Personal AI Could Break Customer Loyalty Built on Friction
Meta’s Muse has been downloaded more than 5 million times. Personal AI agents could turn customer loyalty into an outcome brands must earn, forcing a rethink of identity and authority.

Why it matters
- Meta’s Muse personal AI agent has been downloaded more than 5 million times.
- Personal AI objectives include cutting monthly expenses by 10% and keeping travel within budget.
- The commentary’s example authority boundary allows flight rebooking up to $500.
- The Personal Agent Protocol aims to standardize agent identity and delegated authority across businesses.
- Mastercard and Visa are developing infrastructure for secure AI-initiated transactions.
By Rebecca Stone
Meta’s Muse personal AI agent has been downloaded more than 5 million times, according to the commentary “Personal AI Is Coming for Customer Loyalty and Brands Aren’t Ready.”
The commentary reports that OpenAI has introduced Dots, an always-on agent designed to work toward consumers’ goals in the background. It adds that Google and others are racing toward the same model.
The commentary’s author writes: “Much of what brands call loyalty is really inertia that personal AI will expose.” The argument matters for carriers, banks, insurers and subscription businesses whose retention depends on customer effort.
Consumers often remain because switching requires research, comparing complicated offers, negotiation, forms and a judgment about whether a small saving justifies the hassle, the author says. Personal AI could turn that labor into an instruction.
A consumer could authorize an agent to “Cut my monthly expenses by 10%, keep my travel within budget or ensure I’m always getting the best deal,” the commentary says. An agent that continuously checks whether the customer reaches the best outcome would make inertia visible.
The author says their company is working with Meta, Sierra and others on the recently announced Personal Agent Protocol. The project aims to create a common way for personal AI agents to identify whom they represent and what they may do when they interact with businesses.
The commentary treats the protocol effort as evidence that personal AI is moving from consumer experimentation toward a requirement for business infrastructure. Millions of agents could arrive at front doors that assume a human customer and human authentication. The author expects this collision to reshape how experiences are designed, relationships are built and value is created.
What changes when an agent becomes the customer interface?
Personal AI inserts a new layer between a business and the person it wants to acquire and keep. The consumer may never see the redesigned website, navigate checkout or consider a retention offer designed to keep the customer.
Instead, the consumer states an objective, and the agent determines how to pursue it. The agent becomes the interface to the brand. The business must understand that intent, work with the customer’s AI and orchestrate its own people, AI and systems to deliver the outcome.
The measurement changes with it. The commentary says asynchronous interactions shift the metric from containment to intent-to-outcome: whether the customer achieved the goal and how much effort that took.
Brands need answers to a new operating checklist:
- Who or what is interacting with us?
- Who does it represent?
- What is the customer trying to accomplish?
- What has the AI been authorized to know and do?
- When should we engage the customer rather than their AI?
Why does identity alone not establish authority?
Many companies cannot reliably distinguish an authorized AI acting for a customer from the customer itself, the author says. Identity does not settle the second question: what the agent may access, decide and do.
The commentary proposes these consumer boundaries: “Negotiate my bill, but don’t cancel my service. Rebook my flight, but don’t spend more than $500.”
The author compares that model to a digital power of attorney for AI, built on identity, authentication and delegated authority. Permissions need to travel with the agent. Actions need to be traceable so consumers and businesses know what happened and under whose authority.
Control returns to the human when the agent reaches a permission boundary or faces a decision it lacks authorization to make. That rule matters across businesses because each company cannot solve delegated authority in isolation.
Standards such as the Personal Agent Protocol therefore carry weight beyond one vendor’s infrastructure. The commentary also says Mastercard and Visa are developing infrastructure for secure AI-initiated transactions. Without a shared trust layer, the author warns, personal AI could act everywhere while remaining authorized almost nowhere.
Should AI keep browsing websites built for people?
Personal AI enters business through interfaces built for humans. The commentary says agents browse websites, fill out forms, navigate chat and can even make phone calls.
Enterprises need an agent-native front door, according to the author: a secure route for authorized personal AI and enterprise AI to communicate directly, establish identity and authority, exchange context and coordinate action.
Open standards such as A2A can help agents communicate and collaborate across different vendors and architectures. The alternative, in the author’s view, is forcing AI to navigate an interface designed for people instead of allowing agents to coordinate around the customer’s intent.
The canceled-flight example shows the proposed path. A traveler’s personal AI contacts the airline’s AI, which evaluates alternatives against the traveler’s preferences. An exception brings an AI-assisted employee into the exchange. The traveler returns only when approval is required.
The sequence runs from personal AI to enterprise AI, from AI to employee, from employee to customer and potentially back again. The experience no longer belongs to one channel or participant, the commentary says. Businesses must preserve identity, intent, context and authority as it moves.
What will orchestration have to do with loyalty?
Customer experience has moved beyond human-to-human interaction, the author says. People, enterprise AI and personal AI may now work toward one outcome. Interoperability alone cannot make that exchange trustworthy.
The author says orchestration has to connect the right resources and coordinate action around what the customer wants to accomplish. It becomes essential for a journey that can cross companies, AI systems and people.
When personal AI makes comparison, negotiation and switching easy, businesses will no longer be able to rely on inertia, the commentary argues. They will have to earn loyalty through each outcome instead.
The Personal Agent Protocol and the Mastercard and Visa payments work now underway will test whether personal AI can operate across businesses with delegated authority intact.
Source: unite.ai
More from Rebecca Stone
Show full bio
Correspondent covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.
728 articles
Related articles
- Meta's Muse Agent Takes Aim at the Subscription Economy's Debt to Forgetting
- JPMorgan, Amazon and Kohl's Are Blocking AI Agents. Experts Say They Can't Win That Fight
- OpenAI's $100-a-Month Dots Agent Misheard Me, Said It Loved Me
- Meta Claims Millions Shop With Muse. Executives See 'Hundreds' of Orders
- OpenAI Dot vs. Meta Muse: the always-on AI agent showdown