September 04, 2024

Adobe AI Assistant – a new AI star performer 

Tan Hong Tam
Manager - Licensing Solutions NSW & ACT, Data#3

From pioneering PDF technology to industry-leading creative platforms, Data#3 and Adobe help organisations embrace digitalisation and transform the way that work gets done.

With a relationship spanning 20 years, Data#3 has become Adobe’s largest Australian partner. It means we’re uniquely positioned to simplify licensing and help customers maximise their software investment.

AI is a topic that prompts a lot of opinions beyond the realms of the IT industry. From mainstream and specialist media to politicians and business leaders, a lot of voices are weighing in. There are the scare stories and dire warnings, the hyperbole and promises and, somewhere in between, the truth.

It is with this backdrop that Adobe AI Assistant has entered the fray. True to their creative roots, Adobe has approached both the challenges and potential of AI with their customary blend of design genius and out-of-left-field thinking. For anyone using Adobe products – yes, from Acrobat Reader upwards – the working day might be about to get a lot better.

Partnership with Microsoft

Adobe’s strength lies in designing essential digital tools for anyone who wants to create, no matter if that’s pulling together a perfectly presented report or dreaming up an award-winning ad campaign. They opted to partner with Microsoft for the AI engine behind their offering, basing AI Assistant on Microsoft Azure OpenAI Service. It is both an acknowledgement that Microsoft is the best AI platform of the moment, and a choice to focus on delivering something uniquely Adobe that integrates beautifully into the world of their users.

There are some important differences to the way Adobe uses Microsoft Azure OpenAI Service compared to the Microsoft offering. This positions AI Assistant as complementary to Copilot for Microsoft 365 (M365), rather than in competition. When you query Copilot, it will source input from a vast ecosystem including M365 documents, the web, emails, and Teams chats. In contrast, AI Assistant sources information solely from files that you open in Acrobat. That information is not used to train AI Assistant or Microsoft, and what you create is only retained for 12 hours on the Adobe system.

Rapid introduction to AI

We know from experience that the way change is managed is usually the deciding factor between resounding success and underwhelming results. Given the extraordinary shift that AI promises, it is no surprise that most technology-progressive organisations are in a phase of assessment and preparation for Copilot. As one of just a handful of Australian businesses selected to be part of Microsoft’s Copilot Early Access Program (EAP), we’ve been fortunate to get hands-on experience of the phenomenal capabilities it brings.

At the centre of this is understanding risk and preparing a safe introduction. All this takes a bit of time. Copilot for M365 preparedness assessments and plans have been keeping our consultants busy as they help customers understand the implications to their data, the opportunities unique to their businesses, and the elements to making their rollout a success. Because Copilot may harvest source documents from across your ecosystem, the right rules must be put in place to protect sensitive information.

Adobe’s AI Assistant is a far quicker and simpler decision, making it very quick and easy to put in place. It is a great way to get users started on AI while you prepare for Copilot for M365. The primary reason for this speed is that users are working with sources they provided, meaning they can reasonably be expected to have permission to access that information. Their sources are not used to train Adobe’s model, so whatever data they’re working with goes no further than their own piece of work.

Working with Adobe AI Assistant

Adoption might be one of the quickest decisions you make but that certainly doesn’t mean the effect will be trivial. How? The way you use AI Assistant will depend on your industry, your role, and what you’re trying to achieve. We’re seeing individual users find interesting applications for AI Assistant that relate to their own workflows.

From a user perspective, it is very easy to get started. Say, for example, someone wants to create a report based on a lengthy document. They can open the document in Acrobat, and upfront get a summary of the key insights in next to no time. The individual or business can then engage with their documents using natural language to significantly accelerate their time to knowledge. The user might add documents – in Word, PDF, even PowerPoint – that can lend extra information on their topic. They effectively have a conversation with AI Assistant as it guides them through the process.

One of the very handy features of AI Assistant is that it includes citations to the exact source of every bit of information, so it is simple for the user to check accuracy. This is designed to make it quicker to understand and work with documents, not to replace the human element.

Another feature coming soon is that when the user is happy with the content they have created, they can choose a template and their own brand logo to make the report look professional. Because it is designed to integrate seamlessly with Adobe Creative Cloud, users can also extend to create professional web pages, eye-catching social posts or just about anything, ready to publish in seconds.

In education, an admin might use AI Assistant to review multiple lengthy policy documents to create a summary and build a perfectly presented report, saving themselves hours or days of reading to summarise the main points. The faster time to understanding will give a lot more time to dedicate to the million other tasks they are likely juggling.

For government, whose path to Copilot for M365 is likely to be among the most lengthy due to the nature of the data in their ecosystems, AI Assistant is a chance to dip their toes in the AI water. When proposing policy, for example, a government user may typically wade through numerous documents and data sources, where trying to find the most relevant information is like the proverbial needle in a haystack. AI Assistant puts in the hard yards of hunting down those gems, so the user can work faster and put more focus into interpreting key points and presenting them in the most effective way. It is in this collaboration between human and AI where the magic happens.

Getting started

Adobe has a few sample prompts you can play with if you want to get a quick look at what it can do. Just ask your Data#3 Account Manager, and that hundred-page report on your to-read list can be smashed out by lunchtime.

Interested in finding out more about how Data#3, in partnership with Adobe, can help your organisation make the most of the latest AI opportunities? If the answer is yes, get in touch with us today.