AI in Marketing (2026): How Marketers Use It, What Works and What the Evidence Shows
AI is now routine in marketing, but the evidence on results is mixed and the landscape is shifting rapidly. According to HubSpot’s 2026 State of Marketing survey, 86.4% of marketing teams use AI in at least a few areas. Salesforce reports that 75% of marketing organizations deploy some form of AI, while McKinsey finds that 88% of respondents use AI in at least one function. Despite this widespread adoption, search and advertising are being rebuilt around AI, changing how content is discovered and purchased. This guide looks at how teams actually use AI marketing tools today, what controlled experiments show about results, the leading tools by category with prices, how AI search and AI ads are changing the channels, the rules that now apply, and what it all means for marketing jobs. We also ran our own test of how an AI assistant recommends marketing tools, which matters as more buyers ask a chatbot before they search.
How marketers use AI today
Adoption has moved from experimentation to integration, though governance and training lag behind. HubSpot’s 2026 State of Marketing survey of 1,500+ global B2B and B2C marketers found that 86.4% of teams use AI in at least a few marketing areas. The most common uses are content creation (42.5% extensively, 38% occasionally), media creation (37.2% / 37.7%), ad automation (34.1% / 36.5%), and admin task automation (35.6% / 40.5%). 68.2% say they understand how to use AI in marketing (up from 47% a year earlier) and 67.5% know how to measure its impact, but the workload is rising: 83.5% say they are expected to produce more content because of AI. Time savings are reported, with about one-third of teams saving 10–14 hours per week and another third saving 15+ hours.
Salesforce’s 10th edition State of Marketing, surveying 4,450 marketers across 26 countries, found that 75% of organizations use at least one form of AI, with 13% using agentic AI. High performers are nearly twice as likely as underperformers to use agents. Despite this, 61% say full AI integration into their martech is still a work in progress, and 48% haven’t figured out how to adapt their strategy to broad AI use. Reported gains from AI deployment include a 20% increase in marketing ROI and a 19% increase in conversion rates, though these are self-reported figures.
McKinsey’s State of AI survey points the same way across the whole economy: in its 2025 edition 88% of respondents said their organization uses AI regularly in at least one function, marketing and sales was among the functions where revenue gains were most often reported, but only about 6% of organizations counted as “AI high performers” earning 5% or more of EBIT from AI.
Gartner’s 2025 CMO Spend Survey of 402 marketing leaders found that marketing budgets have flatlined at 7.7% of company revenue. GenAI benefits cited include time efficiency (49%) and cost efficiency (40%). Notably, 22% say GenAI has let them reduce reliance on external agencies, and 39% plan to cut agency budgets. Gartner also forecasts that traditional search engine volume will drop 25% by 2026 due to AI chatbots.
The Marketing AI Institute’s 2025 State of Marketing AI Report, based on 1,882 respondents, found that 74% say AI is critically or very important to their marketing in the next 12 months. 60% are piloting or scaling AI. However, 68% effectively have no AI training, and 75% lack an AI roadmap.
BCG’s survey of 200+ CMOs found that 83% are optimistic about GenAI, and 71% plan to invest more than $10M/year in it. Leaders using scaled personalized offers see 3x higher returns than with mass offers.
Content Marketing Institute’s 2026 B2B Content and Marketing Trends survey of 1,015 B2B marketers found that 95% use AI-powered applications. For content creation, productivity improved for 87% and operational efficiency for 80%. However, only 39% saw improved content performance.
Does AI marketing actually work?
Rigorous evidence on AI in marketing shows mixed but often positive results, particularly in ad performance, though consumer perception varies. A study by Lee, Todri, Adamopoulos, and Ghose from NYU Stern and Emory University, published in November 2025, analyzed fully AI-generated ads on the Google Display Network. The study found that fully AI-generated ads raised click-through rates (CTR) by up to 19% compared to human-expert ads. However, disclosing that an ad used generative AI cut CTR by 31.5%. Human ads modified with AI showed no significant gain.
