The state of agentic advertising

Agentic AI, in the context of advertising specifically, is the use of autonomous and semi-autonomous AI agents to plan, transact and optimize media.

While Chief Innovation Officer at People Inc. (formerly Dotdash Meredith) Jon Roberts recently told Digiday, “there is a trust gap to be solved before this really gets deployed,” agentic AI’s overall promise — saving time so creatives and media planners can use more of their time for thinking than for manual, tedious tasks — is likely to push advertisers, agencies and publishers toward further adoption. 

In this State of the Industry report, Digiday and Optable surveyed 180 agencies, publishers, brands and retailers about how they engage with agentic advertising and AI-driven media execution — from investment and readiness to barriers and impact.

Most respondents (publishers, agencies and advertisers) noted that their respective industries are mainly in the early stages of using agentic advertising, with 31% selecting “experimenting but fragmented” and 42% choosing “early but moving quickly.” 

Eighteen percent of respondents said the industry is ready to start scaling, but the majority are still skeptical of that preparedness. While agentic advertising has potential and is part of the AI wave, the ad industry is still figuring it out. 

In terms of the industry’s overall investment in agentic AI, it’s primarily media buying optimization (70%) and measurement and reporting (58%). 

Respondents noted less of a focus on investing in agentic AI for governance (56%) and human-in-the-loop controls (45%) across their industries.

Interestingly, respondents also wanted more governance for agentic AI, counter to what they want for the industry as a whole — more on that in the following sections. 

This could be why the technology is still in its early stages, given that the needs for agentic advertising are at odds with what the industry at large wants.

Publishers are fairly evenly split among three stages of agentic AI application. While 42% of publishers are either not prioritizing agentic advertising investments or have no long-term plans in the space, 29% are either in early planning or piloting stages, with another 29% actively deploying the technology.

Among those currently deploying AI agents is The Weather Company.

“Building on decades of AI and modeling expertise, we’ve fully embraced the agentic era by deploying AI agents across our core operational pillars,” said Benny Pang, product leader, audience data at The Weather Company. “This evolution allows us to serve customers in new, beneficial ways — a shift that is particularly evident within advertising.

“Here, we utilize agentic workflows for audience discovery, identifying weather triggers that link real-time conditions to consumer behaviors and even creative design,” he added. “These agents act as a force multiplier: they handle the heavy lifting of discovery while our team provides the final verification to ensure every output meets our standards for accuracy and brand safety.” 

However, there are barriers that prevent some publishers from actively deploying agentic AI, primarily technical complexities (68%) and internal organizational readiness (51%). These responses indicate that publishers aren’t necessarily averse to agentic AI. Rather, interoperability and tech stack issues seem to be keeping them from more widely adopting the technology. Publisher teams likely need more education to gain further internal buy-in and resources.

Because agentic AI is a more nascent application of AI, it can be more complex to implement than other AI applications. Some publishers are turning to third-party companies to help them adopt AI rather than building applications in-house. 

“The industry is currently on a broad spectrum; some publishers I’ve spoken to are deep in the trenches, while others are waiting for the landscape to firm up,” Pang said. “I see a future where ownership is hybrid, with room for multiple agents to coexist. Publishers will likely utilize specialized agents provided by technology partners for scale and specific tasks, while building proprietary in-house agents to act as coordinators. 

“This internal ‘manager’ agent is key, as it better understands the organization’s unique workflows, data and internal processes,” he added. “Interoperability through standards like MCP, AdCP and AAMP is essential here, as it allows the central coordinator to manage a fleet of sub-agents across different partners seamlessly.” 

Pang recommended taking the first step now as the base AI models are always going to evolve, but the core use cases driving the value and user experience are stable. “Integrate agents into your workflows today and adjust as the technology matures,” he said. “A thriving agent ecosystem requires collaboration across the buy and sell sides. This is key to the success of the AdCP and similar protocols.”

Other prevalent barriers to wider adoption of agentic AI are unclear standards (41%) and uncertain regulations (37%). Because agentic AI can act on behalf of the user, this raises questions around regulations. 

