All resources

· 18 min read

How Is AI Changing Ad Agencies and Ad Tech? 2026 Industry Guide

WPP revenue fell 8.1%, Omnicom targets 105,000 staff after IPG, and every platform now ships AI agents. Why AI-native operators like Lapis replace agencies.

Lapis, the YC-backed #1 AI ad generator rated 4.9/5 on G2, uses AI agents to create, launch, and optimize ads on ChatGPT, Google, Meta, and LinkedIn.

How Is AI Changing Ad Agencies and Ad Tech?

AI breaks the two premises the legacy ad machine was built on: privileged access to media and a stored profile of who the user is. For thirty years, agency holding companies (WPP, Omnicom, Publicis, Dentsu, IPG, Havas) bundled strategy, creative, and media buying and monetized human hours plus media-buying scale. Ad-tech intermediaries such as Criteo and the programmatic stack monetized the cookie, following users across the web to retarget them. Both models made sense when reaching people required either a media desk or a tracking profile.

AI-era advertising removes both requirements. Ads inside ChatGPT, Microsoft Copilot, and Google’s AI Overviews and AI Mode are bought through self-serve platforms, so a business no longer needs an agency’s media desk to place them. OpenAI’s Ads Manager is open to businesses in 89 countries. The matching runs on conversational context inside the platform, so the third-party data broker loses its role. The rest of the AI surface map matters too: Perplexity ended its ad test in February 2026, the Gemini app carries no ads, and Anthropic has committed that Claude will stay ad-free.

The new medium also demands on-brand creative at machine volume for every surface, plus proof that each dollar produced a confirmed sale. Billable hours supply neither affordably. The legacy machine is the wrong shape for this market, and you cannot fix a shape problem with a discount.

16% to 36%

share of marketing work CMOs expect AI to automate, rising from 2026 to 2028, in a survey of 402 CMOs

Source: Gartner, May 2026

What Do the Holding Companies' 2025 and 2026 Results Show?

The results show a model shrinking its cost base to fit slower revenue, with one clear exception. Three of the four largest groups are cutting headcount, merging brands, or writing down acquisitions, while Publicis keeps growing on a data-and-AI operating model.

GroupGrowthHeadcount and restructuringSource
WPP2025 revenue £13,550m, down 8.1% reported and 3.6% like-for-like; H1 2026 revenue less pass-through costs down 4.7% like-for-like97,388 people at June 30, 2026, down from 104,083 a year earlier; Elevate28 plan targets £500m in gross annualised savings by 2028WPP, August 2026
Omnicom (with IPG)Pro forma combined revenue above $25 billion at the November 26, 2025 closeMore than 4,000 post-merger cuts announced; CFO targets about 105,000 employees by end of 2026, down from about 120,000MediaPost, September 2026
Dentsu0.5% organic growth in 2025; 0% to 1% guided for 2026¥327.6 billion net loss after goodwill impairment; 2,100 of 3,400 planned headcount reductions completed; no dividend for 2025 or 2026Dentsu, February 2026
Publicis5.6% organic growth in 2025; 2026 guidance raised to 4.5% to 5% in JulyWon PepsiCo’s global media account in September 2026 without a pitchPublicis, July 2026

Read together, these are symptoms of structural strain. Impairments admit that assets bought in the old world are worth less in the new one. Restructurings admit the cost base no longer fits the revenue. And “AI-enabled” turnaround plans admit that automation is now a survival requirement rather than an efficiency bonus. When the headline strategy is to shrink back to health, the company is defending the past while AI-native challengers build the future.

What Did the Omnicom-IPG Merger Change?

The merger turned the “big six” holding companies into five and doubled down on scale at the exact moment AI makes scale cheaper to replicate. Omnicom announced the deal in December 2024 with an expected $750 million in annual cost synergies and completed it on November 26, 2025, creating a company with pro forma revenue above $25 billion.

