Silicon Valley’s AI Personal Assistant Gold Rush Is On

A 23-year-old founder. A startup with just 14 employees. A personal AI assistant that can book dinner, change flights and order groceries.

And a $10 billion valuation.

On September 28, 2026, Instinct announced a $1 billion funding round led by Sequoia Capital, Benchmark and Coatue, bringing its valuation to $10 billion. Just a month earlier, investors had valued the company at $2.5 billion.

Its invitation-only AI assistant had launched in August.

That kind of trajectory tells you something about where Silicon Valley believes the next big opportunity lies.

For the past few years, technology companies have competed to build increasingly powerful AI models. ChatGPT, Claude, Gemini and others have changed how we research, write, code and analyze information.

Now the competition has shifted toward AI that can take action.

Personal AI assistants are beginning to manage calendars, monitor email, make phone calls, book appointments and complete complex requests across applications. Some work in the background without waiting for instructions.

The ambition goes much further than replacing a virtual assistant. These companies want to become the primary interface through which people manage their digital lives.

I’ve been exploring several products competing for that role, and my current obsession is Instinct.

The experience is refreshingly simple. No new platform to learn or dashboard to manage. I just text or email my assistant, much as I would another person. It knows the context, understands the request and gets to work.

I’m also using Catch AI for executive administration, Meta’s Muse is becoming an increasingly attractive option for personal tasks, and I’ve just assembled a small team of specialized Grok Bots. I’m pretty impressed with what I’ve seen so far.

Then there’s Anthropic’s new Claude Duet, which could become one of the strongest contenders for professional AI assistance.

The pace of development is remarkable. The competitive stakes are even bigger.

Mark Zuckerberg is spending billions to compete

Meta’s aggressive investment in AI provides some perspective on what the largest technology companies believe is at stake.

In June 2025, Meta invested approximately $14.3 billion in Scale AI, acquiring a 49% stake and bringing its 28-year-old founder, Alexandr Wang, into the company to help lead its superintelligence efforts.

Zuckerberg followed with an expensive recruitment campaign, hiring researchers and engineers from competing AI companies. Former OpenAI researcher Shengjia Zhao joined as chief scientist.

Meta needed to strengthen its AI capabilities after facing criticism over the performance of its earlier models.

The company introduced Muse Spark in April 2026, followed by its personal AI agent, Muse, in September.

Muse attracted millions of downloads within weeks. It allows users to delegate tasks through an application or WhatsApp, connecting with email, calendars and other services.

The advantage Meta brings to this competition is distribution.

Billions of consumers already use WhatsApp, Instagram, Facebook and Messenger. Meta can introduce autonomous AI capabilities through products people already have on their phones.

For smaller startups, competing with that reach presents an enormous challenge.

But Meta has competition from companies with significant AI expertise and existing customer relationships of their own.

Elon Musk is building something different: An entire team of AI agents

Elon Musk’s xAI is another company I wouldn’t underestimate.

Its Grok Bot platform takes an interesting approach to autonomous assistance. Rather than creating one assistant responsible for everything, users can assemble teams of specialized AI agents that work together.

Each bot has its own identity, memory, responsibilities and access to a persistent computer. It can navigate websites, use connected applications, create documents and execute workflows.

The bots can also communicate with one another.

Imagine assigning one agent to market research, another to content development and a third to reviewing the work before publication. They can exchange information, coordinate assignments and bring you in when they need a decision.

In August, xAI launched Grok Bot with the ability to work across applications and execute recurring tasks. Its November 2 release of Grok Bot 1.0 added stronger collaboration, more proactive task management and improved browser capabilities.

The latest version also allows bots to participate in Slack, Microsoft Teams and video-call chats. Teams can share specialized agents with access to relevant company information.

This moves the concept closer to a virtual team with defined responsibilities.

I’ve just created my own small team of expert Grok Bots, each with a particular role, and I’m impressed with the experience so far.

