Everyone in AdTech is talking about AI agents now.
In October 2025, a consortium of AdTech companies led by Scope3, Optable, Swivel, Yahoo!, PubMatic, and Triton Digital launched AdCP, an open protocol that connects AI agents to advertising platforms.
The new protocol sparked discussions about the future of agentic AdTech. As Emma Newman, CRO of EMEA for PubMatic, put it, the industry is at the dawn of “ agentic era, creating a common language for AI systems to collaborate across planning, optimisation and measurement”.
If AdCP takes off, it could solve many industry challenges. Data silos and bloated supply chains waste up to 55% of total programmatic spending.
The potential of AI agents goes beyond replacing OpenRTB. AI agents can successfully automate data analysis, creative production, audience targeting, and retail media campaign management.
In this post, we’ll explore successful advertising AI agent pilots. Publishers, brands, agencies, and AdTech vendors have all put these into action.
What is agentic advertising?
The key difference between AI agents and generative AI (GAI) products like ChatGPT lies in agents’ ability to act proactively (whereas LLMs typically respond to user queries). Besides, AI agents follow a multi-step process to complete tasks autonomously.
- Goal interpretation: The system receives objectives through natural language input. It determines the necessary steps to achieve them and plans its approach proactively.
- Tool and data access: The agent connects to relevant platforms, software systems, and data sources. It gathers the capabilities needed for execution, where the Model Context Protocol (MCP) becomes critical for standardized access.
- Autonomous execution: The agent acts independently. It adapts behavior based on real-time feedback and changing conditions without requiring human intervention.
- Output generation: The agent uses GAI to produce the required deliverables. This includes text, images, code, or video as part of completing its assigned tasks.
In AdTech, AI agents can help AdOps teams automate routine tasks such as data analysis, media buying, and creative A/B testing.
Although the range of applications for agentic advertising is still evolving, four distinct agent categories are emerging.
Campaign monitoring agents track bidding activity, budget pacing, and campaign metrics across programmatic platforms in real time.
Creative optimization agents generate, evaluate, and expand variations of creative assets aligned with specific audience characteristics and contextual signals.
Targeting agents refine audience definitions and maintain unified identity resolution across advertising channels to reach optimal consumer segments.
Measurement agents consolidate attribution data across connected TV, programmatic advertising, and retail media networks to deliver comprehensive performance insights.
Now let’s examine how publishers, brands, agencies, and technology vendors apply agentic advertising technology to solve day-to-day operational challenges.
Publishers
Publishers are adopting generative AI and AI agents to handle the repetitive AdOps work that eats up their teams’ time. The demand for agentic AI comes from media leaders wanting to run leaner operations. They need to stay competitive against platforms that have bigger budgets and more automation.
AI agents in sell-side AdOps have many uses. They manage yield optimization by checking fill rates and eCPMs with demand partners. They also handle inventory packaging by creating and updating programmatic deals based on advertiser needs. Additionally, AI agents navigate negotiations with direct clients and generate performance reports for advertisers.
#1. Negotiations with buy-side partners: The Sun
An average publisher relies on SSP partners to access diverse sources of demand. An SSP connection, in turn, opens the door to hundreds of demand partners and increases the workload on a publisher’s AdOps team.
DoubleVerify estimates that a publisher making $100 million in direct-sold revenue needs to spend $13.7 million each year. And also run a dedicated team of over 140 professionals.
How The Sun uses agentic media buying to solve this problem
To reduce the strain of programmatic media buying, The Sun created an AI agent that automatically connects and communicates with the tabloid’s buy-side partners.
Dominic Carter, EVP, Publisher at The Sun, told Digiday that the publisher is building an AI agent that will autonomously communicate with buy-side AI agents. The Sun’s team wants the agent to respond to buyer needs quickly. It will match brands with available inventory, negotiate terms, and close deals.
Under the hood, The Sun’s agent is powered by the Ad Context Protocol, making the company one of the protocol’s first adopters.
To support agentic advertising, The Sun improved its supply chain and created direct links with key partners – The Trade Desk and Ozone.
What this means for the industry
The Sun’s early adoption of AI agents via the Ad Context Protocol makes it clear: publishers are ready to commit to agentic media buying.
