A wolf in sheep’s clothing? Zuck just kicked the AI platforms off WhatsApp
Plus: Agentic commerce 'cat and mouse' just got real - and Amazon is not happy
Hi everyone, thanks for coming back to Customer Futures.
Each week I unpack the disruptive shifts around Empowerment Tech. Digital wallets, Personal AI and the future of the digital customer relationship.
If you haven’t yet signed up, why not subscribe:
Join us in person - Christmas Customer Futures Meetup!
Are you curious about Personal AI, digital ID and digital wallets? Or are you already working on Empowerment Tech?
Wednesday 10th December
Join us for the annual Customer Futures Christmas bash. We’ll be at the main bar, The Hoxton Hotel Holborn, from 6pm (here)
Looking forward to seeing you there!
Hi folks,
My goodness, it’s been a week.
Perplexity is now in a legal fight with Amazon over AI Agents and shopping. And Zuck just fired the AI platforms from WhatsApp.
Lots to dig into.
Meanwhile, if you’re going for a walk or a run, or just doing the washing up, why not catch up with the latest Customer Futures Conversations:
Episode 01: Making Sense of Digital ID in the UK WATCH
Episode 02: Empowering Care Workers with a Digital Passport WATCH
More episodes coming soon.
I can’t think of a more exciting time to be working on digital trust, customer data and digital customer relationships. Because we’re watching a bunch of real-time battles play out for the future of being a digital customer.
So welcome back to the Customer Futures newsletter.
In this week’s edition:
Agentic commerce ‘cat and mouse’ just got real - and Amazon is not happy
Is OBO the next big thing for AI?
A wolf in sheep’s clothing? Zuck just kicked the AI platforms off WhatsApp
… and much more
Grab a warming hot chocolate, a cosy spot on the sofa, and Let’s Go.
Agentic shopping ‘cat and mouse’ just got real - and Amazon is not happy
Well, we just got front row seats to the latest battle in agentic commerce: AI browsers and online shopping.
Because Amazon just filed a cease-and-desist lawsuit against Perplexity.
Apparently, it violates Amazon’s terms of service to use Perplexity’s new AI browser on their shopping marketplace. Specifically, using CometAI’s agentic features to search and make purchases.
They argue that it’s ‘computer fraud’. Hiding the AI agent’s activity as if it were a real human on the website. They say that CometAI “degrades the Amazon shopping experience” and introduces potential privacy risks.
“Rather than be transparent, Perplexity has purposely configured its CometAI software to not identify the Comet AI agent’s activities in the Amazon Store.”
Well, guess what. We’re back to AI agent identity again. Who knew. The latest fault line in the race to dominate Agentic Commerce.
Perplexity’s CEO has responded by saying that Amazon’s AI restrictions are anti-competitive. But then he fired a surprising shot back, saying:
“Agents should have the same rights and responsibilities as real users”
Wow. Did you see that coming?
And just to make the battle lines even clearer, Perplexity has now published a zinger of a retaliation post: “Bullying is not Innovation”.
I hope the Perplexity PR and policy team get a raise for that one.
They open with this:
“The point of technology is to make life better for people. We call it innovation, but it’s just the constant process of asking how to make things better.
“Bullying, on the other hand, is when large corporations use legal threats and intimidation to block innovation and make life worse for people.”
The knives are out.
But with a twist.
Because it now seems that Perplexity is going after ‘Personal AI’ that works directly for people, not just the corporate ‘Personalised AI’ version that serves the business, not the human.
Check this out:
“User agents are exactly that: agents of the user. They’re distinct from crawlers, scrapers, or bots. A user agent is your AI assistant—it has exactly the same permissions you have, works only at your specific request, and acts solely on your behalf.
They go on to share the three pillars of AI Assistants, and boy does it sound like a Customer Futures post.
Here’s the text in full (bold mine):
For user agents to serve their true purpose, they must be:
“1. Private.
“Your AI assistant must be indistinguishable from you. When Comet Assistant visits a website, it does so with your credentials, your permissions, and your rights. (It’s also unable to do anything you can’t).
“Publishers and corporations have no right to discriminate against users based on which AI they’ve chosen to represent them. Users must have the right to choose technologies that represent them. Privacy and freedom of choice depend on this.
“2. Personal.
“Your user agent works for you, not for Perplexity, and certainly not for Amazon. For decades, machine learning and algorithms have been weapons in the hands of large corporations, deployed to serve ads and manipulate what you see, experience, and purchase.