The ability to give agents specialized responsibilities and have them collaborate interests me, particularly for business activities involving research, content, marketing and project coordination.

I’m still evaluating how well they work independently and how much supervision they need. But there’s something compelling about assigning work to a group of specialists rather than managing every step yourself.

Given Musk’s resources, xAI’s technical ambitions and the company’s growing product ecosystem, I can’t imagine Grok Bot remaining a minor contender in this market.

Anthropic’s Claude Duet could become the one to beat

On October 21, Anthropic introduced Claude Duet, a proactive AI assistant designed for professionals and small business owners.

Duet runs on Claude Opus 5.5 and connects to more than 6,000 applications and tools through Anthropic’s marketplace.

Its capabilities extend from scheduling meetings and managing email to preparing financial forecasts, creating dashboards and building software.

Duet can monitor connected applications, identify items requiring attention and prepare a daily briefing. It can continue working on assignments using its own virtual computer while the user is away.

Most professionals already use several applications to manage work. Email lives in one place, projects in another, customer information in a CRM and financial data somewhere else.

Duet aims to work across those systems and determine how to execute a request using the available tools.

Anthropic has also built in approval requirements for sensitive actions, including sending messages, making payments and signing documents.

The company acknowledges that Duet remains an early product. It can misinterpret priorities, and its personal errand capabilities are less developed than its professional features.

Access began through a research preview for eligible US subscribers, and its background activity counts toward existing Claude usage limits.

Even with those limitations, Duet is the product I find most interesting from a professional standpoint.

Anthropic has combined advanced reasoning with persistent context and substantial application connectivity. If the company makes those capabilities reliable, Duet could become the assistant executives and business owners use throughout their working day.

OpenAI is building agents that work around the clock

OpenAI entered the competition in September with Dots, its new generation of persistent AI agents.

Powered by GPT-6 Astra, Dots operate through their own cloud computers and connect to thousands of applications using OpenAI’s plugin ecosystem.

Users can interact with their agents through ChatGPT and supported workplace platforms.

The idea is to assign ongoing responsibilities rather than start a new conversation for every request.

An agent can retain project context, learn preferences and continue working toward established goals.

OpenAI already has an advantage among people who use ChatGPT for research, business planning, writing and everyday work. An assistant that understands a user’s existing projects could become valuable without requiring another platform to learn that context from scratch.

The competition between OpenAI, Anthropic, Meta and xAI will depend on more than model intelligence.

Each company needs to demonstrate that its assistant can handle actual work, recognize mistakes and manage permissions without constant supervision.

The battle for your personal context

One of the most interesting aspects of the emerging AI assistant market is how personal these systems are becoming.

Traditional chatbots respond to questions using the information available in a conversation.

Persistent assistants can develop a more detailed understanding of someone over time.

They may learn which meetings take priority, how you prefer to travel, which contacts matter, when you typically schedule appointments and which outstanding commitments require attention.

That history can make them more useful.

It can also make switching providers harder.

Today, many AI users move between ChatGPT, Claude and Gemini depending on which model performs better for a particular task.

An assistant that has spent months learning your preferences and connecting with your applications creates a different relationship.

You have invested time granting permissions, establishing workflows and teaching the system how you work.

A competitor may offer a more powerful model, but moving your history and recreating that context may require effort.

For Meta, OpenAI, Anthropic, xAI and the smaller startups, that accumulated personal knowledge could become a significant competitive advantage.

It also makes data privacy and portability important questions for users.

The 10 AI personal assistants worth watching

The market includes everything from personal concierges to AI employees capable of executing business workflows.

I’ve included the products I’m using or testing alongside contenders I believe deserve attention. Their maturity, availability and capabilities vary considerably.