Following The Sun’s playbook, publishers who want to explore agentic media buying must modernize their infrastructure and build a leaner supply chain to ensure reliable, straightforward machine-to-machine programmatic negotiations.
#2 Automating deal management and negotiation: Hearst
For larger publishers, manually managing direct client negotiations across an extensive product portfolio creates inefficiencies and shifts sales teams’ focus away from revenue-generating activities.
At Hearst Communications, a media holding headquartered in New York City, account research would take employees 40 minutes per account on average.
Onboarding sales talent came with its own challenges. New reps had to deeply understand each of the 30+ products in Hearst’s portfolio.
Michael McCarthy, the company’s Senior Director of AI, Sales, and Business Solutions, found traditional methods too slow and ineffective. He claims standard approaches don’t help sales reps with no media experience get up to speed.
How agentic AI improves publisher productivity in deal management
McCarthy calls Hearst’s advertising agent a ‘computer use agent.’ Because it can navigate systems, open apps, browse the web, input data, and manage files all on its own.
Sales reps only guide the agent with commands like “research sales accounts for me” and log it on LinkedIn.
Hearst’s agent works by searching the publisher’s internal databases. It finds client info, pricing details, audience insights, past campaign data, and budget rules. The tool uses this information to generate media plans, complete CRM updates, conduct account research, perform pre-call planning, and create media proposals.
Hearst has also built an agentic knowledge base for training and onboarding. New sales reps can consult this platform to get personalized answers about products and processes.
Although Hearst’s agentic AI pilot is recent, AdOps, sales, and business departments are already seeing ROI uplifts both in employee productivity and deal negotiations. The time needed for account research dropped from 40 minutes per task to 2 minutes per task when supported by the AI agent.
Sales executives report that AI agents help them “show up better to customers and answer their objections more effectively”. Hearst documented a 153% increase in average sale value since implementation. Multiple sales reps have credited the AI assistant with helping them close six-figure deals by enabling data-informed conversations with advertisers.
What this means for the industry
Hearst’s successful pilot shows that AI agents provide quick, clear ROI. They do more than automate tasks; they also help boost revenue. In Hearst’s case, AI agents support sales teams and improve client negotiation by providing SDRs with up-to-date prospect research and relevant offerings.
#3. Improving cross-department workflows: DPG Media
Cross-border publishers, like DPG Media, a European media group active in Belgium and the Netherlands, face significant challenges. The companies struggle with a gap between regional teams and fragmented workflows.
DPG Media executives told Digiday that editorial teams struggled to coordinate tasks and access real-time information across the publisher’s newspaper, magazine, TV, and podcast portfolio.
Sales reps faced issues with disconnected systems. They had to switch manually between their order management system, ad server, and email clients to talk to prospects.
Using a mix of different tools slowed response times. It also raised the chance of mistakes when sharing campaign details.
DPG has operations in many countries, product lines, and departments. So, the company needed a single solution. The one that would help share information and automate everyday tasks.
How AI agents helped DPG Media connect 3000+ employees
As a solution, the publisher deployed an internal AI assistant, ChatDPG. The tool allows employees across all departments to build custom AI agents and query them for information or task execution.
As of October 2025, over 3,000 DPG Media employees are creating and using custom AI agents, and 1,500 interact with them daily.
The system connects with DPG’s order management and ad server. As a result, sales reps can create client emails using the latest campaign data.
The agent cut out the need to switch between platforms. It also automated workflows that once needed manual coordination.
What this means for the industry
ChatDPG is yet another promising use case. It showed how AI agents in AdTech help connect the global publisher team effectively. McKinsey reports that, on average, employees waste 20% of their time searching for information across disconnected systems.
DPG Media’s agentic pilot proves that agents can solve this problem. AI agents automate workflows, connect disparate databases, and break down operational silos.
#4. Reducing campaign reporting time: LG Ad Solutions
Campaign reporting bottlenecks are an industry-wide productivity drain in adtech and programmatic advertising.
Teams on all sides of the pipeline—media, brands, or agencies—report spending up to 40 hours on monthly reporting, rather than allocating that time to strategic campaign optimization and nurturing client relationships.