“The transformative promise of LLMs is that they put power back in the hands of people. Agentic AI marks a meaningful shift: users can finally regain control of their online experiences.
“3. Powerful.
Your AI assistant must be capable of any task that matters to you. Users have a right to select high-performing AI agents from the cutting edge of innovation. The technology available to users can’t be hamstrung just because it threatens some public company’s pressure to deliver more ad revenue.
“The future of AI, like all technology, is for people.”
Woof. Important statements from an important company at an important time.
“The rise of agentic AI presents a choice. Will this technology empower users to take control of their digital lives? Or will it become another tool for corporations to manipulate and exploit?”
I’m kinda stumped.
And quietly delighted.
Let’s now look over at the other gorilla in the room: Anthropic.
Their Chief Product Officer is Mike Krieger, previously co-founder and CTO of Instagram. On a panel earlier this year, he was asked about Anthropic’s competitive moat. When there are so many new releases of other faster, cheaper and more powerful open source LLMs.
His answer?
Personalisation. “We’ll know the user better than the other platforms,” he said.
Notice that it’s a completely different position from Perplexity. Who argue that Personal AIs should have the same rights and privileges as the user. That AI agents should act with and for the customer.
Not for Amazon. And not - remarkably - for Perplexity.
So forget ChatGPT integrations with wallet providers and shopping companies. Forget the agentic payments jazz hands. This is the new AI battle. Where AI agents start showing up in existing customer channels.
This Perplexity + Amazon legal battle is a first skirmish between AI Cat and Retailer Mouse.
But make no mistake. It’s also now the new battleground for Agentic Commerce, personal data, and ultimately the new digital customer relationship.
AMAZON LAW SUIT, PERPLEXITY PR ON BULLYING
Is OBO the next big thing for AI?
Agents are naturally decentralised things. And therefore, so should be the tools and infrastructure we use to:
Identify our AI agents
Give them permissions and scope
I’ve long believed the answer will be a new category of services: OBO
‘On behalf of’
We already do OBO for people: we call it ‘delegated authority’.
But today it only works at human scale. When people only change roles every few years. And for temporary assignments that might only last a few weeks.
What happens when an AI agent is spun up for a specific task that takes only 8 minutes, then is removed or deleted?
We are going to need a new approach to handle delegation for AI at scale. For dynamic authority. Something flexible enough that can handle any number of people, organisations and things - including AI agents.
We need ‘OBO’.
It’s a term that was born in the Microsoft and OpenID world. But it’s ready for primetime. For wider use. And to represent all sorts of cryptographically provable objects.
Yes, OpenID tokens. But also verifiable credentials. And UCANs. And ZCAPs. And I’m sure others to come.
For an employee’s OBO, it could be given to someone stepping into a new role temporarily. Proof that they can represent their organisation, proof of their job and proof of their authority e.g. to sign documents.
And for an AI agent, an OBO credential could be issued to the AI platform for a short procurement process. To onboard a supplier, for example. Proof of the agent’s owner (the organisation), the agent’s permissions (onboarding businesses), and their scope of activity (setting up new vendors on the system).
We are more digital, more remote, and more automated than ever. OBO is an idea whose time has come.
I’ll need OBO. My business will need OBO. And my AI agent is certainly going to need OBO. Because AI Agents won’t be let out of the building without their proof of authority.
But why does OBO need to be decentralised? Because a centralised OBO approach simply can’t scale or be trusted across businesses or geographies.
Yes, there will be a ton of centralised agents - including Zuckerburg’s planned ‘superintelligence’ for us all in WhatsApp (more on that in the next piece below). But with OBO, I’m making a more general point about the wider agentic marketplace. And the coming ‘A2A economy’.
Brands will have agents. People will have them. Governments will have them. Devices will have them. So there won’t and can’t be ‘one master agent platform’, even if you include the BigTech providers.
And remember, digital delegated authority is going to be especially challenging, because it’ll need to work across chains of multiple agents, chains of multiple organisations, and represent people and businesses across multiple jurisdictions.
A good example is the traveller use case.
Yes, you can have a simple ‘book my hotel’ agent run centrally by Booking.com and ChatGPT.
But the rest of the trip - the check-in, the travel disruption, the loyalty, the receipts, the upgrades, the tours, the entertainment, all of it - can’t be run by one company.
So even in a simple travel scenario, and even in only the traveller ‘happy path’, we might still need to deal with several agent platforms.
There’s lots more to come on OBO for agents.
