did LinkedIn just quietly win ai search?
Plus: AI deepfakes went industrial for beauty influencers, customer reviews feed algorithms, and Anthropic launched an institute.
what you need to know this week in ai + marketing in 15 seconds:
AI deepfakes are stealing beauty influencers' glow - creators dealing with AI-generated lookalikes and stolen aesthetics. Beauty was always vulnerable to copycats, AI just turned that into a volume business (Morning Brew)
LinkedIn is the most-cited domain for professional queries in AI search - matters for B2B marketers. If your LinkedIn presence is sleepy, your AI search presence probably is too (ProFound)
Customer reviews become battleground as AI revolutionizes product discovery - reviews feed the model’s understanding of what a product is and who it’s for. Reviews used to influence humans, now they influence the machine that influences the human (Modern Retail)
Marketing needs a decision infrastructure for AI - not just prompts and policies, but an operating system: who approves what, what gets automated, what gets escalated (MarTech)
Anthropic launched the Anthropic Institute - focused on economic and social questions around advanced AI. Will study long-term effects on labor, markets, public life (Anthropic)
Claude Marketplace launched - enterprise app store where orgs use existing Anthropic spend to purchase Claude-powered partner tools. GitLab, Snowflake, Replit, Harvey, Rogo (Anthropic)
Consumers still prefer talking to people in customer service - especially for complicated or emotional issues. “Human-in-the-loop” remains undefeated (CX Dive)
🗞️ top ai + marketing news
labor, layoffs, and the cost story
Meta planning sweeping layoffs as AI costs mount
Reuters | ~5-min read
Reuters reports Meta is considering cuts that could hit 20% or more of staff as it tries to offset the cost of its AI buildout, including huge data center spending and expensive recruiting for top researchers. Executives have also pointed to efficiency gains from AI-assisted work, which gives the whole thing a familiar tone.
My take: We’re now firmly in the era where AI is both the investment story and the layoff story. Convenient…
Consumers prefer talking to people in customer service
CX Dive | ~3-min read
Despite all the AI rollout, most consumers still say they’d rather deal with a human for customer service, especially when the issue is complicated or emotional. AI may be fine for the quick stuff, but people still want a person when the stakes feel even mildly annoying.
My take: “Human-in-the-loop” remains undefeated.
the stack, the org chart, and the boring stuff that matters
AI is repricing the marketing stack, not collapsing it
MarTech | ~5-min read
This piece argues AI is not wiping out the stack so much as changing what each layer is worth. Execution gets cheaper, orchestration gets more important, and the center of gravity moves toward data, workflow, and decision quality.
My take: The stack is not disappearing. It’s just becoming less forgiving of useless software.
Why marketing needs a decision infrastructure for AI
MarTech | ~5-min read
The argument here is that AI does not just need prompts and policies. It needs a real operating system for decisions: who approves what, what gets automated, what gets escalated, and how tradeoffs get made. Otherwise teams get speed in random places and chaos everywhere else.
My take: Most teams do not need more AI strategy decks. They need an answer to “who decides?”
5 design skills to sharpen in the AI era
Figma | ~6-min read
Figma’s view is that the designer’s value shifts toward judgment, systems thinking, prototyping, and working well across functions as AI absorbs more of the mechanical work. It is a design article, but the point applies to marketing too.
My take: The jobs are not getting less creative. They are getting less tolerant of people who only know one lane.
Top resources for Claude Code
Notion | ~4-min skim
This is a practical roundup of Claude Code resources for people trying to use it for workflows, experiments, and real work. Useful if you are still in the “what do I even do with this” phase.
My take: Resource lists are not glamorous, but they save people from three weeks of fake productivity.
discovery, search, and the new shelf space
LinkedIn is the most-cited domain for professional queries in AI search
ProFound | ~4-min read
ProFound found that LinkedIn is showing up more than any other domain for professional questions in AI search results. For B2B marketers, that matters. A lot. It means the content layer around your company, your executives, and your category is increasingly being pulled from a platform many brands still treat like a posting obligation.
My take: If your LinkedIn presence is sleepy, your AI search presence probably is too.
AI mode results are personalized to user behavior
Search Engine Roundtable | ~3-min read
Google’s AI mode results appear to be shaped by user behavior, which pushes search even further toward personalization. That makes discovery less universal, less stable, and a lot harder to “rank” for in the old-school sense.
My take: Search keeps becoming more like a feed, which is bad news if your whole strategy depends on one clean result page.
Customer reviews become a key battleground as AI revolutionizes product discovery
Modern Retail | ~4-min read
As AI systems increasingly summarize products for shoppers, reviews matter more because they feed the model’s understanding of what a product actually is, whether it works, and who it’s for. Reviews used to influence humans. Now they also influence the machine that influences the human.
My take: The humble review is having a very strong year.
brands, content, and the trust problem
An AI avatar of Andy Cohen will start name-dropping brands on Peacock this summer
Marketing Brew | ~3-min read
Peacock is using an AI avatar of Andy Cohen for branded, vertical-video-style engagement on mobile. It’s one more example of media companies trying to make AI-driven content feel familiar by wrapping it in a known personality.
My take: We are getting dangerously close to “celebrity QVC, but synthetic.”