Dave Rudnick, CTO at LG Ad Solutions, shared that compiling reports for advertisers would take the company’s AdOps team an average of 2 full business days.
With no automated system in place, teams had to manually pull data from multiple advertising platforms, consolidate metrics across channels, run calculations, and create visualizations.
As a result, advertisers faced delays in getting performance insights. This made it hard for them to optimize in real-time.
They risked wasting budgets on weak campaigns or missing opportunities to build successful strategies.
How AI agents helped LG Ads Solutions cut reporting time
To make reporting easier, LG Ad Solutions launched Agentiv. It is an agentic AI platform that coordinates up to 30 specialized agents. These collect and organize data from brands, agencies, and adtech partners.
As a founding member of the AdCP protocol, LG designed the system to enable interoperability between agents from different supply chain partners.
The platform has already delivered tangible productivity gains for LG Ads Solutions. Campaign reporting AI agents reduced the time needed to prepare a report from 16 hours to 5 hours.
The publisher is now discussing integration opportunities with agency holding companies and brands to enhance their agentic media-buying operations through partner DSPs.
What this means for the industry
The agents at LG Ads Solutions are proving reliable enough to produce complex client reports. The company is a frontrunner in agentic advertising in the CTV space, and other major publishers are likely to follow LG’s efforts in the coming months.
Advertisers
Advertisers are deploying AI agents across their marketing operations to cut campaign management costs and timelines.
Large FMCG brands with global operations use agents to support localized campaign execution that their regional teams lack resources for, while smaller companies and startups deploy agents to access enterprise-level capabilities without committing to expensive agency services.
#1. Automating personalized cross-platform: Coca-Cola
Setting up targeting campaigns without agentic automation requires heavy involvement from marketing teams.
According to a DoubleVerify study, marketers spend 10 hours per week on manual campaign tasks, and this time increases exponentially as targeting criteria become more granular and campaigns go cross-platform.
Coca-Cola had to grapple with a labor-intensive campaign setup when it planned to target fast-food fans across Saudi Arabia on seven social media platforms and deliver personalized coupons through third-party mobile apps.
The campaign required monitoring each platform to spot people posting about fast food, matching those users to their mobile advertising IDs, and serving them targeted coupons on partner apps.
Executing this manually at scale—hundreds of thousands of coupons — would have required weeks of human effort.
How an AI agent helped Coca-Cola automate hyperpersonalized targeting
To monitor social media activity 24/7 and identify frequent fast-food posters, Coca-Cola built an AI agent that analyzed user posts across LinkedIn, X, Reddit, Tumblr, TikTok, YouTube, and Pinterest and identified fast-food chain goers.
The agent extracted user geographic and demographic data from each platform’s API and matched this information to mobile advertising IDs provided by ad tech partners, Index Exchange and Sharethrough.
Users identified by the agent would receive personalized meal coupons through third-party apps like the Huffington Post.
The agent autonomously ran the campaign for over 2 months, executed 8 million actions, and delivered 828,000 coupon ads to a highly personalized audience.
#2. Automating cross-platform creative optimization: L’Oreal
Industry research shows that personalized creatives drive a 50% higher brand lift compared to generic messaging.
However, 70% of marketers struggle to create more compelling DCO units to personalize their creatives better.
The traditional approach to creative optimization is manual and time-intensive. AdOps teams have to create several asset versions, run multiple rounds of test campaigns, export, organize, and compare results in spreadsheets without a unified source of truth.
For brands with global reach like L’Oreal, the bottlenecks of DCO are even more apparent.
To create messages that resonate with L’Oreal’s 1 billion users, the brand is under constant pressure to adapt campaigns to diverse cultural contexts (e.g., Japanese gardens or the streets of Paris).
Working with agencies to create localized creatives led to longer revision cycles. It kept L’Oreal from achieving the speed and personalization needed to capture TikTok and Instagram users in multiple regions.
How AI agents helped L’Oreal personalize creatives
L’Oréal addressed dynamic creative optimization bottlenecks by deploying an intelligent agent that automatically generates and localizes creative assets.
The company’s engineers used Google’s Imagen 3 and Gemini models through its CREAITECH lab to generate localized visuals and campaign assets from text prompts.