Daviet and Nishimura, publishing in Marketing Science in 2026, used generative AI with Bayesian neural nets to screen for performance and brand fit in a live campaign for an outdoor-activities company. In a live campaign the AI visuals achieved a higher average click-through rate, with less variation, than both a professional designer’s work and an AI tuned only for aesthetics.
Dubé and Xu from Chicago Booth conducted a randomized controlled trial with Wine Access, involving 27,500 newsletter subscribers. They found that all email arms beat the control and were generally indistinguishable in profit. In a trial using Claude, AI beat human writers on gross profit, projecting about $5.5M vs $5M annually. Full automation would cut the $375k/year writer cost by 83–90%.
Fang, Yuan, Zhang, Donati, and Sarvary conducted field experiments in online retail, testing seven workflows including advertising. Sales effects ranged from 0% to +16.3%, with about $5 incremental per consumer per year. Gains came from conversion rather than basket size, and returns and ratings did not worsen.
A Taboola-funded study with researchers from Columbia, Harvard, TUM, and CMU analyzed over 500 million impressions. Raw CTR was 0.76% for AI vs 0.65% for human. Under tight controls, they performed comparably. AI ads that did not look like AI performed best, with large human faces serving as a key trust cue.
Consumer reaction remains a critical factor. NielsenIQ’s 2024 study found that consumers identified most AI ads and found them less engaging, carrying a negative brand halo. High-quality AI ads produced weaker memory activation than traditional ads. The IAB’s 2026 “AI Ad Gap Widens” report surveyed 505 US Gen Z/Millennial consumers and 104 ad executives. 71% of consumers believe they have seen an AI ad, up from 54% in 2024. While 82% of executives think consumers feel positive about AI ads, only 45% of consumers agree. 39% of Gen Z feel negative about AI ads. 73% say knowing an ad used AI would increase purchase likelihood or make no difference, but fewer than half of advertisers always disclose.
The ad platforms’ own numbers look even better, but they are the platforms’ claims, not independent results. Meta’s Q3 2025 earnings call noted that advertisers running Advantage+ lead campaigns see a 14% lower cost per lead on average. In April 2025 Meta also claimed its Advantage+ sales campaigns returned $4.52 per $1 spent, 22% more than its usual campaigns. Google’s Performance Max data shows 27% more conversions or value at similar CPA/ROAS for non-retail advertisers. Google’s AI Max for Search typically delivers 14% more conversions/value. An independent test of 250+ campaigns found a median revenue increase of 13% but a median CPA increase of 16%. OpenAI reported that ChatGPT Ads reached a $1B annualized run-rate in under 200 days, with one e-commerce advertiser achieving 3x ROAS over 28 days.
Brand programs mostly report savings in production time and cost, again by the companies’ own accounts. Klarna reported that AI accounted for 37% of sales and marketing cost savings, about $10M annualized, with external agency spend falling 25%. Unilever used digital twins to make imagery 2x faster and 50% cheaper, with images holding attention 3x longer. L’Oréal says its CREAITECH lab, built on Google’s Imagen, Gemini and Veo models, cut concept creation from weeks to days, while the company does not use AI-generated people or life-like faces, bodies, hair or skin in its campaigns. Mondelez invested $40M+ in a gen-AI content tool, projecting 30–50% lower content production costs.
Consumer backlash has occurred. Coca-Cola’s AI-made holiday ads drew “slop” criticism online in both 2024 and 2025, although the company said the 2025 film was cheaper and faster to make. McDonald’s Netherlands pulled its AI Christmas ad after three days of social media backlash.
AI marketing tools by category
AI tools for marketing have matured into specialized categories, each addressing specific workflow needs. General assistants like ChatGPT, Claude, and Gemini are used for copy, research, and image generation. ChatGPT has introduced ads on its Free and Go tiers, while Claude and Gemini offer varying levels of access and integration. These tools are foundational for brainstorming and drafting.
