PIN AI launches an encrypted AI assistant with personal loyalty and privacy as its main focus

There’s a focus on protecting our personal data now, perhaps more so than ever before, be it from foreign powers, Big Tech, or Elon Musk’s DOGE. Consumers are wary about where their data is going, what it’s being used for, and how, or if, they can put up safeguards, especially while using tech tools like search engines or AI assistants.

With several AI tools and assistants hitting the market over the past couple of years—Microsoft’s Copilot and Meta’s Llama, among many others—much of our data, queries, and summaries are being fed to large tech companies, used to train their artificial intelligence models.

While some users may not mind, it could turn others off, which is why one company is trying a different approach.

PIN AI, a Bay Area-based AI startup, is rolling out what it calls the “personal AI network” with the launch of a mobile application this week, complete with a native large language model (LLM) that is designed to become a “personalized intelligence network,” or “PIN.”

In a nutshell, PIN AI’s app and LLM “live” on a user’s smartphone, accessing the personal data via other apps and platforms that only the user authorizes, and then provide specific and tailored insights or answers based on that information. It’s similar, in some ways, to Apple Intelligence.

That’s important, say PIN AI’s founders, because AI tools are the next great online frontier, even if they’re not ubiquitous quite yet.

“These AI products are becoming the universal interface to access online services,” says Davide Crapis, one of the company’s cofounders, along with Bill Sun and Ben Wu. “It’s becoming the new way that we access the internet.”

Getting personal

Personalization is going to be incredibly important going forward—not only to protect users’ data and information but also to provide relevant information.

For example, suppose a user allows PIN AI to connect to their Gmail account, social media accounts, banking information and payment apps, text messages, and internet browsing data. Tapping into all of that information, PIN AI will be able to deliver personalized, in-depth answers or responses, unlike any or most other AI tools on the market.

In other words, while other AI models utilize all of the data they’ve scraped and ingested to train themselves to generate responses for users, PIN AI only uses the personal data and information a user has given it access to.

Crapis, Sun, and Wu, collectively, have been working on AI technology for many years, and say that it may sound complicated but that users will quickly realize it’s a fairly simple product.

“There’s an LLM packed with the app,” says Sun, “so you get a personalized AI.”

As an example of how “personal” the AI tool can get, Sun tells Fast Company that a hypothetical user could simply bark a voice command at the app such as, “order my mom a birthday present.” From there, PIN AI would sift through past communications and other information, discerning what the user’s mom might want, where to get it, and at the best price. It can also place the order and have it delivered—with no additional effort from the user, other than the initial voice command.

It’s all about sovereignty

Another example: If a user were to ask PIN AI a more complicated question, such as, “can I afford to buy a new car,” the tool could likewise analyze the user’s bank accounts, figure out what types of vehicles the user may want or need (incorporating family needs, etc.), their favorite color or other potential specifications, financing needs, and deliver up an answer of either yes or no, along with a list of potential buying options from nearby dealerships.

While it’s all pretty powerful, users may rightfully be wary about giving an app and AI tool so much access to their information. That’s another thing the PIN AI team says they’re focusing on: encryption and security.

The tool is built on blockchain protocols and users’ data and information are encrypted, so if a user loses their phone with PIN AI installed, a bad actor still shouldn’t be able to access it.

Resonating with early users

After rolling out a trial app, Sun says that it has attracted around 2 million users, almost exclusively through word of mouth, as well as an active community on platforms like Discord.

In terms of monetization, the company plans to drive revenue by charging the various agents from different apps and platforms that will interact on the model a per-use fee—so users themselves don’t pay anything. The monetization takes place within an agentic AI marketplace, and those agents will fork over a slice of applicable commission fees.

And it’s also gotten the attention of investors. The company is backed by a16z Crypto (CSX), Hack VC, and Sequoia Capital U.S. Scout. So far, PIN AI has raised more than $10 million, according to data from PitchBook.

With the app launch, the team hopes to attract more users, refine the product, and see how PIN AI ultimately slots in on the market among other AI tools. 

“It’s about the sovereignty of the user and their data,” says Crapis. “With PIN AI, you have full control.”

View the full article