AI deepfakes are stealing beauty influencers’ glow
Morning Brew | ~4-min read
Beauty creators are dealing with AI-generated lookalikes and stolen aesthetics, which is creating a new layer of identity theft in an industry already built on image, trust, and imitation.
My take: Beauty was always vulnerable to copycats. AI just turned that into a volume business.
An AI doppelganger stole my voice
The New York Times | ~6-min read
This opinion piece is about the creeping weirdness of AI replicas and voice imitation, and how unsettling it is when your style, tone, or likeness gets reproduced without your consent. The article taps into a broader feeling that AI tools are starting to mess with authorship in a way that is less theoretical and more personal.
My take: The deepfake problem gets more disturbing the closer it gets to ordinary people.
creative moving upstream
As AI creative moves upstream, one production firm is pitching brands a model built on that trend
Digiday | ~5-min read
Digiday looks at a production company pitching brands on using AI earlier in the creative process, not just in post or versioning. In other words: AI is moving from “help us make more assets” to “help us shape the idea.”
My take: This is where people get nervous, for good reason. Upstream is where taste lives.
the capital race and the “who gets to shape the rules” phase
Introducing the Anthropic Institute
Anthropic | ~3-min read
Anthropic launched a new institute focused on the economic and social questions around advanced AI. The company says it will study AI’s long-term effects on labor, markets, and public life.
My take: Every major AI company eventually starts building the think tank version of itself.
The Top 100 Gen AI Consumer Apps, 6th edition
a16z | ~8-min skim
The latest a16z ranking tracks which consumer AI apps are actually getting used, not just talked about. The useful bit is not the list itself. It is the reminder that daily utility is separating the sticky products from the novelty products.
My take: Plenty of AI products are impressive. Fewer are part of anyone’s Tuesday.
👩💻 thought-leader highlights
Richard King – “Mastering Claude for Product Marketing”
Richard argues Claude is no longer just a chatbot but an ecosystem; to get real value product marketers must learn how Projects, Cowork, Skills, Connectors and Code work together. He promises concrete workflows for creating positioning docs, battle cards and launch briefs rather than generic prompting advice.
Ann Wehren – “Building with Claude Code: Start with Daily Tasks”
Ann suggests auditing two weeks of your calendar, listing every task and ranking them by impact and enjoyment to see where Claude Code can help. She uses it for inbox triage, calendar invites and drafting proposals and plans to build a personal CRM and family OS next.
Allie K. Miller – “Awesome Non‑coding Claude Code Loop Ideas”
Allie shares a toolkit of loops: monitor email every 15 minutes for specific projects, prep meeting summaries, watch deal threads for legal issues, track competitor announcements or viral posts and scan Slack for blockers. She notes recruiters, teachers and event planners can all use these proactive monitors to spot important actions instead of manually checking.
🛠️ latest ai + marketing tools
Claude Inline Visuals Anthropic rolled out interactive charts, diagrams, and visualizations that render directly inside Claude’s chat — not in a side panel.
Claude Marketplace Anthropic launched an enterprise app store where organizations use existing Anthropic spend commitments to purchase Claude-powered partner tools. Launch partners: GitLab, Snowflake, Replit, Lovable, Harvey, and Rogo. One contract, consolidated billing. Currently in limited preview.
Microsoft Copilot Cowork New autonomous execution layer for Microsoft 365 Copilot. Describe an outcome — “prep me for Thursday’s customer meeting” — and it builds a plan, pulls context from emails/meetings/files, creates deliverables, and runs in the background with checkpoints. Built in partnership with Anthropic’s Claude. Research Preview now, broader rollout late March.
Google Gemini Workspace Upgrades Major Gemini updates across Docs, Sheets, Slides, and Drive. Docs gets “Match writing style” and file-grounded drafting. Sheets builds entire spreadsheets from prompts and auto-populates tables with live web data (”Fill with Gemini”). Slides generates themed slides from context. Drive adds “Ask Gemini” — AI Overviews that answer questions across all your files and emails. Beta for AI Ultra and Pro subscribers.
Perplexity Personal Computer Software that turns a Mac mini into a 24/7 AI agent with full local file and app access, orchestrating 20+ frontier models. Works continuously as a digital proxy — controllable from any device, with audit trails and a kill switch.
Webflow Acquires Vidoso Webflow bought four-person AI content-generation platform Vidoso, which turns keynotes and panels into short clips, blog posts, social assets, and presentations — all brand-governed.
💼 cool ai + marketing jobs
Anthropic - Community Marketing Lead
Lovable - Senior Solutions Marketer
Canva - AEO & Organic Growth Lead
Qualcomm - Director, AI Product and Technology Marketing Lead
Snap Inc - AI Community Program Manager
Hit reply anytime and let me know:
Topics you’d like covered in future weeks
Sections you love or could do without
Questions about AI, marketing, or anything else
I read every reply and use them to shape what I write.
Love y’all,
Carley
Great issue as always!
Re: the Anthropic Institute -- can any company-owned 'think tank' be truly neutral? I've worked before on think tank-esque teams in tech, and in my experience companies are able to recruit top researchers because that's the only way to access critical data (also.... stock options). But then we end up with a world where the leading research on societal impact and ethics is owned and shaped by the companies.