For marketing copy, tools like Jasper and Copy.ai specialize in brand voice consistency and GTM workflows. Jasper includes GEO agents to optimize for generative engines, while Copy.ai offers structured workflows for sales and marketing content. These tools help teams produce more content at scale, addressing the workload increases noted in adoption surveys.
Design tools such as Adobe Firefly and Canva’s Magic Studio generate images and videos. Firefly offers commercially safe outputs with Content Credentials, while Canva integrates AI directly into its design platform. These tools reduce the time and cost of visual asset creation, as seen in Klarna and Unilever’s reports.
Video generation is dominated by Runway, Synthesia, HeyGen, and Google’s Veo. Runway offers high-fidelity video models, while Synthesia and HeyGen specialize in AI avatars and dubbing. These tools are increasingly used for personalized video ads and training content, though human faces remain a key trust cue.
For SEO and AI visibility, Semrush and Ahrefs have added AI toolkits. Semrush includes an AI Search toolkit, while Ahrefs offers Brand Radar AI to monitor AI-generated search results. HubSpot’s AEO beta helps brands optimize for answer engines. These tools are critical as search traffic shifts to AI summaries.
CRM and agent platforms like HubSpot Breeze and Salesforce Agentforce automate customer interactions. HubSpot’s Breeze agents handle content, prospecting, and customer data, while Salesforce’s Agentforce provides autonomous agents across sales and service. These platforms are central to the shift toward agentic AI.
Ad automation tools include Google’s Performance Max and AI Max, Meta’s Advantage+, Amazon’s Creative Agent and OpenAI’s ChatGPT Ads. These platforms automate creative generation, audience targeting, and bidding. They represent the shift toward end-to-end AI advertising, where marketers supply assets and budget, and AI handles the rest.
| Category | Tools | What marketers use them for | Price from (official page, 24 Oct 2026) |
|---|---|---|---|
| General assistants | ChatGPT, Claude, Gemini | Drafting, research, images, agents | ChatGPT Go $8/mo (US); Claude Pro $20/mo |
| Marketing copy | Jasper, Copy.ai | Brand-voice copy, campaign workflows, GEO agents | Jasper Pro $69/seat/mo; Copy.ai Chat $29/mo |
| Images and design | Adobe Firefly, Canva Magic Studio | Ad visuals, social graphics, product imagery | Firefly Standard $9.99/mo |
| Video | Runway, Synthesia, HeyGen, Google Veo | Short ads, avatar explainers, dubbing | Runway Standard $15/mo; Synthesia Starter $29/mo; HeyGen Creator $29/mo |
| SEO and AI visibility | Semrush, Ahrefs (Brand Radar AI), HubSpot AEO (beta) | Keyword research, tracking brand mentions in AI answers | Semrush from €134/mo; Ahrefs from €27/mo; Brand Radar AI from €179/mo |
| CRM and agents | HubSpot Breeze, Salesforce Agentforce | Email, lead handling, customer and prospecting agents | HubSpot Marketing Hub Starter from $7/seat/mo; Agentforce $2/conversation |
| Ad automation | Google Performance Max and AI Max, Meta Advantage+, Amazon Creative Agent, ChatGPT Ads | Automated targeting, bidding and creative | No fee beyond media spend |
Monthly prices are list prices billed monthly where both options exist; Semrush and Ahrefs showed euro prices from our location. Canva and most ChatGPT plan prices could not be read from their pages.
Search is changing: AI Overviews, AI Mode and GEO
The landscape of search is undergoing a structural shift that is already impacting how marketers acquire traffic. According to a July 2025 Pew Research study, 18% of 68,879 tracked searches produced an AI summary. When those summaries appeared, click-through rates (CTR) on traditional organic results dropped to 8% from 15% when no summary was present. Furthermore, 58% of respondents encountered at least one AI summary during their browsing sessions. The effect is even more pronounced for longer queries: 53% of searches containing ten or more words triggered a summary, compared to just 8% of one- or two-word searches.