L’Oréal instantly generates photorealistic, localized shots of a product in culturally relevant settings across 20 EMEA markets.
The company is on track to integrate the AI agent, including Google’s Veo 2. Google’s state-of-the-art video generation model will enable the brand to convert static images into 8-second animated video clips with audio elements, trained on brand-specific styles.
The agent also uses Tidal to automate paid media buying across platforms.
This integrated agentic workflow brought L’Oreal 22% higher media efficiency and a 14% increase in campaign conversions in Nordic markets.
What this means for the industry
With AI agents directly involved in creative production and testing, advertising is moving from the automation of isolated tasks to end-to-end autonomous campaign execution. In this model, agents handle everything from asset generation to localization to media placement with minimal human intervention.
AI agents support brands in achieving personalization at scale that was previously impossible, but simultaneously threaten traditional agency models built on manual creative production and media planning.
#3. Orchestrating disconnected MarTech tools: Bayer
Industry surveys show that marketing teams allocate up to 30% of their marketing budget to inefficiencies caused by siloed tools.
Even AI agents, when disconnected from each other, fail to deliver the automation benefits that team leaders hoped to harness.
For Bayer, disconnected AI tools added layers of complexity instead of simplifying marketing operations.
Instead of successfully slashing repetitive work and enabling increased scale, the company struggled to manage a patchwork of disconnected systems. AI-assisted campaigns demanded extensive manual effort and human oversight across every stage.
Since Bayer operates in a regulated industry with strict privacy regulations and brand guidelines that must be embedded throughout the advertising process, the lack of control and fragmentation was particularly problematic.
How Bayer uses intelligent orchestration to manage AI agents
Bayer tapped Innovid Orchestrator’s orchestration layer to coordinate specialized AI agents across the advertising lifecycle
The pharmaceutical company uses Innovid’s platform to coordinate AI agents that automate basic advertising activities, including ad creation, delivery, measurement, and optimization.
Advertisers using the platform can create detailed guidelines for each step of campaign setup and specify exactly how AI agents should interact with each other and with third-party systems.
This connected framework delivers what Bayer prioritizes: faster insights, improved automation, and built-in compliance and governance.
What this means for the industry
Orchestration tools help advertisers move from deploying isolated AI agents to creating interconnected systems that work together across the entire advertising lifecycle.
As the AdTech industry begins to explore AI agents as helpful automation tools rather than a novel emerging trend, these orchestration layers will separate companies that achieve quantifiable efficiency gains from those stuck with fragmented systems.
Agencies
When it comes to AI agent adoption, holding companies are leading the race.
Basis Technologies reports that, in 2025, agencies have been more advanced than advertisers in using AI to reach audiences, from segment definition to campaign personalization.
However, this rapid adoption has outpaced governance and safeguards: over 70% of marketers have experienced AI-related incidents, including hallucinations, bias, and off-brand content, yet fewer than 35% plan to increase investment in AI governance or brand integrity oversight.
Even amidst these challenges, the financial pressure to adopt is intense. 50% of agencies worry that brands will bring AI capabilities in-house and reduce their reliance on partners.
#1. Campaign planning automation: Omnicom
67% of CMOs acknowledge they are overwhelmed with data.
This was the case for Omnicom teams as well. Every day, they have to track up to 10,000 data attributes across 2,000 individual client accounts, and as Jonathan Nelson, CEO at Omnicom Digital, put it for AdExchanger, “give up in frustration” after 30 minutes.
Manual audience research required teams to brainstorm marketing objectives from scratch and spend hours sifting through spreadsheet cells to identify relevant audience segments.
How agentic AI helped Omnicom automate campaign management workflows
Since 2024, Omnicom has supported campaign managers with AI agents that analyze 10,000 data attributes from sources like Experian and PlaceIQ and automate specific agency tasks.
In a new workflow, a “chief strategist” agent generates marketing objectives, an “audience intelligence” agent creates audience segments and profiles, and targeting agents recommend influencers and social platforms based on audience behavior.
Each agent is customized with brand-specific guidelines and matched to the right knowledge base, requiring strategists to craft precise prompts that define the persona, context, and desired output format.
