The News: Adobe announced it is integrating generative AI functionality into its venerable Acrobat product, enabling users to intuitively summarize, organize, and query documents using natural language queries. The Acrobat AI assistant is available now in its Acrobat Standard and Pro editions and will be rolling out to enterprise versions shortly.
According to Adobe, the key differentiators of the Acrobat AI Assistant from other generative AI-based assistants include the ability to leverage generative AI within the flow of work with a standard document format (PDF); the power of Adobe’s proprietary approach to extracting information from within a PDF, even when the information is contained within complex tables or sections; and the clear attribution of source data when content is summarized.
Adobe also highlighted the following attributes of its generative AI offerings:
- Curates best-in-class for the job to be done; LLMs are prohibited from training on Adobe customer data
- Provides proprietary AI services with deep understanding of PDFs driving high quality, verifiable outcomes
- Incorporates AI ethics to counter bias and harmful content
- Utilizes enterprise-grade security and information governance; only operates on documents provided by user; documents are not stored
- Provides scaled value; delivering value to billions of active Adobe Acrobat users within the flow of work
Adobe Embeds Generative AI Into Acrobat
Analyst Take: Adobe is the latest SaaS company to launch a generative AI-based assistant, joining the likes of Microsoft, Salesforce, Google, Amazon, and others that have been marketing these tools as companions that can automatically extract information from documents and create summaries and insights, and permit users to query source documents in a conversational manner to get specific answers to questions.
Leveraging the Ubiquity of the PDF as a Standard Document Format
Adobe’s Acrobat AI Assistant, however, is somewhat unique in that it is specifically tuned and designed to extract information from within Adobe Acrobat, the most commonly used document reading and sharing format in use today, with more than 500 million monthly Acrobat users, and more than 400 billion PDFs opened within Acrobat in the past year. By incorporating generative AI within Acrobat, users can get insights and summaries without needing to switch platforms, enabling them to incorporate generative AI within their workflow.
Furthermore, documents that have been saved or distributed in the PDF format are generally completed works, or works that are ready for review. Adobe says that as a result, these types of documents are generally considered to be good sources of information that has already been checked and verified, and often contains information that can be considered as a source of record, such as a company’s annual report, a court filing, or a knowledge base article. Incorporating generative AI within Acrobat enables workers to generate insights using the original source material more easily, without needing to convert it to another format, or manually extract the information.
In particular, Adobe says that its generative AI tool can handle more complex information, such as data contained within tables, as well as understand other data based on elements embedded into Acrobat, such as formatting cues, including headers and section markers.
Establishing and Maintaining Trust with Generative AI via Annotations
One of the interesting aspects of Acrobat AI is the way it handles content attribution. When an Acrobat AI Assistant-created summary of insights is delivered, or a specific, natural-language query is run, the results are provided within the AI Assistant pane in the application, and are annotated to refer back to the specific place within the document where the insight or answer has been sourced.
These annotations are a key catalyst for establishing and maintaining trust in generative AI. As illustrated below, when key insights are generated, numbered annotations that point and link directly to where a piece of information was found within the document are clearly visible (see the purple numbers next to each insight in the example captured below).
By making it easy to quickly identify where the information contained in an insight or summary is derived, users can quickly verify that the information is correct (and can submit feedback to Adobe if it is not), and can also quickly copy the information with the annotation for use in another document, saving time and effort. Most importantly, including an attribution to the source information helps to build and engender trust that generative AI is working as it should, rather than hallucinating.
Driving Efficiency by Allowing Users to Incorporate Other Sources of Information
Adobe said it is currently working on features that will allow users to augment the information contained within a PDF from sources on the Web, including from within other document types. While Adobe did not commit to a firm release date for that technology, we expect that functionality will arrive in beta within the next quarter or so, based on comments made from the product team to analysts.
In addition, the product team said it is working on automation of workflows using generative AI, such as redacting personally identifiable information from a document, or classifying a document based on certain attributes. These features likely will be incorporated within the application itself in the near future, as well as exposed through APIs to enable more seamless workflows with other applications.
Continuing Adobe’s Legacy of Responsible AI
One of the other key strengths is Adobe’s commitment to deploying generative AI responsibly. The company is one of the pioneers with using AI, and its extensive work supporting Firefly (largely around compensating the creators who contribute to the company’s Stock business on which Firefly relies) is proof that it takes data, intellectual property, and privacy rights seriously. This should further help solidify Adobe’s reputation as a reliable and trusted partner in the market as it rolls out Acrobat AI Assistant.
Disclosure: The Futurum Group is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of The Futurum Group as a whole.
Other Insights from The Futurum Group:
Adobe Revenue in Q4 Hits $5 Billion, Up 12% YoY, As Growth Continues
Adobe Delivers Digital Experiences via E-Commerce and Workflow Apps
Generative AI Spurs Adobe’s Investment In Digital Content Transparency
Author Information
Keith Kirkpatrick is VP & Research Director, Enterprise Software & Digital Workflows for The Futurum Group. Keith has over 25 years of experience in research, marketing, and consulting-based fields.
He has authored in-depth reports and market forecast studies covering artificial intelligence, biometrics, data analytics, robotics, high performance computing, and quantum computing, with a specific focus on the use of these technologies within large enterprise organizations and SMBs. He has also established strong working relationships with the international technology vendor community and is a frequent speaker at industry conferences and events.
In his career as a financial and technology journalist he has written for national and trade publications, including BusinessWeek, CNBC.com, Investment Dealers’ Digest, The Red Herring, The Communications of the ACM, and Mobile Computing & Communications, among others.
He is a member of the Association of Independent Information Professionals (AIIP).
Keith holds dual Bachelor of Arts degrees in Magazine Journalism and Sociology from Syracuse University.