Data from Ahrefs reinforces this trend. In an April 2025 analysis of 300,000 keywords, the presence of an AI Overview correlated with a 34.5% lower CTR for the top position. Its February 2026 update found the click-through rate of the top result on AI Overview keywords fell from 7.3% in December 2023 to 1.6% in December 2025, a 58% drop after allowing for the trend; the finding is a correlation, not proof of cause. Seer Interactive’s September 2025 update showed organic CTR on AI Overview queries dropping from 1.76% to 0.61%, while paid CTR on those same queries fell from 19.7% to 6.34%. However, there is a clear upside for visibility: brands cited within an AI Overview received 35% more organic clicks and 91% more paid clicks than those that were not cited, according to Seer Interactive.
Traffic from generative AI tools is also growing rapidly. Adobe reported that traffic to US retail sites from generative AI tools surged 693.4% year-over-year during the November-December 2025 holiday season. That is why marketers are turning to generative engine optimization (GEO): making content easy for AI assistants to find, quote and cite. The foundational KDD 2024 paper on GEO demonstrated that adding citations, quotations, and statistics can boost visibility in generative engine responses by up to 40%. Adoption is accelerating: the 2026 CMO Survey found that 40% of companies now use GEO, while Salesforce reports that 88% of marketers are optimizing for AI-driven search. To capture this traffic, marketers must structure content to be easily cited by AI models, with clear facts, statistics and sources rather than keyword repetition. Our test below shows how unstable AI recommendations still are.
Ads inside AI: ChatGPT ads and fully automated campaigns
The integration of advertising into AI interfaces is moving from experimentation to scale. In January 2026, OpenAI outlined its advertising approach for ChatGPT, launching ads on the Free and Go tiers while keeping Pro, Business, and Enterprise users ad-free. By August 2026, ChatGPT Ads had reached a $1 billion annualized run-rate with tens of thousands of advertisers across 40+ countries. OpenAI’s Ads Manager allows for self-serve campaigns using CPC and outcome bidding, with a pixel and Conversions API available for measurement.
Meanwhile, major ad platforms are automating the creative and targeting process. Meta plans to fully automate ad creation and targeting by the end of 2026. According to a June 2025 Wall Street Journal report, advertisers will soon supply only a product image and budget, with AI generating the image, video, and text, selecting audiences, and personalizing creatives in real time. Meta already offers end-to-end Advantage+ automation for sales, app, and lead campaigns by default. In Q3 2025, Meta reported that advertisers running Advantage+ lead campaigns saw a 14% lower cost per lead on average, and the annual run-rate for its end-to-end AI-powered ad tools exceeded $60 billion.
Google launched AI Max for Search in May 2025 and took it out of beta in April 2026. Google claims it typically delivers 14% more conversions at a similar cost per acquisition, and 27% for campaigns that mostly use exact and phrase match; an independent test of more than 250 campaigns found median revenue up 13% but median cost per acquisition up 16%. Dynamic Search Ads are scheduled to migrate to AI Max in September 2026. Amazon introduced its Creative Agent at the unBoxed 2025 event, an agentic tool that produces display, video, and audio assets at no cost.
This shift offers efficiency but reduces control. Marketers gain speed and scale but lose granular oversight of where and how their creative appears. The NYU Stern and Emory research found that disclosing an ad as AI-generated cut CTR by 31.5%, suggesting that transparency can impact performance. The Taboola-funded study with Columbia and Harvard researchers points the same way: AI ads that did not look like AI performed best, and a large human face was a key trust cue.
Our test: which marketing tools does an AI assistant recommend?