The system allows users to verify agent-generated insights by downloading underlying data into Excel and automating the time-consuming task of sifting through spreadsheet rows that would otherwise require hours of manual analysis.
Omnicom piloted its in-house AI agents for the “Ah, Nuts!” campaign for The Planters, automating influencer matching and enabling teams to prioritize strategic tasks.
What this means for the industry
In the past, AdOps and data analytics teams had to manually coordinate to process vast volumes of audience and campaign information. AI agents are rewriting the playbook by becoming on-demand data analysts for AdOps teams.
Fully agentic workflows allow teams like Omnicom to instantly synthesize insights from thousands of attributes and shift from data processors to prompt engineers who guide automated intelligence.
#2. Building self-service agentic solutions for clients: WPP
Historically, agents have struggled to find offerings that meet the needs of two underserved market segments that couldn’t access full-scale agency services.
One of those is global brands whose local marketing teams lack resources to manage complete campaigns.
The others are growing businesses, like tech startups, with small teams that are not ready to commit to large marketing departments.
Large agencies charge monthly retainer fees ranging from $1,000 to $10,000, averaging around $2,500 per month. At the same time, 45% of small businesses spend less than $1,000 annually on marketing due to limited financial resources.
Up-and-running businesses need affordable access to enterprise-level AI tools, data capabilities, and campaign execution without the commitment and overhead costs of engaging a full-service agency.
Creating WPP Open Pro, an agentic self-service offering, enabled WPP to tap into these segments without jeopardizing profitability.
How agentic AI enables WPP’s self-service capabilities
WPP launched WPP Open Pro, an agentic AI platform that centralizes campaign workflow in a single interface.
The agency’s engineers leveraged WPP’s $400 million investment in Google’s AI models, Gemini, and Veo. This tech empowers platform users to create and localize campaign creatives with minimal involvement from the agency.
WPP Open Pro operates on a “pay for what you use” pricing model rather than traditional agency retainers, which average $2,500 per month.
WPP positions Open Pro as both a client-acquisition tool for businesses constrained by small marketing budgets and a way to extract value from its existing AI infrastructure investments.
What this means for the industry
Now that clients are seeking greater autonomy and flexible pricing, agencies are exploring “Agency as a Service” models.
In the future, holding companies are likely to commoditize campaign execution through self-serve platforms.
On the one hand, platforms like WPP Open Pro allow agencies to tap into underserved market segments. On the other hand, agencies will have to go the extra mile to retain high-grossing accounts and focus heavily on strategic tasks such as brand positioning and customer experience design.
#3. Cutting programmatic waste: Butler/Till
Due to the mechanics of the programmatic marketplace, agencies typically have to accept the misalignment between their programmatic buying platforms and campaign performance goals.
DSPs are engineered to surface bid requests for impressions that agencies are most likely to purchase, rather than for impressions that will actually convert or perform well for the campaign.
Legacy algorithms also tend to be biased toward cheap reach and direct budgets toward MFA sites. Conversion data coming from these placements is unreliable because users are more likely to accidentally click on an excessive number of ads on the page.
Butler/Till realized the magnitude of this inefficiency when analyzing performance metrics. One of the agency’s campaigns was running across a bloated inventory that included 52% more domains than necessary, many of which were nonperforming or made-for-advertising (MFA) sites.
Without more sophisticated tools to identify truly performant inventory, Butler/Till lacked meaningful control over how client budgets were allocated.
AI agents helped Butler/Till prioritize performance over the number of impressions
Butler/Till deployed the SCaLE (Smart Curation and Learning Engine) optimization tool by SWYM.ai to create a programmatic buying system aligned with actual campaign performance rather than platform bidding incentives.
The AI agent operates by analyzing both sell-side optimization signals from Index Exchange and buy-side signals from Google DV360. It will simultaneously pull attribution data from Google Marketing Platform’s Floodlight conversion and event tracking to understand which impressions drove conversions.
Butler/Till’s agent identified the bid request characteristics that correlate with strong performance: domains, geos, ad sizes, device types, and channels. The platform used this data to curate private marketplaces that contained only impressions from purchase-ready users. Through an API integration with the supply-side platform, the AI agent dynamically packages lookalike impressions and updates the curated PMPs daily, continuously refining its selection based on real-time optimization signals.
