The AI boom may have started in 2022-2023, but Adobe’s experimentations with AI go back way further.
It’s been a decade now since Adobe first launched Adobe Sensei on November 3rd 2016, during the annual Adobe MAX conference. Sensei was introduced as Adobe’s first artificial intelligence and machine learning framework, integrated across the entire Creative Cloud suite.
Since 2016, Sensei and its frameworks have evolved into Adobe Firefly, which Adobe refers to as their ‘AI-native platform’. Firefly was a major focal point of Adobe MAX 2025, particularly during a presentation on ‘The Future of Creative AI’. Positioning Firefly as key infrastructure within the intersection ‘where creativity meets intelligence’ particularly in the realm of eCommerce and digital marketing, Adobe rolled out a collection of Firefly updates at MAX last year specifically designed to boost the commercial viability of Firefly AI tools.
So how far has Adobe come in the year since, and what innovations might we be able to expect at MAX this year?
What is Adobe Firefly?
Adobe Firefly is Adobe’s proprietary generative AI platform. Designed to act as an ‘AI-infused creative toolkit’, Firefly offers a plethora of generative AI tools and features, ranging from text to image or even image to video AI generation.
● Text-to-image generation
● Text-to-video generation
● Text-to-audio generation
● Sketch-to-image generation
● Image-to-image generation
● Image-to-video generation
● AI face generation
● AI animation
● AI avatar and character generation
● AI music generation
● AI translation and dubbing (in 15+ languages, including Hindi)
Alongside these AI generation tools, Firefly also offers dynamic AI-powered editing capabilities, with tools like Generative Fill for seamless removal or replacement of image elements and expanding on image or video files, to Generative Match for applying the same stylistic edits across multiple different designs for easier brand editing.
Designing Firefly to be the definitive AI-native platform
The reason why Firefly’s catalogue of AI features is ever-expanding is simple: Adobe wants Firefly to be the definitive ‘all-in-one’ creative AI platform for all global users. Where other tools may only offer traditional text-to-image generation or only excel at particular outputs, Adobe is determined to make Firefly a high-quality jack-of-all-trades and a one-stop shop for creative AI capabilities.
They’re achieving this goal with a three-pronged strategy in play:
Developing AI features and edit tools that align with what users need (i.e. based on customer feedback and extensive AI research powered by Adobe Research)
Integrate Firefly tools across the wider Creative Cloud suite for comprehensive functionality
Expand Firefly’s compatibility with other AI partner models
Firefly as a hybrid model ecosystem
Alongside improving their own AI models over time, Adobe is also investing heavily into partner model integration, ensuring Firefly maintains compatibility with other AI models like Gemini, ElevenLabs, Runway, and OpenAI’s GPT Image. Last year, Firefly was only compatible with only about half a dozen other AI partner models. Today, that number has virtually doubled, and Firefly is now compatible with:
● ElevenLabs
● Flux 1.1 by Black Forest Labs
● Flux.1 Kontext by Black Forest Labs
● Gemini 2.5 Flash Image (Nano Banana) by Google
● GPT Image generation model by OpenAI
● Ideogram
● Imagen by Google
● Kling AI
● Runway Gen-4 by Runway
● Topaz Astra by Topaz Labs
● Topaz Bloom by Topaz Labs
The primary advantage of offering compatibility with all of these other prominent AI models is that Adobe users can easily migrate their generative projects over to the Firefly platform and iterate in one central space. This positions Firefly to become a ‘creative command centre’ for generative AI projects.
Firefly integrations across the Creative Cloud suite
On top of expanding their partner model integrations, Adobe is also continuing on with their original strategy with Sensei integration across the Creative Cloud. This means you can use Firefly AI-powered editing tools both within the Firefly app itself, plus within other Creative Cloud apps like Photoshop, Illustrator, Acrobat, Premiere Pro, Express, and more.
The reason Firefly integration works so well is simple: Adobe’s universal user interface. Given a good majority of creative professionals around the world already maintain proficiency with Creative Cloud tools, mastering the Firefly platform may not come with as steep a learning curve as other similar generative AI tools. And with Firefly editing already accessible through your favourite Creative Cloud apps, Firefly is well on track to being the most widely used creative AI tool on the market today.