AI assistant	Best suited for	My perspective
Instinct	Everyday assistance, travel, reservations and personal coordination	My current obsession. I just text or email it.
Catch AI	Executive administration, email, scheduling and calls	My dream executive assistant
Meta Muse	Personal errands, shopping and reservations	Fast becoming a personal favorite; haven’t tried it for work
Claude Duet	Complex professional work and proactive assistance	My strongest contender for future leadership
Grok Bot	Specialized AI teams, business tasks and autonomous workflows	Just built my own expert team; impressed so far
OpenAI Dots	Persistent projects, research and cross-application work	Strong potential for existing ChatGPT users
Viktor	Autonomous business operations and recurring workflows	Evaluating for business and marketing automation
Lindy	Custom AI workflows across business applications	Useful for repeatable processes
Martin	Voice, reminders and conversational assistance	Personal coordination
Must AI Assistant	Specialized agents for writing, planning and learning	General productivity toolkit

Instinct: My latest AI obsession

The best thing about Instinct is how little I have to think about using it.

I send a text or email. It gets to work.

For someone who spends much of the day moving between clients, emails, meetings and projects, that simplicity has a lot of appeal.

Instinct can connect with calendars, email and other services. It can handle reservations, purchases and personal coordination while remembering context from previous interactions.

And here’s where its business model gets interesting. Instinct is free.

The company plans to make money from transactions rather than subscriptions or advertising. The idea is to charge merchants a fee when Instinct brings them a paying customer, while keeping the assistant free for users.

Think about the implications. Instinct could become the interface through which consumers make purchases, with merchants paying for the business its assistant delivers.

Its founder reports that the platform is approaching $1 billion in annualized transaction volume. Travel already represents approximately 50% of that activity.

The company hasn’t disclosed how it calculates that volume or begun collecting transaction fees, but the opportunity is hard to ignore.

Hotels and travel companies already pay substantial commissions to intermediaries that deliver bookings. Instinct could eventually participate in those economics, and its founder has identified travel as a particularly attractive category.

Given my background in hospitality and travel technology, that’s fascinating. I’m interested in how Instinct handles bookings, itinerary changes and the complications that come with travel.

For now, I’m enjoying discovering what happens when I give it a request that would normally require several emails or a few browser tabs.

The fact that I can do all of this for free makes the experiment even more appealing.

And it helps explain why investors are willing to value a company with just 14 employees at $10 billion.

For now, Instinct has my attention.

2. Catch AI: My dream executive assistant

Catch AI concentrates on administrative work that executives would traditionally delegate to another person.

It can manage email, coordinate meetings, work across calendars and make outbound telephone calls.

The phone capability is particularly interesting. Catch can contact restaurants, hotels, airlines and service providers, navigate automated systems and provide a summary of the conversation.

At its reported launch price of $99 a month, it costs considerably less than hiring administrative support.

I’ve been using Catch for executive administration and appreciate its focus on familiar tasks that consume time during the working day.

Its value lies in whether it can coordinate those activities without requiring me to step in for every exception.

Catch remains an attractive option for professionals who want something resembling a traditional executive assistant.

3. Meta Muse: Fast becoming my choice for personal life

Muse is becoming a serious contender for my personal activities.

Meta has designed the assistant around everyday requests, including shopping, scheduling, reservations and managing personal commitments.

Users can communicate with Muse through a messaging-style interface or WhatsApp. It can navigate websites, complete tasks and continue working after the conversation ends.

I like the convenience of having an assistant that can handle personal errands without requiring me to configure complicated workflows.

Muse also offers a free tier, with paid plans for additional usage.

There are legitimate privacy considerations. In September, Reuters reported that Meta had tested human contractors to help complete some telephone calls, prompting concerns about disclosure and access to private information. Meta acknowledged problems with the experiment and temporarily withdrew the capability.

I haven’t explored Muse for professional work yet, but I’m interested in seeing how its personal assistant capabilities develop.

4. Claude Duet: The potential professional category leader

Claude Duet deserves a prominent place in this comparison because it aims to manage complex professional work.

Its ability to connect with thousands of tools allows it to move between email, documents, financial systems and business applications.

