1 MoEngage acquired Aampe — one AI agent per customer, in cash
On June 23, MoEngage acquired Aampe, a San Francisco AI infrastructure company that provisions a dedicated autonomous AI agent for every individual customer of a brand. The all-cash deal brings Aampe's reinforcement-learning engine natively into MoEngage, creating a single system where workflow agents (for marketers) and decisioning agents (acting for each user) operate together. Aampe's founders — Paul Meinshausen, Schaun Wheeler, Sami Abboud — join MoEngage to lead Agentic Decisioning.
The headline isn't the price; it's the architecture. "One agent per customer" is a fundamentally different bet than "segments at scale" — it replaces audience cohorts with per-user decisioning, which is what 1:1 personalization actually requires. If you've been buying segmentation tools for ten years, this is the genre breaking. Ask your engagement platform whether it's heading toward per-user agents or just better segments. The answer determines whether you compete with the next generation, or merge with it.
2 DISQO shipped AI Search Lift — the first measurement product for LLM advertising
DISQO launched AI Search Lift, the first outcomes measurement product built to trace the incremental effect of advertising inside large language models and AI-powered search interfaces. The system uses a consumer-consented research panel and an exposed-versus-control methodology to determine whether media drove brand exploration in AI environments. Early measurement across automotive, insurance, beauty, CPG and travel found higher-consideration categories show deeper engagement inside LLM conversations.
Measurement is the unlock everyone's been waiting on. Without it, the "ChatGPT as third gatekeeper" story stays anecdotal — you can see citations, but not whether your media drove them. AI Search Lift puts a real lift number against AI-mediated discovery, which is what budgets actually move on. Get this on your test plan this quarter; the first brands with proven AI-search ROI will defend share while the rest are still arguing about the methodology.
3 LiveRamp opened LAB — third-party agents on its rails, via MCP
LiveRamp launched the LiveRamp Agent Builders (LAB) program, a formal initiative letting third-party AI companies bring purpose-built agents onto LiveRamp's platform and make them available to its full customer base via APIs and MCP servers. Alongside it, LiveRamp piloted agentic programs tailored for food delivery, specialty retail, and grocery commerce media networks, layering orchestration protocols on top of existing data clean rooms.
LAB is LiveRamp choosing the App Store path: be the surface where agents distribute, not just the company that builds them. For independent agent startups, this is the most credible distribution channel in adtech — you get LiveRamp's enterprise customer base without a sales cycle. For buyers, it means the procurement question shifts from "which DSP?" to "which agents from which builders, hosted where?" Read the catalog like a media plan, not a vendor list.
4 Horizon Media wired its buying agent to 12 partners and every major media owner
Horizon Media announced agentic buying capabilities inside HorizonOS' Blu platform, with an open ecosystem of partners: Affinity Solutions, Innovid, Kartel, Kochava, Magnite, Predactiv, Sightly, Smartly, Vidmob, Vurvey Labs, ZeroToOne.AI on the tech side; and Clear Channel Outdoor, Disney, Fox, iHeartMedia, NBCUniversal, OUTFRONT Media, Snap and TikTok on the media side. Agents make real-time decisions across channels while humans keep strategy oversight.
This is the independent agency answer to the holding-company agent stacks Adobe rolled out at WPP and Omnicom. Horizon's bet is open ecosystem beats proprietary tech — same pitch SSPs made against walled DSPs a decade ago, and it largely worked. For mid-market advertisers, this is the path: an agentic stack you don't have to renegotiate when one vendor merges with another. Ask your agency which integration layer their agents speak.
5 Zeta Global tied its data to Palantir's infrastructure — martech meets defense-grade ops
Zeta Global formed a strategic alliance with Palantir Technologies to build an AI infrastructure layer that ties customer and operational intelligence to marketing execution. The pitch is "agentic marketing benchmarks" — using Palantir's industrial-strength data plumbing to host Zeta's decisioning across the enterprise.
Palantir doing marketing infrastructure should land harder than it has. The company built its business making operational data usable in environments where mistakes are expensive, and that's now the bar for AI agents touching ad budgets. The takeaway for CMOs: the next-gen martech stack is going to feel more like an industrial data platform than a SaaS dashboard, with audit trails, lineage and access controls front-and-center. If your incumbent CDP doesn't have those, this is the conversation that ends the contract.