“Companies that are going to open up their web experiences to these kinds of agentic flows, they’re eventually going to want to know that they’re safe from legal recourse from consumers or consumer groups, based on how they receive and store these kinds of data,” Clive Henry, head of partner solutions at Adobe, said in a recent interview with Digiday.

As for what publishers want from agentic AI, most respondents are focusing their testing and planning efforts on making improvements and increasing use rather than cost savings. 

Nearly half are interested in increased demand discovery for premium or niche inventory (49%), followed by higher yield from buyer-seller matching (46%), improved transparency (46%) and direct relationships with advertisers (44%).

For publishers, agentic advertising usage is primarily focused on attracting more advertisers or finding ways to become more appealing to them. 

Many conversations during the Digiday Publishing Summit Europe conference in October 2025 focused on publishers exploring AI agents for both media and content. Mario Lamaa, managing director of data and revenue operations at Immediate Media, described how the company’s AI-powered agent was built to give its sales team instant access to information, like campaign performance and historical post-campaign analysis, to enable its reps to respond to client briefs in real time. 

“Our sales team can query [the AI agent] whenever they are in conversations to help them respond to briefs,” Lamaa said onstage at the conference. “If we were speaking to a car client, they could speak to the agent and ask it: ‘What kind of segments do we have around this? What size are they? How could we deliver a campaign over a month?’ That’s the stuff that typically we would have to ask one of our analysts….[with] a turnaround time of around 48 hours. It means that we can be responding to our clients …in a much shorter time than we ever could in the past.” 

When Digiday and Optable asked respondents which publisher capabilities they were most likely to prioritize in the coming year, only 5% of respondents said they were focused on building or enhancing AI agents.

However, publishers indicated they are focused on optimizing for external AI agents. Forty-nine percent of respondents said they are improving first-party data and signals for AI agent discovery. 

Publishers have been heavily affected by the rise of zero-click search and its impact on their ability to control content scraping. On the ad side, they are likely seeing an increase in advertisers using AI agents to optimize for digital ad placements and buys.

Advertisers are more keen to use agentic AI than publishers. While only 29% of publishers said they were actively deploying agentic AI, 89% of advertisers said they were using or building agentic AI.

“We’re seeing advertisers, mostly through their agency partners, use AI agents to automate audience discovery, media planning and campaign activation,” said Bennett Crumbling, head of marketing at Optable. “Instead of manually configuring campaigns, buyers define objectives in natural language and agents handle the discovery, matching and execution. The 41% using out-of-the-box AI agents from tech vendors tracks with what we see: platforms like Scope3 are building brand-side agents, PubMatic is layering AI into SSP workflows and Newton Research enables AI-powered audience modeling.

“On the agency side, many agency groups are building agentic tools for things like automated inventory matching and PMP creation,” he continued. “On the sell side, tools like Optable’s Audience Agent automate RFP-to-audience workflows in minutes rather than days, while the Sales Agent connects publisher inventory to buyer agents. The critical enabler is open protocols, especially AdCP, which Optable co-founded, that give all these agents a common language across organizational boundaries.” 

While advertisers are more likely to use agentic AI than publishers, there are some barriers to adoption for agent-driven buying. 

Advertisers are concerned about brand safety (63%) and loss of control (61%). In this sense, advertisers mirror publishers’ concerns about how to best align technology and compliance. 

Additionally, 43% of advertisers said they have difficulty integrating agent-driven buying with their existing tech stacks. Similar to publishers’ challenges with technical complexity and internal organizational readiness, this is likely connected to advertisers’ interoperability and a need for additional resources.

“When a system is making decisions autonomously at speed, the same thing that makes it powerful is what makes it risky — it doesn’t slow down when something goes wrong,” said Meher Patel, founder and CEO at Hector Ai. “Beyond that, there’s a real industry-level problem: two competing standards frameworks are emerging right now, and that fragmentation is slowing enterprise adoption. Brands can’t confidently build on infrastructure when the infrastructure itself is still debating what it is. That needs to be resolved. Until it does, the governance concern isn’t a perception problem, it’s a legitimate one.”

Similar to what publishers are looking for, advertisers are seemingly looking to agentic AI to increase the success rates of their campaigns. When asked where they see the most potential value from agentic advertising, nearly 8 in 10 advertisers cited improved performance KPIs (79%) and increased marketing success (79%). Nearly half (48%) saw more precise audience insights and discovery.

With 89% of advertisers using or building agentic AI, Patel described some of the implications for publishers.

“The real gap is on the publisher and sell-side, and it’s structural. Publishers are being asked to rebuild data infrastructure and expose their inventory in machine-readable ways while simultaneously dealing with AI-driven traffic declines eating into their core revenue,” Patel said. “That’s not a mindset problem — it’s a bandwidth and infrastructure problem. The bottleneck in agentic advertising isn’t advertiser appetite — it’s ecosystem readiness.” 

Ultimately, advertisers are looking toward AI rather than away from it. In the next 12–18 months, more than 8 in 10 advertisers are expecting to at least maintain current levels of investment in agentic capabilities, including 66% who plan to increase spend in agent-enabled buying environments. 

This is particularly noteworthy, as publishers are much more hesitant about adopting the technology but are setting up their infrastructure to better match external AI agents — indicating an awareness that advertisers are adopting agentic advertising and attempting to accommodate that.

“There are real misunderstandings slowing things down [investment], mainly the assumption that agentic means fully autonomous and unsupervised. The reality is semi-autonomous: agents handle the tedious work of audience discovery and campaign setup while humans retain strategic decision-making,” Crumbling said. “The bigger drag on adoption isn’t misunderstanding; it’s structural: tech stack integration challenges, measurement uncertainty and the fact that investment in governance and interoperability has lagged behind investment in execution tools. The industry is building the car while figuring out the rules of the road, and that trust gap slows adoption regardless of how well the technology works.”

Agencies are further along the adoption curve for AI, with the vast majority of agency respondents saying they are using AI to some extent — 92% to advertisers’ 89%. The majority of agencies are either building AI agents internally (45%) or using out-of-the-box, ready AI tools (35%).

“Right now, it’s less about one specific agentic platform and more about combining the AI capabilities already built into the major ad platforms with other automation and generative tools or building on top of them to solve for custom opportunities on behalf of clients,” said Tom Swierczewski, vp of media investments and partnerships at Goodway Group. “We’re using AI to help with things like optimization, audience insights, creative iteration and analyzing performance data faster. The bigger shift is starting to connect those systems so more decisions can happen automatically while teams stay focused on strategy.”

As far as the prevalence of agency clients using LLMs, 63% of agencies found that at least 41% of their clients are engaging with or are interested in using LLMs. So, while advertisers themselves may not be heavily investing in AI technology, they are looking for agency partners with access to it. 

“Clients are definitely asking about AI, but usually the conversation starts with performance and outcomes rather than the technology itself,” Swierczewski said. “They want to know how AI can help connect media, commerce and measurement so they can drive things like store visits, sales or other real-world results. A lot of the opportunity is using AI to build a more connected commerce infrastructure where those signals can actually inform strategy in a more unified way and drive real business results. AI is creating a higher bar for what agencies expect to deliver in terms of customization (i.e., creative), data-driven decisions and measurement.” 

Agency respondents agreed that advertisers are not looking to manage AI tools themselves. When asked about the primary role of agencies in an agent-driven ecosystem, 62% said the primary role is managing AI agents and tools. 

Agency respondents also recognize their role in handling AI agents and workflows on behalf of advertisers, including providing reporting (48%), integrating agents into existing workflows (45%) and managing client intent (45%).

“I think the role shifts from doing every step of the campaign manually to enabling the strategy that drives performance,” said Swierczewski. “AI can handle more of the operational work — things like optimization, signal processing and analysis — while agencies focus on defining goals, aligning media with commerce outcomes and making sure the right data signals are feeding those systems. In many ways, agencies become the layer that connects media, retail data and measurement into a more cohesive, connected commerce ecosystem.