More and more buyers ask a chatbot which tool to use before they ever reach a search results page, which is why “AI visibility” has become a marketing goal. We wanted to see how stable those recommendations are. On 24 October 2026 we asked an open-weight model, Qwen3.6-35B-A3B running on our own server, for the best AI tools for 12 common marketing jobs, from copywriting and SEO to paid ads and personalization. Each question came in two versions, a plain one (“What are the best AI tools for SEO and keyword research? List your top 5”) and one from a small business (“I run a small business with a marketing budget of about $100 a month…”), and each was asked five times: 120 answers in all.
What we found:
The answers are a long tail, not a shortlist. The model named 140 different tools. The ten most recommended accounted for only 26% of all mentions. Canva led (28 mentions), followed by Semrush (21), Jasper (16), Ubersuggest and HubSpot (14 each).
The same question gives different answers. Asked the same question five times, the model named about nine different tools per question, and on average fewer than two appeared in all five answers. For paid-ads tools, no tool appeared in every run of either version of the question.
Context changes everything. Adding “small business, about $100 a month” did not just reorder the list, it replaced it. For SEO, design, CRM and paid ads, none of the tools the model recommended consistently for the plain question survived; for SEO, Semrush, Ahrefs and Surfer SEO gave way to Ubersuggest, Google Keyword Planner and AnswerThePublic. Copywriting was the most stable category, with 60% overlap.
Assistants recommend things that no longer exist. In 5 of the 10 personalization answers, the model recommended Google Optimize, an A/B-testing tool Google shut down in September 2023. An assistant answering from its training data can send buyers to products that are gone.
What this means for marketers: being “recommended by AI” is not a fixed position you win once. Brands the model reaches for across many tasks, such as Canva (named in design, video, social and analytics answers) and Semrush (SEO and research), get the most exposure; most other tools appear in some answers and not others. Visibility also depends on how the buyer phrases the question, including their size and budget, which is another reason to publish clear, specific information about who your product is for and what it costs.
Limits: one open-weight model, which answers from its training data without searching the web; commercial assistants that browse may behave differently. Twelve tasks, 120 answers, tool names read from the answer lists (8 answers could not be parsed). We measured consistency, not which tools are actually best.
Rules and risks
The regulatory environment for AI in marketing is tightening. In the US, the FTC’s Consumer Reviews and Testimonials Rule, effective October 2024, bans fake reviews, including those generated by AI. Knowing violations can bring civil penalties of up to $53,088 each. The FTC has already sent warning letters to ten companies regarding this rule. Its Operation AI Comply (September 2024) went after deceptive AI claims, including DoNotPay’s “robot lawyer” and AI “passive income” schemes, and 2025 cases targeted “AI-washing”, such as unsupported accuracy claims for an AI-content detector.
In the EU, Article 50 of the AI Act has applied since 2 August 2026. It requires providers to inform users when they are interacting with an AI system. Generative AI providers must mark outputs with machine-readable data, and deployers must disclose deepfakes. A Code of Practice on marking was finalized in June 2026.
Copyright remains a complex area. The US Copyright Office concluded in January 2025 that prompts alone do not constitute authorship, meaning purely AI-generated ad assets may not be protectable. Adobe offers IP indemnification for select Firefly outputs to eligible enterprise customers to mitigate this risk.
Liability for AI hallucinations is also being tested. In the 2024 case Moffatt v. Air Canada, the airline was held liable for negligent misrepresentation by its chatbot regarding bereavement fares, paying C$650.88 in damages. Brand safety is another concern: NielsenIQ found that AI ads often carry a negative brand halo, and Coca-Cola’s 2024 and 2025 AI holiday ads faced significant public criticism. Disclosure practices are inconsistent; the IAB found that while 89% of advertisers using generative AI disclose it at least sometimes, fewer than half always do.
