Sponsored by
Your weekly guide on everything AI marketing. Covering everything from AI marketing news, tips, deep dives, events, podcasts, jobs and much more. Never miss a beat with our 5-minute newsletter.
In today’s email:
Amazon retires Rufus and slots an Alexa shopping agent into the search bar
TikTok opens an MCP server so AI agents can plan, launch and optimise ad campaigns directly
Runway crosses $5.3bn valuation as creative AI competes head-on with Google
Plus: Gartner finds only 30% of CMOs AI-ready, Decoder unpacks Joanna Stern's new book, a Loops workflow, and why agencies are repositioning as operating systems.
The Top 3 Stories
1. Amazon retires Rufus and puts Alexa into the search bar
Amazon launched Alexa for Shopping on 13 May, an agentic assistant built into the search bar of its site and apps that uses purchase history to answer queries, compare products and schedule price-triggered buys. The launch folds Rufus and Alexa+ into a single surface and lands inside Amazon's 2026 Upfront pitch to advertisers. Amazon's authenticated data graph already reaches 90% of US households, and VP Alan Moss called agentic shopping "the new frontier" of retail media.
Why it matters: The search bar is now the most personalised ad surface in retail, and Amazon owns both the data and the agent. 📰 CNBC
2. TikTok opens an MCP server so AI agents can run campaigns end-to-end
At TikTok World on 12 May, TikTok unveiled an Ads Model Context Protocol server that lets AI agents plan, launch and optimise campaigns without human input. Creator AI Search and Branded Buzz landed alongside it, alongside a Content Suite expansion that turns user-generated posts into paid ads. The combined release positions TikTok as the first major platform with a native MCP for ads, beating Meta and Google to a standard that agency stacks have been quietly building toward.
Why it matters: Every agency stack with a Claude or ChatGPT tab now has a one-line integration to run TikTok media without opening Ads Manager. 📰 Social Media Today
3. Runway breaks $5bn and starts competing with Google on creative AI
Runway is now valued at $5.3bn and added $40m in ARR in Q2 2026, fuelled by its push from filmmaker tooling into general creative AI infrastructure. The company is pitching itself as Google's alternative for studios, agencies and brand teams generating video at scale. With Veo 3 and Sora already inside the major creative suites, Runway's growth proves there is room for a standalone vendor when production volume and licensing certainty matter more than raw model benchmarks.
Why it matters: A standalone creative AI vendor at this valuation puts a "Runway or Veo" call on every CMO's roadmap before the next campaign cycle. 📰 TechCrunch
More trending AI marketing news from last week
Gartner finds CMOs put 15.3% of budgets into AI; only 30% can scale.
AdExchanger argues first-party data will define the agentic ad era.
Higgsfield Supercomputer went live, running full creative pipelines from chat.
Recursive Superintelligence emerged with $650m to build self-improving AI.
What 2,000 SaaS Companies Reveal About Growth in 2026
Is your growth in-line with your peers in B2B SaaS & AI?
Benchmark yourself against actual billings data for Maxio’s 2000+ global customers, alongside firsthand company perspectives to understand how growth varied by company size, business model, and strategic focus.
Key takeaways from the report:
Average growth across 2,000 companies
Growth by revenue band
AI-led vs AI-enhanced. Who performed better?
Quote of the week
Marketing leaders are stuck spending more on AI than their organisations can absorb. Without operating-model investment, those budgets will not translate into measurable performance."
Trending AI tools for marketers
Here are 4 trending tools to explore this week. Quick, practical upgrades to your AI marketing arsenal:
Zen Reports: Attribution for traffic from AI assistants, slotted inside GA4.
Magic Patterns: Design-system-aware AI prototyping that ships real components.
Loops: Email built for modern SaaS with AI-assisted lifecycle workflows.
Sparkle: AI outbound agent with verification, warmup and pipeline tracking baked in.
TV, podcasts & streaming
Decoder with Nilay Patel: I am not a robot, with Joanna Stern
Patel sits down with Joanna Stern on the 11th May, the day before her new book launches, to unpack what AI is doing to daily life and where it leaves marketers, consumers and creators. Stern argues the next consumer battleground is "proof of humanity" inside every interface, from email replies to dating app messages to customer service chats. For marketers, the conversation lands on a real question: if every brand touchpoint can now be generated by an agent on either side, what signals does a customer actually trust before they buy? Watch it to it here.
AI training
How to launch an AI-personalised lifecycle email campaign using Loops
The overview:
Loops is a SaaS-focused email platform that lets marketing teams ship AI-personalised lifecycle campaigns in under 30 minutes using one prompt and your existing customer events.
Step-by-step:
Connect your event source Pipe product events into Loops via the API, Segment or a native integration. Pick events that signal real lifecycle moments such as signup, first value, or cancellation, because the cleaner the event, the cleaner the segmentation downstream.
Define the loop trigger Create a new loop and pick the event that fires the journey, for example "no login for 14 days". Keep the trigger narrow on the first send so you can attribute revenue cleanly, then widen it once you have a baseline number to compare against.
Prompt the AI builder Describe the email in one line, naming audience, offer and tone. Name the customer segment, the discount, the call to action and the brand voice rule in one sentence, because the more specific the prompt the closer the first draft.
Review and edit the generated send Check subject line, body and CTA against your brand voice before approval. Pay closest attention to the CTA copy, which is where AI defaults feel most generic and where small edits move click-through the most.
A/B test on a 20% holdout Send variant A and B to a small slice and let Loops route the winner. Pick one variable to test, usually subject line or first line, so the winning version teaches you something repeatable for the next prompt.
Monitor and iterate Track open, click and revenue per send, then tune the prompt for the next cycle. Save the prompt and its result in a shared doc so the team builds a prompt library instead of starting blank each week.
Pro tip: Save your winning loops as templates, then fork them for win-back, dunning and post-purchase flows so new campaigns ship in minutes.
The weekly deep dive
Why agencies are repositioning themselves as AI operating systems
A wave of holding companies and creative shops is pitching themselves as AI operating systems, not service businesses, and it is changing how brands buy and budget for marketing.
The pitch is shifting from headcount to infrastructure
Agencies are selling proprietary AI platforms with embedded data, models and workflows, then charging for outcomes the platform produces. Pricing is moving from retainer to consumption, with hours billed disappearing into model usage and per-asset fees. WPP Open Pro, Publicis CoreAI and Omnicom's agent stack are all variations of the same playbook, and the agencies that ship platforms first are setting the pricing reference for the rest of the market.
Data and IP become the moat
The agencies winning briefs are the ones with first-party data partnerships, proprietary creative analytics and vertical IP that no AI lab can show up with on day one. Generic strategy decks are losing in pitch, and creative networks are racing to acquire predictive analytics and audience-data businesses to close the gap. Publicis spent $970m of its 2026 deal budget on this kind of capability, which signals where the spend is moving and what the next twelve months of agency M&A will look like.
Procurement is changing the conversation
Brand teams are being asked to defend agency spend against in-house AI builds and lab-led services arms inside OpenAI and Anthropic. CMOs need a clear answer on what they are buying that prompt-and-tool budgets cannot replicate. The answer that survives procurement scrutiny is usually a mix of proprietary data, category expertise and creative judgement, and the agencies who can name those three lines on a pitch deck are the ones renewing in 2026.
Our takeaway: If your agency relationship is priced on hours and headcount, renegotiate before procurement does it for you on AI maths.
📖 Read the full article at Ad Age [paywall]
