Every year during Adobe Summit, the company gives the public a sneak peek into its latest experimental innovations, with 40% of them historically making it to actual rollout. In the past, some of Adobe's most popular features were first revealed as part of Sneaks, including Generative Fill, then dubbed Project Fast Fill.

Also:Adobe reveals 10 purpose-built AI agents -- here's what they can do for your business

The Sneaks are proposed by Adobe employees and build on Adobe's existing offerings to optimize creatives' workflows. Unsurprisingly, this year, most of the Sneaks finalists involved AI agents that can help marketers and brands alike. To get a glimpse into Adobe's future, keep reading below.

Project Perfect Context

This experiment would help teams better understand their customers through an AI agent that can combine external insights, such as economic data and global weather, with first-party behavioral data within the Adobe Customer Journey Analytics in the Adobe Experience Platform, according to the post.

Also: AI tools are increasingly driving traffic to retail sites, Adobe Analytics finds

Using the AI agent, users can learn more about customer experiences by conversationally asking questions that leverage more information about both datasets. These insights can ultimately inform the building of future experiences to cater more to the audience's wants and needs, ultimately improving campaign performance.

Project Slide Wow

Slide decks are a great medium for presenting information, but building them can be time-consuming. Senior leaders, in particular, often bear the burden of synthesizing robust data insights into presentations to share the findings with other teams. Project Slide Wow automates the entire process, converting data in Adobe Customer Journey Analytics into an engaging visual Microsoft PowerPoint presentation that includes visualizations and even speaker notes.

Also: Forget Sora: Adobe launches 'commercially safe' AI video generator. How to try it

The presenter can then ask a question and get an answer pulled straight from the dataset, even if it wasn't originally included in the slides. In the live demo at Sneaks, the presenter converted a dataset of favorite ice cream flavors into a visually engaging and well-organized PowerPoint presentation with different sections, charts, speaker notes, and even a concluding slide with questions. It kept a theme and natural order that drove the point.

Project Site Leap

Content management systems play a central role in many businesses' core operations. Yet even when these systems are outdated or no longer the best fit, companies often struggle to make a change due to the high costs and potential disruption to their workflow. As the name implies, Project Site Leap makes the transition easier using an AI agent that can take existing brand pages, import the content into Adobe Experience Manager (AEM), and align it with new design mocks, according to Adobe.

Also: Adobe Acrobat's AI Assistant can now decipher complex contracts for you

In the demo, the presenter switched an open-source blog into AEM by asking Project Site Leap to "Onboard the site to AEM," and was able to import all of the content into AEM within seconds. Then, the second step was to reproduce the design of the blog into AEM, and again, it was able to do it within a second. Lastly, the presenter asked the AI agent to freshen up the images and the typography, and again it was able to do so. If you want to change the design entirely but keep the content, you can also upload a mock-up and have the agent transform the old content into the new style.

Project Get Savvy

This experiment is the ultimate tool for marketers, helping them through their kickoff campaigns by aiding the brainstorming, creation, and deployment of a campaign through the use of various AI agents. With Project Get Savvy, marketers would be able to interact with agents in real time to create content and visuals, find the best ways to reach audiences, and even create personas that resemble the target audience and get their feedback.

Project Frame Sense

Taking a step away from agentic capabilities, the next couple of experiments focus on using generative AI to deliver value to marketers and their audiences. Project Frame Sense helps marketers create tailored experiences in the travel and hospitality sector, where creating tailored experiences is specifically worthwhile.

Also: Adobe Photoshop adds Live Co-editing for more productive team work

The experiments connect customer insights from the Adobe Experience Platform with video content so that customers can receive a highly personalized video tailored to their preferences. The videos even use AI-powered digital avatars. As seen in the image below, in the demo, the presenter took an AI avatar of Sneaks host Eric Matisoff, Principal Evangelist of Analytics & Data Science at Adobe, to create a personalized video for co-host Ken Jeong that took into account Ken's interests, such as what food he wanted to eat on the trip.

Project Panorama

Project Panorama helps brands make sense of user interactions on mobile applications. This task is typically challenging because of all the different elements found in apps, such as screens, offers, and messages. Instead, Project Panorama would offer a map that brands can use to visualize the user's journey.

Also: Adobe's new AI feature lets you edit stock images on the fly -- no Photoshop needed

The map would update in real time, track performance metrics and user behavior, and outline each page and screen of the experience, according to Adobe. Ultimately, marketers could use those insights to deploy new offers and send messages from the Project Panorama interface without additional code, making optimization to increase user engagement easier.

Project Vision Cast

This project focuses on making the collaboration between marketers and creatives more seamless. Using generative AI, it combines data insights and brand imagery to create rough interpretations of new product concepts. These interpretations are then sent to Adobe's Project Concept so both teams can brainstorm and collaborate.

Get the morning's top stories in your inbox each day with ourTech Today newsletter.

Artificial Intelligence

The best AI for coding in 2025 (and what not to use - including DeepSeek R1)

I tested DeepSeek's R1 and V3 coding skills - and we're not all doomed (yet)

How to remove Copilot from your Microsoft 365 plan

How to install an LLM on MacOS (and why you should)

• The best AI for coding in 2025 (and what not to use - including DeepSeek R1)
• I tested DeepSeek's R1 and V3 coding skills - and we're not all doomed (yet)
• How to remove Copilot from your Microsoft 365 plan
• How to install an LLM on MacOS (and why you should)