Anthropic has also given Duet a proactive role. It can review connected information, identify priorities and prepare work before someone asks.

The product is still in research preview, so I’m reserving judgment on its day-to-day reliability.

But I expect it to put considerable pressure on established executive assistants and business automation platforms.

5. Grok Bot: I’ve built my own team of AI experts

Grok Bot is one of the newest products I’m experimenting with, and I’ve already created a small team of specialized agents.

The concept is interesting because each bot can take responsibility for a particular type of work.

Instead of asking one assistant to research, write, analyze and manage everything, I can create specialists with different skills, instructions and access to relevant tools.

Grok Bot allows those agents to coordinate with one another, share context and work on assignments in parallel.

Each bot operates through a persistent computer, where it can navigate websites, work with applications and produce documents or other deliverables.

Users can also demonstrate a workflow once and ask a bot to remember the steps for future use. Recurring activities can run on a schedule.

The November 2026 Grok Bot 1.0 update introduced more proactive capabilities. Bots can recognize unfinished work, suggest next steps and participate in shared conversations through Slack and Microsoft Teams.

For a business owner or consultant, the possibilities are intriguing.

You could have one agent monitoring industry developments, another preparing research and a third reviewing work against established guidelines.

I’m still in the early stages of testing my expert team, but the ability to give each bot a defined role and have them coordinate work has made a strong first impression.

The question I’ll be watching is whether these agents can produce consistently useful results without requiring the same level of management as a human team.

Given Musk’s investments in AI and Grok Bot’s rapid development, I expect xAI to remain a serious player in this competition.

6. OpenAI Dots: Persistent assistance for existing ChatGPT users

OpenAI Dots builds on the work many professionals already conduct inside ChatGPT.

Dots can retain context, connect with applications and continue working on ongoing assignments.

The ability to remember projects and priorities could be valuable for consultants, executives and business owners managing multiple responsibilities.

The question is how effectively these agents translate existing knowledge into completed actions across applications.

For someone who already uses ChatGPT extensively, Dots could provide a natural extension of that workflow.

7. Viktor: An AI employee inside your business

Viktor takes a different approach.

It operates inside Slack and Microsoft Teams, where users assign work through conversation.

Viktor can work across connected tools, update customer records, build reports, manage workflows and create internal applications.

It also supports scheduled and recurring tasks.

I’ve been evaluating Viktor as a potential way to automate business and marketing work across multiple client accounts.

Its pricing structure is based on credits used to complete assignments, making it important to assess the cost of running complex or frequent workflows.

Viktor interests me because it aims to execute business tasks rather than simply recommend what someone should do next.

8. Lindy: Flexible automation for recurring work

Lindy lets users create AI agents that perform specific processes across connected applications.

It can handle email workflows, meeting coordination, CRM updates and other routine activities.

The flexibility appeals to businesses that want to automate processes around their existing systems.

However, more complicated workflows require configuration and testing, while usage-based pricing can increase with activity.

I’d consider Lindy for predictable, recurring work that can be defined and measured.

9. Martin: Personal assistance through conversation

Martin emphasizes voice and text interaction.

Its capabilities include reminders, scheduling and personal coordination.

For users who prefer communicating through conversation rather than managing another dashboard, that approach has appeal.

I’d want to test how well it handles multi-step tasks and unexpected complications before depending on it for important commitments.

10. Must AI Assistant: Specialized help for everyday productivity

Must AI Assistant provides specialized AI agents for activities such as writing, planning, learning and creative work.

Its approach differs from autonomous personal assistants capable of making calls or executing transactions.

It may appeal to users looking for several productivity tools in one application.

For now, I see it as a broader AI toolkit rather than a direct competitor to Instinct, Catch or Claude Duet.

The privacy question nobody can afford to ignore

The more useful these assistants become, the more access they require.

An assistant that manages your calendar needs to understand your schedule. One that handles email needs access to conversations. An agent making purchases may need a payment method and permission to interact with accounts.