“In 2026, there is even more complexity for brands to navigate, more tools that can add incremental value, more data to make informed decisions and this just increases the role and responsibility of agencies,” he continued. “At this stage of AI growth, there is a lot of value for agile agencies that adopt ways for faster test and learn, and a system for vetting and integrating new capabilities. Our belief is SMEs now have an even more unfair advantage when powered by AI tools to lift up organizations and clients.”

Overall, survey results found that agentic AI benefits most agency workflows. At least 1 in 5 respondents expects agentic workflows to benefit organization and pacing (25%), reporting and insights (23%), client strategy and consulting (22%) or media planning and forecasting (21%). 

This aligns with the high level of agentic AI adoption by agencies, indicating that they see value in technology and are keen on adding it to their tool belt.

Agencies are more focused on using agentic advertising for efficiency purposes — so far, a common demand among all respondents that outpaces their desire for lower costs. Thirty-one percent of respondents said improving task completion speed has the biggest impact, followed by 25% who said lowering business costs is most impactful.

Publishers, in a similar vein, want to use agentic advertising to increase demand discovery and improve yield and transparency, rather than decrease costs.  

While agencies have higher agentic AI adoption rates than both publishers and advertisers, they face a similar issue as publishers. When asked about what would accelerate investment, 47% of respondents said more AI education to lower the learning curve. 

One of the main challenges for publishers was technical complexity. Regardless of the type of company adopting agentic AI, the nascent technology is intricate and businesses are eager for more education about it.

“Formal education around AI should focus on practical use cases,” Swierczewski said. “That means showing how AI fits into things like media planning, optimization, reporting and how commerce or retail data can feed into those decisions. It should also cover things like setting guardrails, validating outputs and understanding where human judgment still needs to play a role. This is all in service of ensuring it has a tangible impact on the final process, and enhances that media-to-business outcome connection.”

Among all respondents, the primary factor they expect to determine the success of agentic advertising is increased investment from advertisers at 29%. This aligns with earlier responses that publishers are not investing in their own agentic advertising tools but are becoming more discoverable by external advertiser AI tools. 

Twenty-two percent of respondents said demonstrated performance outcomes will determine whether agentic advertising succeeds, indicating that advertisers may become more willing to adopt agentic AI once they see more successful use cases. 

“There’s no universal connector between planning tools, DSPs, measurement platforms and the agent ecosystem,” said Optable’s Crumbling. “That’s why open protocols like AdCP and MCP matter: they provide interoperability without bespoke engineering for each integration. The good news is that this is solvable without rebuilding from scratch. The agentic model asks organizations to make their data legible and machine-readable, not tear down what they have.”

While agentic AI is powerful, companies are still wary of the technology and want guardrails and standardization in place. Some respondents believe industry-wide standards (19%) and advertiser trust and governance (16%) are key to the scalability of agentic advertising.

To increase advertiser adoption of agentic advertising, Crumbling recommends: Open standards adoption so agents can interoperate across platforms; demonstrated campaign intelligence that demonstrates agentic workflows drive better KPIs and publisher data readiness, including machine-readable inventory.

For any publishers or agencies who may be waiting for more advertisers to adopt agentic AI or who are looking for further industry standards and guidance first, Crumbling advises not waiting. 

“Two-thirds of advertisers plan to increase agentic spend,” said Crumbling. “The question is whether you’ll be ready. Start by getting audience data organized and enriched, ensure your identity infrastructure supports multiple frameworks and explore technology partners who can provide turnkey agentic capabilities rather than building in-house. For those seeking standards: engage with the bodies defining them now, including the AgenticAdvertising.org, the Prebid Agentic Taskforce and the IAB Tech Lab. If your position is ‘we’ll adopt once standards are mature,’ you’re ceding the design of those standards to your competitors.

“The publishers and agencies who invest in data quality, identity and signal infrastructure, and governance controls today will capture the most value as agentic workflows become the default.”

---

About Optable

Optable is the Agentic Audience Platform for publishers and media companies, giving publishers the infrastructure to build identity, activate premium audiences and transact efficiently in an era of AI-powered advertising. As a founding member of AdCP, chair of the Prebid.org Agentic PMC Taskforce and IAB Tech Lab board member, Optable is built for the AI-powered workflows reshaping premium advertising.