What AI means for marketing jobs and agencies
AI is reshaping the marketing workforce and agency economics. The Marketing AI Institute’s 2025 report found that 53% of marketers believe AI will eliminate more marketing jobs than it creates, up from 40% in 2023. Gartner’s 2025 CMO Spend Survey indicates that 39% of marketers plan to cut labor spend, and 39% plan to reduce agency budgets. WPP, for instance, reported a headcount reduction from 104,083 to 97,388 by mid-2026, with up to 1,000 more cuts planned, tied to both AI adoption and weaker client spend.
Despite these cuts, productivity gains are real. HubSpot reports that about one-third of teams save 10-14 hours per week, and another third save 15+ hours. Salesforce estimates that AI agents will return eight hours per week to marketers. However, workload is increasing: 83.5% of marketers expect to produce more content because of AI. The work is shifting from execution to strategy and brand management. In a Chicago Booth study, Wine Access found that AI automation could cut writer costs by 83-90% while maintaining profit levels, highlighting how routine content production is becoming commoditized.
How to use AI in marketing: a practical plan
- Pick 2-3 workflows: Start with high-volume, repetitive tasks like content drafting or image generation. Avoid trying to automate everything at once.
- Set rules: Establish an AI policy covering data privacy, disclosure, and brand voice. Ensure you have a review process for AI-generated content.
- Train the team: Address the skills gap. The Marketing AI Institute found that 68% of marketers receive no AI training. Invest in prompt engineering and tool literacy.
- Measure against a control: Use holdout groups to compare AI performance against human-led campaigns. Track metrics like CTR, conversion rate, and CPA.
- Protect the brand: Never generate fake reviews, check every factual claim, and test AI creative before scaling it; in a large ad study a prominent human face was a key trust cue.
- Prepare for AI search: Implement GEO strategies by citing authoritative sources and structuring content for easy extraction by AI models.
- Keep humans on strategy: Use AI for execution, but keep humans in charge of final approval, creative direction, and brand narrative.
Frequently asked questions
How is AI used in marketing? AI is used for content creation, media production, ad automation, admin task automation, and strategic planning. According to HubSpot’s 2026 State of Marketing report, 86.4% of marketing teams use AI in at least a few areas, with content creation and ad optimization being the most common applications.
What are the best AI marketing tools? It depends on the job: general assistants such as ChatGPT and Claude for drafting and research, Adobe Firefly or Canva for visuals, Runway, Synthesia or HeyGen for video, Semrush or Ahrefs for SEO and AI visibility, HubSpot or Salesforce for CRM and agents, and the built-in automation of Google, Meta and Amazon for ads. In the Marketing AI Institute survey, ChatGPT was marketers’ favorite tool.
Does AI-generated advertising work? Yes, but with caveats. Studies show AI-generated ads can outperform human designs in CTR when they do not look obviously like AI. However, disclosing AI usage can reduce engagement, and some consumers find AI ads less engaging or annoying.
Will AI replace marketers? 53% of marketers in the Marketing AI Institute’s 2025 survey expect AI to eliminate more marketing jobs than it creates, though only 21% worry about their own role. Production work is most exposed; strategy, brand judgment and review are where people remain essential.
Do I have to disclose AI-generated ads? In the US there is no general rule requiring a label on every AI-made ad, but deceptive claims and fake reviews are illegal. In the EU, the AI Act requires deepfakes to be disclosed and AI-generated content to be machine-readably marked by the tools that make it. Studies disagree on whether labels hurt results, but most consumers say they want disclosure when AI imitates a real person.
Can I copyright AI-generated marketing content? The US Copyright Office has stated that prompts alone do not constitute authorship. Human-authored elements or significant creative selection and modification of AI output may be protected, but purely AI-generated assets may not be.
What is generative engine optimization (GEO)? GEO is the practice of optimizing content to be cited and referenced by AI search engines. It involves adding citations, quotes, and statistics to increase the likelihood that AI models will use your content as a source in their responses.
How do I start using AI in marketing? Start by identifying repetitive tasks that consume the most time. Implement AI tools for those workflows, set clear guidelines for quality and tone, and train your team. Measure results against a control group to ensure AI is delivering value.