What constitutes a ‘commercially safe’ AI tool?
With great output comes great responsibility. Adobe has well and truly absorbed that truth through their ongoing commitment to maintaining AI transparency and ethical training practices for their Firefly AI models. In fact, Adobe is describing Firefly and Firefly Foundry as ‘commercially safe AI’, and inviting brands to see what the platform can do.
Adobe’s approach to delivering commercially safe AI boils down to three major tenets:
1. Copyright-safe outputs
Adobe asserts that all designs generated using Firefly are copyright-safe, ensuring the Firefly user retains full ownership rights over the library of generated assets. This means that anything you produce using Firefly can be confidently used for business or marketing purposes, making Firefly the ‘safest AI tool’ for commercial projects.
2. Content Credentials
AI transparency is the flip side of the coin to ownership rights. This is precisely why Adobe co-founded the CAI (Content Authenticity Initiative), a global industry body dedicated to AI ethics, governance, and transparency.
One of the initiatives put in place through Adobe’s collaboration with the CAI is ‘Content Credentials’, which are essentially metadata labels applied to every AI-generated asset that includes creator details and how the content was made (i.e. 100% AI-generated, photography with AI edits, etc). Using Content Credentials for all assets generated or edited using Firefly, Adobe is ensuring greater AI transparency and ensuring user assets are commercially ready for publication and distribution.
3. Ethical AI training
Adobe trains Firefly’s AI models entirely using public domain data as well as independently managed copyright-safe datasets, predominantly comprising Adobe’s own Adobe Stock library of royalty-free image, video, and audio assets.
Alongside using closed, ethically sourced datasets, Adobe has also outlined that they don’t train their models on user content. This means anything you upload to the Firefly platform or download off of Firefly is 100% private and will never be used by Adobe for AI training purposes.
Firefly Foundry, agentic AI and workflow automation
On top of their commitment to commercial viability, Adobe is also flipping traditional generative AI tool design firmly on its head by integrating agentic models into Firefly and Firefly Foundry. Firefly Foundry is essentially a business-oriented take on a generative AI tool suite, allowing users to not only create their own custom design templates and brand style kit, but even their own proprietary, IP-protected generative AI models.
With Firefly Foundry, commercial users like small business owners can essentially create their own AI models trained directly on their brand’s unique resources, ensuring greater control over creative workflow automations not only across Firefly but other Adobe tools, including GenStudio, Express, and other apps within the Creative Cloud suite.
Alongside supporting proprietary model development for business users in Firefly Foundry, Adobe’s also recently rolled out Firefly AI Assistant for all users. Firefly AI Assistant was originally announced in the leadup to Adobe MAX 2025, first being named ‘Project Moonlight’. Recently, Adobe rebranded ‘Project Moonlight’ to the Firefly AI Assistant we know today, but its purpose has largely stayed the same: to be ‘your personal orchestration assistant’.
In other words, the Firefly AI Assistant helps users transfer their Firefly outputs across other Adobe apps. For example, if you were looking to migrate a design produced in Firefly to Photoshop or Illustrator, Firefly’s AI Assistant can help you get that done.
Adobe’s agentic AI offerings are also designed to offer recommendations, both across design work and automation assistance, helping Firefly users cut down on their project management admin and unlock the sheet automation potential of Firefly and Firefly Foundry as commercial AI tools.
So what’s next for Adobe Firefly?
All of Adobe’s updates and AI innovations over the past twelve months have made one thing abundantly clear: the software giant is gearing towards making Adobe Firefly a shoe-in for sustainable AI integration into creative workflows.
With that goal in mind, we can assume that Firefly updates and developments announced at this year’s coming Adobe MAX conference will showcase more exciting innovations geared towards creative commercial projects. Whether it be a social media ad campaign for a local business, or polished animations or soundtracks for film, television, or web series projects, you can create it to scale with Adobe Firefly.