And it’s not only going to become a new type of ‘delegated authority’ for AI. I’m betting that it’ll become a massive new market category for AI agents, a bit like ‘KYC’ or ‘Power Of Attorney’ is for people.
OBO might just be the next big thing for AI.
A wolf in sheep’s clothing? Zuck just kicked the AI platforms off WhatsApp
Well, there we have it.
Meta is using ‘platform strain’ as the excuse to kick all the other AI platforms off WhatsApp.
At first glance, it’s just another ‘walled garden’ play. But there are some undercurrents here you need to be paying attention to.
First, we can be pretty certain Zuck is planning to give everyone their own AI agent. He wrote a manifesto about it a few months ago, called ‘Superintelligence’.
When he already has billions of daily active users across WhatsApp, Insta and FB, it makes complete sense. It’ll be another smart move, to meet people where they already are. Not force us to download yet another app, browser extension, or switch over to another ‘AI browser’.
Second, we should note that Zuckerberg has already said publicly that he’s betting on ‘WhatsApp for Business’ becoming a main revenue driver globally.
Imagine telling every brand on the planet that they can not only have their own powerful AI customer chatbot, offering highly personalised customer experiences and unlocking a smart new customer service channel…
…but that WhatsApp users now have a new ‘superintelligent’ AI that can DO things for them:
Rebook an appointment
Manage an order
Renew a subscription
Return a product
Upgrade a package
And yes… make payments and check out right from within the WhatsApp chat.
Zoom out, and this fast becomes a dream go-to-market strategy for Zuck AI:
1. Billions of existing users
2. …Connecting their existing product accounts (“please give us your account number…” followed by “we’ve just sent you a one-time password”)
3. …Using their existing daily chat interface that they already use 243 times a day, to interact with businesses to get stuff done using Meta’s AI agents
I expect Zuck will simply wait for the right time to roll it out. Perhaps when AI agent liability gets clarified in the market.
And then he’ll get enterprises to come running, to connect all their business AI platforms - which on the face of it, are his competitors - to his existing customer base. Then watch the money roll in.
It’s why he won’t want all those AI platforms inside WhatsApp. Better to kick them off. Then he can get businesses to connect their AIs to his users, and on his terms. Let enterprises do the hard work inside the business, and he’ll just walk over with a billion DAUs and say ‘let’s connect’.
But third - and this is the big one - this becomes the mother of all data grabs.
Because he won’t just charge businesses for connecting their AI services to WhatsApp. And he won’t just make WhatsApp even more sticky, and harder for users and businesses to leave.
He’ll get access to Every. Single. Customer. Conversation.
All that data.
How?
Because he just updated the Terms of Conditions in ‘WhatsApp for Business’.
Giving him the right to access ALL data across ALL customer conversations via the channel.
WhatsApp for people will still be private and ‘end-to-end encrypted’. But its big sister, ‘WhatsApp for Business’, is now fair game. Non-private, and now Meta’s new data goldmine to go digging in.
So just to recap:
Zuck is planning to give us all our own Personalised AI - ‘super intelligence’
He’s doubling down on WhatsApp for Business
He’s changed the T&Cs so that he gets all the data shared between customers and businesses using that channel
And he’s now just kicked all the other AI platforms off WhatsApp
Looks like a pretty coherent strategy to me.
A (Data Grab) Wolf in (Customer Service) Sheep’s clothing.
So let’s not worry about AGI folks. Let’s worry about Zuckerberg building the most accessible, convenient and powerful customer services platform we’ve ever seen.
And while we all do a victory lap on AI token performance, health discoveries and co-pilots, let’s keep a close eye on where control is shifting.
It’s going to be awfully difficult to change the taxi service once we’ve given Zuck the keys to lock us into the car.
ZUCK’S SUPERINTELLIGENCE, THE REAL WHATSAPP STRATEGY
OTHER THINGS
There are far too many interesting and important Customer Futures things to include this week.
So here are some more links to chew on:
Post: AI Agent Identity is not What You Need READ
Dataset: The latest World Bank ID4D Findex survey on adoption and use of digital ID apps around the world DOWNLOAD THE DATA YOU WANT
Paper: What does “continuous and real-time” data portability actually mean? READ
Post: Jobsites don’t need your certificates, your passport, your visa - they just want assurance that you’re qualified READ
Idea: The imperative: De‐Anthropomorphising an AI Experience READ
And that’s a wrap. Stay tuned for more Customer Futures soon, both here and over at LinkedIn.
And if you’re not yet signed up, why not subscribe: