Get the basics about the opportunities for infusing artificial intelligence into your content management strategy.
Artificial intelligence (AI) is one of the hottest topics in technology, and in the realm of content and its management, AI is of particular interest. Growing volumes of information continue to be a challenge — or opportunity, for those who can harness it — for enterprises of all sizes and industries. When it comes to leveraging AI in the management of enterprise content, there is much to gain.
Let’s explore the rapidly evolving role that AI and machine learning (ML) play in content management, including available AI offerings and their practical application for enabling better access to critical information. Discover real-world use cases and how early adopters get business value out of these technologies, and understand critical enterprise considerations for getting started with AI and content.
Yes, according to 81% of respondents to Forrester Consulting’s 2024 study. To stay competitive, your organization must close the gaps by integrating AI and automation into your content strategy. Read this compelling study, full of stats and insights, for more.
Content — in all of its various forms — has long been a challenge from an information management perspective. The right information can be nearly impossible to find due to complex technology systems, inadequate and inconsistent metadata attributes, limited search functionality within core business applications and disconnected repositories or systems.
Not only is content inherently hard to manage, the volume and types of content are growing at an unprecedented rate. Many enterprise organizations have accumulated billions of pieces of content in recent years, whether it’s documents, scanned images, emails, video or other formats. While historical content is massive, the reality of today’s business is that hundreds of millions of new objects could be entering an enterprise per month, which can literally multiply an entire corpus of content in the next few years.
In general, content isn’t difficult for people to understand. We consume content every day without even thinking about it. The challenge is that the things we do naturally — quickly classifying content to determine what it is, identifying critical information and data within it, and determining if it’s vital information that needs to be preserved — don’t scale well.
Extracting information from content and entering it into fields and tables is work that people inherently don’t like to do. And doing this work across thousands or even hundreds of thousands of new documents every day is challenging, expensive and difficult to do with consistent accuracy. This is why so many organizations have struggled with enterprise content management (ECM) for so long.
Now, with the ubiquity of AI and ML, there’s a way to process content like a human does, but at a massive scale. Enterprises can deploy a range of services to intelligently extract critical data from content and, in doing so, transform content into actionable information that can be easily found, readily used to perform work and accessible anytime, anywhere, and on any device.
— Forrester Consulting, Transforming processes and experiences with content, automation and AI, 2024
Most modern content solution platforms can integrate with a variety of public cloud services for artificial intelligence. Typically, the content solution platform will pass an object, be it a document, image, or even a video file, to a cloud provider, and will then receive a set of data produced by the AI service.
The world of AI continues to evolve rapidly, and a number of large technology companies now offer a variety of commodity AI services that can be leveraged for working with various forms of content. Here are some of the most popular technologies in use, and examples of how they can be employed to work with content:
Popular AI technologies
| Technology | How it works | Example |
| Natural Language Processing (NLP) | A service that employs ML to perform entity extraction, sentiment analysis and language detection on text. It can also perform document classification. | Perform sentiment analysis on customer emails and chat sessions to identify unhappy customers for priority response. |
| Deep-learning image and video technology | A technology that identifies objects, text, people, scenes and activities in videos and images. It can also detect inappropriate content and perform facial recognition. | Identify celebrity images used in advertising content. |
| Optical character recognition (OCR) | A ML service that identifies scanned documents and images, and extracts specific data values. | Recognize and process forms. |
| Language translation | A neural-machine service that uses deep learning models to translate text accurately and efficiently. | Automatically translate sales and marketing collateral into various local languages. |
| Speech-to-text conversion | A deep learning process that uses advanced machine learning algorithms to transcribe audio files into accurate, readable text in real-time. | Transcribe customer service calls, which can then be processed with sentiment analysis. |
| RESTful API image analysis | A ML service that classifies and assigns labels to images, detects embedded objects and extracts text. | Read license plates in an automobile accident photo. |
| Intelligent document processing (IDP) | A ML technology that uses information extraction capabilities to read, recognize and understand content. | Automate the extraction and verification of forms, such as for loans, jobs, healthcare and more. |
>Read more | Essential AI terms you need to know
Many of these ML offerings focus on providing greater insight into and understanding of content, whether that’s text-based documents, photos and images, or audio and video files.
A lot of value can be derived from these generic models and services, particularly in performing routine tasks with high volumes of content. For example, if you need OCR for a large existing set of content, these services are accurate and highly performant. Real-time sentiment analysis of chat sessions, emails or even social media content is another great use case for these services.
Generic services have been trained with a broad range of data. As a result, these models tend to return generic data, which may or may not be helpful depending on the use case.
Generic ML model
In the picture, below, an automobile accident is shown that has been labeled by Google Cloud Vision (a pretrained, generative AI model).
As you can see, Google Vision returned a number of labels or data values related to the image. However, if you were an automobile insurer, would this data really be valuable?
Custom ML model
One of the benefits of ML is that organizations don’t have to rely on generic models and generic data. They can use ML to train their own custom models that will return data tailored to the needs of the business.
Now, let’s consider a ML model that was specifically trained with lots of pictures and data related to automobile accidents. Here’s an example of the kind of data a custom model could extract from the same image:
Note that now the make, model and factory color of both vehicles have been correctly identified. It also identified two Illinois license plates and captured full and partial plate numbers. There is a face present in the image and identified as the operator, Jim Smith.
Assuming this picture was taken with a smartphone or other digital device, it’s likely the ML model could also use the GPS coordinates from the picture and identify the accident location. No, this isn’t AI, but it’s probably useful for processing and verifying the claim!
The benefits of a custom ML model include extracting business-specific data that adds real value (instead of returning generic sets of labels) and bringing greater automation to the process.
Not only are custom ML model-users no longer dependent on a human to enter these values, users can also automatically alert a claims processor that new information is available for this claim. This is a prime example of AI bringing content and data together to ensure the right information is in the right place, at the right time.
— Forrester Consulting, Transforming processes and experiences with content, automation, 2024
Content enrichment is all about extracting data from content and using that data to make the content more accessible, highly contextual and in short — more powerful. Content enrichment takes many different forms, depending on the type of content and what type of AI models are in place.
Powering content with AI
On an AI-powered content solutions platform, the content housed in core applications across the enterprise gains visibility, quality and actionability. Using various AI tools — be it ML, optical character recognition (OCR) or automated metadata tagging, for example — content becomes unified and usable in a way that legacy content platforms simply can’t match.
Information housed within an AI-powered content solution is not only more useful (because of the contextualization of its data points), but AI can also take that valuable content and initiate processes across integrated systems. For example:
Siemens received millions of invoices from vendors worldwide. The OCR solutions they had previously invested in could extract some fields with accuracy, but there were still too many human touchpoints.
By leveraging Hyland’s solution with built-in intelligence, Siemens accelerated process automation and was able to reach a milestone of 90% for number of data fields extracted without human intervention.
According to a 2024 Forrester Consulting study, 67% of businesses leading in modern content management practices are developing their intelligent automation capabilities to automate manual processes and extract deeper insights from data. Although the trend is rising, that still means a lot of enterprises depend on manual work — some even with handwritten paper forms to process. This comes with critical processing challenges like determining what type of form it is, validating the necessary responses, confirming signature placements and checking if confidential information has been provided (not to mention a lag in accurate data availability).
AI and ML can help companies better automate critical business functions and processes, for example:
Intelligent document processing
Intelligent document processing (IDP) uses AI to read, recognize and understand, as a human would, the text and formatting within semistructured and unstructured content, so forms and documents can be processed automatically. Using ML to “teach” the IDP software how to make sense of documents, the software gets smarter and more effective the more it works.
Process optimization
With AI integrated into content-centric processes and systems, organizations can streamline workflows effortlessly. Use cases for this include using a language learning model (LLM) prompt to identify qualified participants for a clinical trial; automating analysis for requirements or checklists to initiate workflows; searching enterprise repositories for similar content/documents to flag for fraud (insurance and government); identifying and eliminating duplicates in a digital asset management (DAM) solution; checking terms and conditions to underwrite insurance policies; and initiating automated processes related to credit review, fraud detection and compliance management.
Data validation
ML models also enable intelligent exception management to quickly identify what’s missing or inaccurate with a provided form and automatically route it to a customer service representative or back to the customer for remediation.
Records and retention management
Many organizations have struggled for years to implement an effective records management approach for their information. The simple reason for this is that most organizations are unwilling to devote the effort required to look at all their existing content to determine if, when and how it should be retained or deleted.
This is painstaking work for humans, but ML can automatically classify content and extract data from it at a bigger scale. As a result, it’s much faster and easier to examine tremendous volumes of content, classify a variety of documents or information, then automatically identify records, apply requisite retention periods and delete nonvital information.
> Read more | The ultimate process automation guide
Drawing new insights and meaningful connections from your content with the help of AI enables beneficial modernization in various ways, such as:
Now that we have explored the difference between commodity and custom models and have examined some real-world use cases for AI, ML and content, let’s look at some key considerations for organizations considering a content management platform with enterprise-class AI and ML capabilities.
Though AI is a hot topic and gaining adoption in commercial and lifestyle spaces — read this Gartner Hype Cycle Report for the latest — it is still a relatively new frontier. Regulation and compliance best practices still trail the technology. Organizations need to assure proper governance for AI/ML is place.
AI capabilities require data governance for several reasons, including to protect sensitive information, address ethical considerations, prevent biased results, identify and mitigate risks, and manage entire data lifecycles properly. Mishandling of data by unregulated AI sources may lead to inaccuracy in reporting, data breaches and noncompliance with governance data protection regulations, ethical concerns such as lack of transparency and eventually, a negative public perception.
When onboarding AI into an enterprise content strategy, the platform should allow teams to:
We also recommend an enterprise content platform that has extensive experience in handling highly sensitive data while adhering to differing industry standards. Enterprise content platforms with stringent data governance policies help protect sensitive information and ensure compliance with data privacy regulations such as GDPR, CCPA and HIPAA, which are enforced in sectors like government, healthcare and the financial industry.
Another critical consideration is how your AI models perform over time.
First, you should consider solutions that utilize continuous training paradigms that enable your ML models to evolve and improve over time as new content and data is added to the system. Human interaction with machine-generated data is also critical to provide data validation and further train ML models. Look for a content solution platform that considers the human role in the machine-learning process and provides specific interfaces for “human in the loop” training.
Your AI solution should also provide real-time performance monitoring for models. ML models can begin to show bias or even degraded performance; therefore, a performance monitoring interface will help identify models that have become corrupted or are showing degradation in performance. Machine-learning models should also be versioned, allowing you to quickly roll back to an earlier version should your model become degraded.
Enterprise content platforms are a uniquely powerful place to embed AI capabilities. Modern platforms unite content across apps and departmental silos, and with AI, the content gains relevance and usefulness. With AI and content, enterprises can capitalize on next-generation information management.
According to a recent Forrester study, 81% of respondents predicted that AI-enabled automation will meaningfully improve content-heavy processes by 2026. Indeed, 66% say they have already significantly evolved their content management approach due to AI.
The key to gaining meaningful information and insights while leveraging AI is to have clean, organized and accessible information in the first place. Augment your content services solution with powerful AI capabilities to derive more value from your content.
AI is transforming the world in unprecedented ways and we’ve yet to unlock its full potential. Partner with a company who understands the value of your content and processes in today’s dynamic and complex environment and who can prepare you for the opportunities that come with AI.
At this point, you might be saying, “This is all great, but how do I get started?” We’d like to help.
Hyland’s intelligent content solutions include cloud-native, low-code, AI-enabled content platforms and solutions that enable you to optimize information management and empower smarter business predictions.
Hyland’s solutions offer intelligent capture, data extraction, process automation and content management — all of which are integrated into unified platforms and enhanced with the latest AI-enabled technologies.
By automating manual tasks and extracting valuable insights from unstructured content, Hyland’s content solutions help accelerate decision-making, enhance customer experiences and achieve operational excellence.
Read more about our services, and how we’re making it easy for customers to train and manage their own custom ML models.
Discover how AI-powered document processing works, as well as its benefits, applications and future potential.
The combination of RPA and AI in intelligent process automation transforms the way organizations operate by enhancing efficiency, improving decision-making and streamlining processes.
Explore how artificial intelligence (AI) and machine learning (ML) work alongside humans to get the most value out of enterprise digital asset management.
Gain a cloud-native digital transformation strategy dedicated to better customer service — and smarter, stronger, faster growth. With Hyland, you can reach your full potential.