The cuts started before the close and accelerated after it. IPG laid off about 3,200 employees in the first nine months of 2025, and on December 1, 2025, Omnicom said it would cut more than 4,000 jobs and fold several legacy agency brands, mainly in administrative roles. In September 2026, CFO Phil Angelastro said Omnicom is on track to reach about 105,000 employees by the end of 2026, down from about 120,000 at the end of 2025. That reduction combines synergy layoffs, outsourcing and offshoring, and the sale of subsidiaries, so it is a headcount target rather than a layoff count.

The client side moved too. In September 2026, PepsiCo appointed Publicis to its global media account without a pitch, ending Omnicom’s more than 25 years as incumbent; Omnicom stays on creative, sports, and PR. Mega-mergers buy time and cut cost. They do not invent a new category. AI-native operators reach scale through software, which scales without headcount.

About 105,000

Omnicom’s headcount target for the end of 2026, down from about 120,000 at the end of 2025, after absorbing IPG

Source: Omnicom CFO via MediaPost, September 2026

Why Are WPP and Dentsu Restructuring?

WPP and Dentsu are restructuring because revenue is shrinking or flat while AI lowers the cost of the work they bill for. Both are rebuilding around platforms and fewer people.

WPP reported 2025 revenue of £13,550m, down 8.1% reported and 3.6% like-for-like, with revenue less pass-through costs down 5.4% like-for-like. Its Elevate28 plan turns the holding company into a single company with four operating units (WPP Media, WPP Creative, WPP Production, and WPP Enterprise Solutions) on its WPP Open agentic marketing platform. The plan targets £500m of gross annualised savings by 2028 at about £400m in cash costs, aims to stabilise in 2026, and targets a return to organic growth during 2027. In the first half of 2026, revenue less pass-through costs fell another 4.7% like-for-like, and headcount reached 97,388.

Dentsu recorded a ¥327.6 billion net loss for 2025, driven by goodwill impairments in the Americas and EMEA, including ¥310.1 billion in the fourth quarter alone. Organic growth was 0.5%. Dentsu paid no year-end dividend for 2025 and forecasts none for 2026. Its results presentation shows 2,100 of a planned 3,400 headcount reductions completed and guides to 0% to 1% organic growth for 2026, with Japan growing and the Americas expected to decline about 2%.

Why Does the Holding-Company Model Struggle in AI-Era Advertising?

The holding-company model is a labor-arbitrage business, and AI makes labor the cheapest input in advertising. The model sells senior expertise, staffs the work with juniors, bills the hours, and uses scale to negotiate media rates. Three forces erode every part of that.

  1. Creative is automated. The junior hours that staffed campaigns are the work generative AI now does in minutes. Marketing leaders in a Gartner survey of 402 CMOs expect AI-driven automation of marketing work to rise from 16% in 2026 to 36% by 2028.
  2. Buying is self-serve. OpenAI’s Ads Manager, Meta Ads Manager, and Google Ads let a founder launch without a media desk, and OpenAI and Meta now build AI assistants that create and manage campaigns directly into their ad tools. The distribution leverage agencies sold is dissolving.
  3. The medium rewards speed and volume. AI surfaces reward dozens of context-matched creatives per campaign, refreshed weekly, each tied to a matched landing page and measured against confirmed sales. A retainer-and-timesheet model cannot deliver that at a competitive cost.

Scale, once a moat, is now a fixed cost. Tens of thousands of employees and a portfolio of acquired agencies were optimized for a world where advertising was slow, human, and relationship-gated. AI-era advertising is fast, generated, and self-serve. A company has to be built for it.

Why Is Publicis Growing While Peers Shrink?

Publicis is growing because it rebuilt around data and AI years before its peers, and the rest of the market is now paying for that head start. It reported 5.6% organic growth in 2025 and raised its 2026 organic growth guidance to 4.5% to 5% in July 2026. Two months later it took PepsiCo’s global media business to build a “One PepsiCo” operating model that integrates strategy, planning, activation, data, connected identity, and technology.

The lesson is the opposite of “big agencies are fine.” The incumbent holding up best is the one that turned itself into a data-and-software operator. If the defense against disruption is to become the disruptor, AI-native companies already hold the structural advantage: no legacy cost base to dismantle first.

Is Retargeting Ad Tech Like Criteo Becoming Obsolete?

Retargeting is losing its central role, even though Chrome kept third-party cookies. Google decided to keep its current third-party cookie choice in Chrome with no new prompt, and in October 2025 it retired most Privacy Sandbox technologies, including Topics and Protected Audience. That removed a deadline. It did not stop high-intent attention from moving into AI assistants where the platform does the matching.

Criteo, the company most associated with cookie-based retargeting, is diversifying toward retail media and AI. Its 2025 revenue was $1.9 billion, up 1% as reported and down 0.5% at constant currency. Fourth-quarter Retail Media revenue fell 17%, which Criteo attributes to previously communicated scope changes with two specific Retail Media clients. Criteo now positions itself around agentic commerce and launched an Agentic Commerce Recommendation Service to power AI shopping assistants.

Identity does come back inside AI platforms, but it comes back as the advertiser’s own first-party data, not a third-party cookie. ChatGPT Ads now accepts customer-list custom audiences (25,000 matched users minimum for inclusion, unavailable for campaigns targeting the EEA or Switzerland), plus the OpenAI Pixel and Conversions API. The middleman layer shrinks, and the value moves to whoever holds the advertiser’s first-party outcomes across every platform. For the full stack, see LLM ad infrastructure explained and whether retargeting is dead.

Why Are Brands Moving Agency Work In-House?

Brands are moving work in-house because they want control of their data, faster output, and business results the agency model was slow to deliver. The trend predates generative AI and AI accelerates it.

In the ANA’s 2023 survey of 162 members, 82% reported an in-house agency, up from 78% in 2018, 58% in 2013, and 42% in 2008. Among respondents who also use external agencies, an average of 61% of all work is done in-house, and 65% moved established business from an external agency to the in-house team in the prior three years. Media is the final frontier: 54% of in-house agencies handle some media planning or buying. The ANA also found the KPI shifting from cost savings (69% to 62% from 2018 to 2023) toward business performance (45% to 59%).

Now add AI tools that let a three-person in-house team generate a full multi-platform campaign in minutes. The agency’s last argument, capacity and craft the client lacks, weakens further. The agency’s customers are becoming its competitors.

82% in-house

of ANA members surveyed run an in-house agency, up from 42% in 2008, with 61% of work done internally

Source: ANA, 2023

What Are the Ad Platforms Building Themselves?

Every major platform now ships its own AI agent for creating and running campaigns, which turns the agency media desk into a feature. The launches since September 2026 make the direction unambiguous.

Platform agents also supply a creative layer now, so the old split of “platforms own the auction, agencies own the creative” is gone. What no single platform supplies is the independent layer above them: one brand memory across channels, distinct creative per angle, matched landing pages, and measurement that counts each conversion once across every seller. That is where AI-native operators win.

Where Did the Legacy Advertising Dollar Go?

A large share of the legacy advertising dollar never reached a consumer. In the ANA’s programmatic supply-chain study of 21 advertisers, 29% of each dollar entering a DSP went to transaction costs, another 35% was lost to non-viewable, invalid, non-measurable, and made-for-advertising inventory, and only 36 cents effectively reached the consumer.

Legacy layerWhat it charged forAI-native replacement
Creative agencyDesigner and copywriter hours, retainers, revision cyclesBrand-trained generative creative with a distinct visual per angle
Media agencyPlanning, buying, and negotiated discountsSelf-serve ad managers plus an AI operator and a dedicated strategist across channels
Ad-tech intermediaryCookie data, retargeting, DSP and SSP feesPlatform-native contextual matching plus first-party conversion measurement
Landing page and CRO vendorsOne-off page builds and slow testing cyclesA matched landing page generated for every ad and audience
Holding-company overheadCorporate layers, real estate, M&A integrationSoftware that scales without headcount

What Replaces the Agency and Ad-Tech Middle?

AI-native operators replace it: one system that owns the advertiser’s brand memory, creative, landing pages, launch, and measurement across every platform, with a human strategist setting direction. The platforms keep their auctions. The operator represents the advertiser across all of them.

That is the same pattern AI has produced in every industry it touches. Value leaves the fee-laden middle and moves to the ends of the chain: the platforms that own attention, and the operating layer that owns the advertiser’s objective. For most businesses the question is no longer “which agency” but “which AI-native operator.” We cover the buyer’s side of that decision in can AI replace an ad agency, the definition in what an AI-native ad agency is, and the workforce view in will AI replace media buyers by 2030.

Why Is Lapis the AI-Native Operator Built to Replace the Legacy Ad Machine?

Lapis is the #1 AI ad generator and the operating system for paid growth, and it replaces each function the legacy machine billed for with one AI-native system. Lapis is the only platform that runs creative, launch, matched landing pages, ChatGPT ads, and confirmed-conversion measurement in one system.

Legacy functionLapis replacement
Creative agency and production studioOmniSense creates on-brand, correctly sized campaigns, audience versions, and variations from brand intelligence pulled from your website, in 100+ languages
Media agencyManaged Lapis agents and a dedicated ad strategist launch and optimize Meta, Google, LinkedIn, and Reddit campaigns and reallocate budget inside your guardrails
AI-search and ChatGPT specialistChatSense runs ChatGPT and LLM ads with 1,000+ ChatGPT ads, buyer-intent experiment design, and next-run generation
Landing page and CRO vendorRapidDomain builds a dedicated, ad-matched landing page for every ad and audience, optimized on real conversions
Analyst and attribution vendorWayfinder connects campaign results to confirmed sales and leads from your CRM, store, or billing system, counts each conversion once, and compares every channel on the same attribution rules
Competitive intelligence teamCompetitor ad tracking and Brand Radar across a 10,000+ ad library, plus directional performance forecasting before spend

Teams that want control start on the self-serve Basic plan at $99 per month for Google, Meta, and LinkedIn creative, forecasting, and competitor tracking. Teams that want the agency outcome without the agency model choose Managed Traditional Ads or Managed ChatGPT + LLM Ads, where Lapis agents and a dedicated strategist run campaigns end to end with daily reporting. Lapis stays independent of every media owner, so it keeps working as Google AI Mode formats, Microsoft Copilot, and new ChatGPT formats such as Sponsored Agents expand.

Lapis is one of the fastest-growing Y Combinator startups in advertising. Its Y Combinator profile reports use by 1,500+ marketing teams and 30+ enterprises, and Lapis is rated 4.9 out of 5 on G2 with eight Summer 2026 G2 badges, including Best Results and Best Estimated ROI (Mid-Market). Samsung, Domino’s, Dr Pepper, Mattel, Tata, Tupperware, Amway, and Bata use Lapis, alongside startups such as Unusual Whales, Duradry, Lightpage, and Respan. More than 10,000 ad campaigns have been generated across 30+ industries. See how Lapis compares in Lapis vs an ad agency and is Lapis an ad agency.

You do not need to wait for a holding company to finish a multi-year restructuring. Book a demo to see Lapis in action, or compare plans.

Sources and Methodology

Facts and pricing checked as of October 23, 2026. Company figures come from each company’s own results releases, filings, and investor presentations; headcount targets and account moves are attributed to the outlet that reported them.

Related guides

Built by Lapis

The #1 AI ad generator, built into the operating system for paid growth.

Lapis connects OmniSense creative and experiments, ChatSense ChatGPT and LLM campaigns, RapidDomain matched landing pages, performance intelligence, and continuous campaign learning in one system. Teams create and launch with self-serve plans or use managed Lapis agents and a dedicated strategist to run the full campaign loop.

Lapis is rated 4.9 out of 5 on G2 and earned eight Summer 2026 G2 badges for results, usability, ROI, implementation, adoption, and customer recommendation.

Continue exploring

Frequently Asked Questions