Snap is advancing its AI strategy with new product features and partnerships, including the launch of Animate It, an AI-powered video generation tool, and a $400 million deal with Perplexity AI to bring conversational search directly into Snapchat.
These developments highlight Snap’s broader effort to weave generative AI into core user experiences, from content creation to discovery, as competition intensifies among social media platforms to retain engagement and attract advertisers.
AI features and partnerships
Animate It allows users to create short, personalized videos from text prompts, marking Snap’s first open-prompt AI video generation Lens. Available exclusively to Lens+ subscribers, the feature is built on Snap’s internally developed AI video model and can be accessed directly from the Snapchat camera.
In parallel, Snap has partnered with Perplexity AI to integrate its AI-powered answer engine into Snapchat’s chat interface beginning in early 2026. Under the agreement, Perplexity will pay Snap $400 million over one year, through a combination of cash and equity, as the product rolls out globally. Revenue contributions from the partnership are expected to start in 2026.
The integration will allow users to ask questions and receive conversational answers drawn from verifiable sources without leaving the app. The deal marks the first large-scale integration of an external AI partner directly into Snapchat and positions Snap as a distribution platform for AI services targeting a young, mobile-native audience.
Snap’s AI and augmented reality patents
This section highlights selected patents that demonstrate Snap’s ongoing investment in intelligent, immersive, and user-focused technologies. The disclosures span advances in conversational artificial intelligence and augmented reality-based communication, offering insight into how the company is shaping more responsive digital experiences with an emphasis on personalization, user control, and real-time engagement.
Enhancing conversational AI through intent recognition
Chatbots often face difficulty determining whether a user is engaging in casual conversation or making a request that requires a specific and actionable response. This can result in vague or incomplete answers. The challenge is particularly pronounced when a prompt could be resolved by drawing on external data sources rather than relying solely on a generative model.
U.S. Pat App. Pub. No. 2024/0356873 addresses this issue by introducing a system that distinguishes between open-ended conversational prompts and intent-driven requests. The invention is designed to identify when a user is seeking general dialogue versus when the input reflects a clear intent that can be fulfilled through an application programming interface (API) call to an external service.
When a definable intent is detected, the system generates an API request to retrieve relevant information, which is then used to guide the response produced by a generative AI model. This approach enables responses that are more accurate, contextual, and actionable. Prompts that do not signal a clear intent are instead handled as conversational input, allowing for more natural, open-ended exchanges.
Titled “Personal AI Intent Understanding,” the patent application was filed on April 18, 2024, and published on October 24, 2024. The listed inventors are Jason Brewer, Wenxiang Chen, Michael James Cunningham, Dmytro Ishchenko, Tanvi Motwani, and Diwakar Punjani.
Facilitating time-limited augmented reality interactions
Real-time augmented reality communication can be immersive, but without clear limits, it raises challenges around privacy, session control, and user fatigue, particularly when multiple participants are involved. Unrestricted live video and audio connections can also place added strain on devices and networks, reducing usability over time.
U.S. Patent No. 11,985,175 introduces a system for time-limited augmented reality communication between multiple users. The invention enables a host to initiate a virtual interaction session and invite selected participants to join a real-time communication link, allowing them to view a live camera feed of the host’s physical environment augmented with digital content.
The system relies on configurable session parameters, including a predefined micro-chat duration that sets a time limit for each real-time interaction. Once the allotted time expires, the communication link is automatically terminated, ending both the live video feed and audio connection. This structure provides built-in controls for privacy, session flow, and engagement, while still allowing multiple users to join sequentially or concurrently during an active session.
Titled “Virtual interaction session to facilitate time limited augmented reality based communication between multiple users,” the patent was filed on March 19, 2021, and granted on May 14, 2024. The listed inventors are Brian Anthony Smith, Yu Jiang Tham, and Rajan Vaish and Hemant Surale. Schwegman Lundberg & Woessner represented Snap in the filing.
Enabling collaborative messaging through asynchronous media contributions
Traditional digital messaging systems are built around linear exchanges, requiring users to send separate messages each time new media is added. This approach fragments conversations, increases file sizes, and often degrades visual quality as content is repeatedly recompressed, particularly when multiple participants contribute images or videos over time.
U.S. Patent No. 11,044,217 introduces a system that allows multiple users to contribute media asynchronously within a single interactive message. Instead of exchanging separate files, participants add images or videos to predefined interactive objects within a shared message, enabling the content to evolve as it is passed from one user to another.
To improve efficiency and preserve quality, the system avoids recompressing the entire message with each contribution. Original media from each user is stored separately and dynamically combined at render time by merging selected features, such as facial regions or visual elements. This method reduces file size, minimizes compression artifacts, and maintains higher visual fidelity across devices.
Titled “Generating interactive messages with asynchronous media content,” the patent was filed on December 4, 2019, and granted on June 22, 2021. The listed inventors are Grygoriy Kozhemiak, Oleksandr Pyshchenko, Victor Shaburov, Trevor Stephenson, and Aleksei Stoliar. Schwegman Lundberg & Woessner represented Snap in the filing.
Snap: Patenting Activity
Snap’s patent filings rose steadily from 2017 through 2020, peaking in 2020 as the company ramped up investment in new technologies. The increase reflects an intensive period of innovation, with filings spanning both granted and pending patents, as Snap sought to protect core intellectual property while expanding its product portfolio.
The surge in patent activity coincided with a period of rapid growth for Snap. In 2019, daily active users increased 17 percent to 218 million, while fourth-quarter revenue rose 44 percent to $561 million. That momentum accelerated in 2020, with daily active users climbing 22 percent to 265 million and fourth-quarter revenue jumping 62 percent to $911 million.
Snap linked its results primarily to continued investment in camera and augmented reality technologies, which by 2020 were used daily by more than 200 million people. After that year, patent filings fell sharply, pointing to a move away from broad foundational development toward a narrower focus on improving existing products and expanding monetization as the company matured.
Snap: Top Law Firms
Snap’s patent filings have been handled by a range of intellectual property law firms, with filing activity unevenly distributed across providers. Schwegman Lundberg & Woessner accounts for the largest share of filings, which points to a prominent and sustained role in the company’s patent prosecution efforts.
Beyond the leading firm, Snap has engaged a broad mix of domestic and international law firms to support its patent filings. Kim & Chang IP, Mathys & Squire, and Culhane Meadows account for substantial portions of this work, alongside a number of other firms with smaller but ongoing involvement. The distribution reflects a diversified legal approach, with Snap relying on multiple firms to handle patent matters across different jurisdictions and technical areas.
Snap: Top Technology Areas
Patent activity at Snap from 2015 to 2025 is concentrated in computing and visual technologies that support its camera-based products. Filings related to electric digital data processing (G06F) form the largest portion of the portfolio, highlighting the importance of core software, data handling, and system-level processing. Image data processing and generation (G06T) also accounts for a substantial share, alongside patents covering the transmission of digital information (H04L), reflecting the technical requirements of creating, processing, and delivering visual content in real time.
A significant share of Snap’s patents also falls under information and communication technology (G06Q) and computing arrangements based on specific computational models (G06N), pointing to ongoing development of algorithmic systems, including machine learning and AI-driven processes.
Additional activity appears across image and video recognition (G06V), pictorial communication (H04N), optical elements and systems (G02B), and wireless communication networks (H04W). Taken together, these categories illustrate the breadth of technologies underpinning Snap’s products, spanning software infrastructure, computer vision, networking, and hardware-adjacent components that support augmented reality and visual communication features.
