Your Brand’s Dumbest Employee Is Now AI

Almost all AI adoption leads back to the same question: what to keep and what to slash on the altar of efficiency. And nearly all AI adoption has the same problem: people keep things that should be cut and slash things that should never be slashed. And nowhere else is this more visible than in customer service.

Of course, it is visible in CS. It is the frontline of our brands. People make buying decisions just based on how CS performs. Any award-winning marketing campaign can tank in numbers if your CS has a bad rep.

This week, we bring you news (among others) on how others are approaching this problem.

Within the fine details, a clear path emerges for everyone to follow to succeed in building better services with AI.

The future of service should stay human. But AI should put superpowers in every CS agent’s hands.

— Peter & Torsten

Before we jump in, a word from this week’s sponsor:

Stop wasting time on repetitive explanations. Guidde’s AI creates stunning video guides in seconds—11x faster.

How it works: Click capture on the browser extension, and Guidde auto-generates step-by-step video guides with visuals, voiceover, and a call to action.

Best part? It’s 100% free.

👉 Download the extension now

The rush to put AI in front of customers is creating some spectacular, public-facing blunders. Taco Bell, for instance, is reportedly having second thoughts about its AI drive-thru after videos of it comically messing up orders went viral (like here). It’s a classic case of a tool that worked in the lab but couldn't handle the noise and accents of the real world.

Meanwhile, clever users are running circles around poorly secured chatbots. A developer on X showed how he tricked the United Airlines support bot into spilling its internal programming and connecting him to a human just by using a simple "prompt injection" technique. It’s the digital equivalent of asking the new social media intern for the CEO's phone number.

These aren't one-off goofs but symptoms of a bigger problem. Companies are deploying AI without the rigorous testing—or “red teaming”—needed to make them secure and reliable. Without it, you’re basically giving a half-trained system a direct line to your customers, and the whole world is watching.

Your customer-facing AI is a direct reflection of your brand's competence. Every failure is public and amplified on social media. This isn't an IT problem. It's a CMO's problem. If your AI agent is easily broken, frustrating to use, or provides incorrect information, the damage to brand trust can far outweigh any cost savings.

The lesson is that the budget for red-teaming and continuous monitoring must be considered a core part of the AI implementation cost, not an optional add-on.

Vivrelle, a luxury club for borrowing designer handbags, just launched an AI personal stylist named Ella (try it here). Instead of just showing users a grid of products, Ella acts like a helpful expert, giving personalized advice to help members choose the right accessory for an outfit or event. It's a smart play to boost both rentals and sales.

Ella works by having a natural conversation with the user to figure out their style and what they need. It's a great example of using AI to add value and create a premium experience, not just to cut costs or deflect support tickets. It makes the service feel more personal and high-touch.

This is an AI strategy you should consider copying. Vivrelle is using AI offensively, to drive revenue and delight customers, not just defensively to save money. For any business with a big product catalog, an “AI expert” can be a powerful tool to help customers navigate their options, feel understood, and make a decision faster.

Apple is rolling out a new AI chatbot for its own retail and support employees. The tool acts like an internal search engine, helping staff get quick, accurate answers about product troubleshooting, repair policies, and other company knowledge.

This is classic Apple: cautious, deliberate, and focused on quality control. Instead of rushing out a consumer-facing product that could go off the rails, they're using AI to improve their own operations first. It's a lower-risk way to test their models and get a real business benefit at the same time.

A better-informed employee provides a better customer experience. The AI isn't replacing the Genius Bar. It's making the Genius smarter and faster. This ensures that the AI's impact aligns perfectly with Apple's premium brand image.

This is the “walk before you run” playbook for enterprise AI. Before you aim an AI at your customers, aim it at your own team. An internal AI expert is one of the fastest ways to get a return on your investment. It makes your sales and support teams more consistent, reduces training time, and eliminates the need for senior staff to answer repetitive questions.

Amazon just rolled out Lens Live, an AI tool in its app that lets you buy things just by pointing your phone’s camera at them. The feature uses computer vision to identify a product in real time and slaps a virtual “Add to Cart” button right on your screen.

It's augmented reality, but for shopping. This isn't just a souped-up version of their old image search. It removes all the friction. You no longer have to snap a picture and wait for results. You just point, and Amazon tells you what it is and how to buy it, instantly.

