one-decimal values.

Timeline

Summary

Who: Channel V Media, a New York public relations and communications agency founded in 2008 that sells AI visibility services, with survey design and fieldwork by Prosper Insights & Analytics. Channel V Media selected the questions and wrote the analysis; the report is written by Kieran Powell, Executive Vice President, and Gretel Going, President, is quoted in the release. The findings concern US consumers, executives and employees, and bear on advertisers, agencies, publishers and platforms selling inventory inside AI answers.

What: The State of AI Adoption in America, a survey-based report finding that 40% of US adults would let an AI agent handle at least one everyday task while 8% use agentic AI and 18% consider agents a good idea. Restaurant booking (12%), grocery buying (11%) and smart home management (10%) lead the task list; 53% of generative AI users search with it; majorities prefer a person for healthcare (84%) and banking (83%); and 60% are very or extremely concerned about AI and privacy. The release publishes no margin of error or question wording, and the accompanying pitch gives a 7% usage figure against the release's 8%.

When: The survey was fielded in June 2026, the release was sent to PPC Land under embargo on September 30, 2026, and the report was published on October 8, 2026.

Where: The United States, across a sample of 7,675 adults aged 18 and older segmented by generation and workplace role.

Why: AI assistants are becoming the point where consumers first meet brands and, for a small but measurable group, the tool that acts on their behalf. That shifts attention to who controls the recommendation layer, how paid placement inside it is disclosed, and how visibility there is measured - questions the survey raises but, as research from a firm selling AI visibility services, cannot settle alone.