That creates real risks.

An assistant might misunderstand an instruction, make an unwanted reservation or send sensitive information to the wrong person.

Developers are responding with isolated virtual computers, encrypted credential storage, audit histories and approval requirements for sensitive actions.

But users still need to pay attention to the permissions they grant.

I’m comfortable experimenting with these platforms, although I prefer starting with lower-risk tasks and expanding access as the assistant proves dependable.

There’s another issue worth considering: how much personal information you want one company to accumulate.

A personal assistant that knows your communications, preferences, appointments and financial obligations could become one of the most revealing collections of information about your life.

Providers need to earn that level of access.

Will one AI assistant eventually replace them all?

Right now, I’m using different products for different purposes.

Instinct has become my current obsession because interacting with it feels so natural. I just send a message, and it gets to work.

Catch remains appealing for executive administration. Muse is gaining ground for personal tasks, and I’ve started experimenting with a small team of Grok Bots for specialized business work.

Viktor interests me for business automation, while Claude Duet is the product I’m watching most closely for complex professional work.

I can see why these companies have attracted so much investment.

A personal AI assistant that handles meaningful work across applications could become an essential part of someone’s day.

Meta can introduce Muse to an enormous consumer audience. OpenAI has an established relationship with ChatGPT users and a growing plugin ecosystem. Anthropic has strong reasoning capabilities and a focus on professional work. Musk’s xAI is developing another model of assistance through coordinated AI teams.

Smaller companies such as Instinct, Catch and Viktor have an opportunity to provide better experiences for specific needs.

I expect prices, capabilities and reliability to change considerably as companies compete for adoption.

The real test is how much work you can hand over

Instinct’s rapid rise to a $10 billion valuation illustrates how much investors believe in this category.

The funding also raises an important question: how much of that valuation reflects demonstrated demand, and how much depends on expectations about future adoption?

The technology is progressing quickly, but autonomous assistants still make mistakes.

They encounter websites they cannot access, instructions they misunderstand and tasks that require human judgment.

For me, the test comes down to a few practical questions.

Can the assistant complete the work I’ve assigned without requiring constant intervention? Can it recognize when something has gone wrong? Does it understand enough about my priorities to become more useful over time?

And perhaps most importantly, does using it feel easier than doing the work myself?

That’s what has impressed me about Instinct. I can communicate with it as naturally as I would another person.

My early experience building a team of Grok Bots has raised another possibility: instead of relying on a single assistant, I may eventually have several specialists working together across different parts of my business.

I wouldn’t be surprised if my preferred setup six months from now looks entirely different.

Silicon Valley is spending billions competing for that position in our personal and professional lives.

I’m enjoying finding out how much of the work I can hand over.

About the Author

Ivana Johnston is a B2B marketing agency founder, strategist, and advisor to SaaS, hotel and travel technology companies, from emerging startups to global innovators. For over two decades, she has partnered with founders, investors, and executive teams on positioning, go-to-market strategy, growth, and bringing new technologies to North America. Her work has contributed to billions in funding outcomes and company valuations reaching 10x revenue.

Today, much of her work focuses on AI and how it is changing discovery, buying behavior, business models, and competitive advantage across travel. Her perspectives have appeared in Forbes, Entrepreneur, and leading industry publications.

Connect on LinkedIn

About Puzzle Partner

Puzzle Partner is a specialized B2B marketing, PR, and go-to-market (GTM) firm serving SaaS, hospitality, and travel technology companies. We provide strategic counsel, fractional CMO leadership, and outsourced marketing for startups and high growth businesses. Since 2014, we’ve helped some of the industry’s most visionary brands enter new markets, accelerate growth, and build authority. Through strategic positioning, authentic storytelling, PR, thought leadership, and AI search visibility, we help companies connect with buyers and stand out in competitive markets. Companies across our client portfolio have collectively achieved over $1.4 billion in funding and acquisition outcomes. Learn more at Puzzle Partner.