6 Dentsu launched Idea Builder on Google Cloud — with full audit trails baked in
Dentsu Creative announced Idea Builder, an enterprise generative AI platform built natively on Google Cloud. Creative teams move from idea to structured development to high-quality concept images and video; the system captures the full conversation history, audit trail and cost log at every step. It's already in active use across the dentsu network.
The interesting part isn't the gen-AI — it's the audit trail. Holding companies are finally building creative tools that solve the part their finance and legal teams actually care about: who decided what, when, using which prompt, at what cost. That makes Idea Builder more defensible to clients than ten dashboards full of variations. If you're being pitched gen-AI creative without a usable audit log, ask why — "we made it with AI" is going to be a question, not an answer.
7 Smartly's $7B memory layer met Salesforce's $59% commerce uplift — same week, same architecture
Two days apart, Smartly's Synapse (an AI orchestration and memory layer behind $7B in ad spend) and Salesforce's biggest Agentforce Commerce release (with retailers running their own agents growing sales 59% faster) both shipped the same idea: agents need a shared memory across functions, or they re-invent context every conversation. Smartly's lives on the buy side; Salesforce's lives at the storefront.
The convergence is the story. Marketing platforms (Smartly), commerce platforms (Salesforce), engagement platforms (MoEngage) and identity platforms (LiveRamp) are all shipping "memory layers" within a single week. That means the architecture is settling: every agentic stack will need a brand-specific memory, and the vendor that owns that memory owns the relationship. Pick the memory you trust most — you'll be living with it for a decade.
8 Forrester & the 4As: 90% of US agencies use gen AI, half use agentic — and creativity is the bill
Forrester's landmark research released at Cannes, co-authored with the 4As, found nine in ten US marketing agencies now use generative AI and half use agentic AI. The same study warned that agencies are systematically trading creativity for cost savings, and that long-term brand growth is starting to show the cost. Veterans like Sir John Hegarty echoed the warning from the festival stage: brands are overly reliant on targeting and technology while under-investing in storytelling.
This is the inconvenient truth of the festival. The same week the industry celebrated agents, the actual research showed creativity is contracting under the same pressure. As a buyer, this is your leverage: the agencies that protect creative quality while running their AI stack at full tilt will become more expensive, and more worth it. As an agency, this is a positioning warning — if your AI story is purely cost-out, you're in a price war you can't win.
9 A24 wrote Google DeepMind a $75M check — for the AI tools, not the content
Outside Cannes but landing in the same week, A24 announced a $75 million partnership with Google DeepMind to co-develop AI tools for film production and distribution — one of the first big-studio deals with a frontier AI lab. The framing was carefully chosen: tools to support creative workflows, not models that generate finished films, with filmmakers keeping creative control.
For brand marketers, A24 is the canary. If the most creatively-led studio in America is putting $75M into AI workflow tools, "AI for creative ops" is no longer a niche debate — it's standard infrastructure for anyone making premium video. The brand parallel: stop treating AI as a "do we / don't we" question on individual campaigns, and start budgeting it as a line item in production. A24's framing — tools, not content — is the cleanest defensible position out there.
10 Infosys, the ANA's CMO Growth Council and LIONS launched the CMO AI Hub
Late in the week, Infosys, the ANA's Global CMO Growth Council, and LIONS launched the CMO AI Hub — an AI-powered, peer-learning platform for the council's 1,200+ chief marketers (chaired by P&G's Marc Pritchard). Members type business challenges in plain language, get context-rich answers drawn from curated research, executive perspectives and case studies, and connect to other CMOs through the platform. It runs on Infosys Aster with governance and privacy-by-design.
"A chatbot for CMOs" sounds twee until you read the membership list. The strongest competitive moat in marketing strategy isn't a vendor — it's access to what other senior buyers are doing, and this is now a queryable database of that. If you're a CMO at one of the 1,200, the Hub is your unfair advantage. If you sell to them, expect a more informed buyer with fewer patience for vendor pitches that aren't grounded in real peer outcomes.