Since introducing the agent in early 2025, Butler/Till saw a 56% increase in conversion rate and a 26% decrease in cost per conversion compared to the control portion of the campaign that didn’t use the tool.
The agent reduced the number of domains the campaign ran on by 52% by eliminating nonperforming inventory and avoiding MFA placements.
What this means for the industry
AI agents can help agencies break free from the fundamental conflict of interest in programmatic platforms, where DSPs prioritize impressions that buyers will purchase rather than those that convert.
Intelligent bid managers can create custom algorithms tailored to campaign goals and independent of legacy systems biased toward cheap reach and MFA placements.
AdTech vendors
#1. Optimizing retail media budgets: Skai
EMarketer’s data shows that marketers going all-in on retail media are struggling to manage reports from over 200 retail media networks (RMNs).
Each RMN partner requires marketers to allocate up to a day of their time for analysis and optimization.
This pain point is becoming more critical as US commerce media ad spending surges 21.8% this year, pressuring marketers to optimize significantly larger budgets across an expanding number of data sources.
Skai addresses RMN data fragmentation with AI agents
Skai’s new agentic AI solution, Celeste, is marketed as an around-the-clock analyst that analyzes performance data from over 200 retail media partners.
The tool brings in competitive intelligence and historical cross-channel insights to deliver strategic recommendations and automated reporting.
Early adopters of Celeste are reporting performance improvements of 30% to 50%.
Skai’s President, Gil Sadeh, also believes the agent can accomplish tasks that took marketing teams “an entire day” in “less than a minute.”
#2. Enabling interoperability between AI agents: AdCP by a consortium of AdTech vendors
Once AI agents become more widespread in the AdTech ecosystem, the industry will have to face the challenge of integrating them.
At the moment, AI agents are predominantly developed in isolation by different teams using disparate frameworks and deployed across varied infrastructures.
In a few years, this may create a digital “Tower of Babel” in which hundreds or thousands of industry agents are siloed, unable to communicate, collaborate, or share knowledge effectively.
How AdCP helps solve AdTech AI agent fragmentation
AdCP (Ad Campaign Protocol) is an open standard that enables AI agents from different advertising platforms to communicate and coordinate through standardized message formats and interoperability frameworks.
Developed as a joint initiative by Yahoo, Optable, PubMatic, Scope3, Swivel, and Triton Digital, AdCP establishes common methods for agents to discover each other’s capabilities, exchange campaign data, delegate tasks, and collaborate on media buying workflows.
At the time of writing, the AdCP protocol is still new with no published performance metrics or adoption rates. However, by the end of 2026, the founding members plan to accelerate adoption and add capabilities for creative generation and performance attribution.
What this means for the industry
AdCP steps in to address the risk of AI agent fragmentation that already plagues traditional martech stacks.
If the protocol gains broad adoption, it can help publishers and advertisers power away from walled gardens and intermediary-heavy programmatic infrastructure toward direct agent-to-agent transactions between buyers and sellers and slash the “AdTech tax”.
Bottom line
The advertising industry is already seeing tangible wins from AI agent adoption across all stakeholder groups.
Brands are using agents to enforce global creative consistency and accelerate asset production at scale. At the same time, agencies leverage them for audience segmentation, competitive intelligence, and keyword strategy that once required days of manual analysis.
Publishers have deployed agents that transform first-party data access for sales teams, enabling real-time responses to client briefs. AdTech vendors are embedding specialized agents for media mix modeling, forecasting, and performance optimization directly into data environments.
Early adopters report cost reductions and efficiency gains in campaign execution and faster decision-making, proving that AI agents deliver a measurable operational and financial impact when properly implemented.
However, these wins remain fragile without addressing three critical caveats that threaten to derail broader adoption.
Interoperability stands as the most pressing challenge: the majority of AI agents operate in isolation, unable to communicate across platforms or share context between creative, media, and analytics systems.
While agentic AdTech is capable of delivering results to everyone in the pipeline, without focus and investment in orchestration layers, governance frameworks, and compliance safeguards, early wins could give way to fragmented chaos and regulatory backlash that stifles the technology’s potential.