It’s the shortest possible path from seeing something you want to owning it. Lens Live also connects to Rufus, Amazon's shopping assistant, so you can ask questions about the object you're looking at. You can point at a pair of sneakers and ask, “Do these come in black?” or “What are the reviews?” It's like having a personal shopper in your pocket who can see what you see.

The line between the physical and digital shelf is vanishing. Every product—on a store shelf, in a magazine, at a friend’s house—is now a direct link to a sales funnel. This means your product's visual design is more important than ever. If Amazon's AI can't recognize your product, you don't exist in that moment of impulse.

Salesforce CEO Marc Benioff just pulled back the curtain on one of the biggest AI wins we’ve seen. The company is now handling 50% of its customer support with AI, which has resulted in a reduction of the support team by 4,000 people. It’s a massive efficiency gain that proves the business case for AI in high-volume service operations.

But the real story isn't just about cost savings. Benioff revealed that Salesforce had a backlog of over 100 million leads from the last 26 years that they simply never had the manpower to call back. Now, an “agentic sales” system is working its way through that list, turning a dusty old database into a fresh pipeline.

This move reframes AI from a simple replacement tool to an engine for tackling operational debt. It's about finally doing the work that was always too expensive or time-consuming for humans to handle. By automating the routine, Salesforce can focus its human talent on higher-value work and unlock revenue that was previously written off.

This is the slide you bring to your next board meeting. The story isn't “we can cut headcount.” It's “we can monetize our biggest untapped asset: our own data.” Every business has a version of this 100-million-lead problem sitting in its CRM. This is your green light to frame AI projects as direct P&L drivers that expand your reach without scaling your team.

Stripe and crypto firm Paradigm just launched Tempo, a new payment network built so AI agents can pay each other with stablecoins (essentially digital dollars). The project is backed by a who's who of tech, including OpenAI, Anthropic, Shopify, DoorDash, and Visa. This isn't an experiment; it's the foundation for a new machine economy. Tempo is designed to let autonomous AI agents hold their own money and pay for things without a human clicking “confirm.”

Imagine an AI that re-stocks your office supplies, books travel, or pays for API access all on its own. This requires a new kind of financial plumbing—one that's fast, cheap, and programmable. By getting the major AI labs (OpenAI, Anthropic) and major commerce platforms (Shopify, DoorDash) on board from day one, Tempo is being built directly into the ecosystem where it will be used. They're not waiting for agentic AI to arrive; they're building its wallet.

While we’re busy teaching AI to write emails, the smartest people in tech are teaching it how to spend money. Soon, your customers won't just be humans; they'll be bots working on behalf of humans. If your products and services aren't available via an API, these new machine customers won't be able to find or buy them. You'll be invisible.

Small Headlines. Big Shifts.

» Warner Bros. Sues Midjourney.

The movie studio is suing the AI art generator for copyright infringement, claiming it was trained on their films without permission. Get the popcorn ready; the outcome of this lawsuit will help write the rules for creative AI for the next decade.

» ElevenLabs Launches AI Sound Effects,

The AI voice company just dropped a new version of its SFX tool, which can generate any sound effect you can imagine from a text prompt. It's a huge time and money saver for creative teams working on podcasts, social videos, or ads.

» Kimi Gets a Creative Writing Upgrade.

Moonshot AI has updated its Kimi model, one of our favorites for long-form content creation, thanks to its unique writing style. Worth giving it a try.

» Satya Nadella Shares His Favorite Prompts.

The CEO of Microsoft revealed the five AI prompts he uses to run his day, from summarizing documents to prepping for meetings. The big takeaway? The AI is only as smart as the data you give it.

» WordPress Teases "Telex" AI Builder.

The popular website platform showed off an experimental tool called Telex that builds sections of a webpage from a simple text prompt.

AI adoption is a challenge on every level, not just in marketing. Decisions on what to keep and what to cut can make a huge difference in net positive budgets and ROI, and can even drive brands into a downward spiral.

People making these decisions need support and help. That’s what we’re building inside the AI Ready CMO community.

Apply for your membership: join.aireadycmo.com

See you next week,
Peter & Torsten

PS: Know someone who would benefit from this newsletter? Refer them and get rewarded: