In the past, we’ve discussed the importance of adding quality metadata for several reasons. Of course, better metadata will improve all aspects of search performance. Also, the right identifiers and tags can provide you with business intelligence, audit support, ideas for extra revenue streams and more. The real gold mine is in converting unstructured data into structured data.
Here’s a quick one-minute primer on what exactly metadata is:
Adding this additional information to describe your documents may seem like a big job, and that’s certainly true if you don’t use the best tools. Not to fear. Today’s artificial intelligence will help create and add better metadata with less effort. Plus, it will work with all types of files, including text, graphics, audio, and video.
How Intelligent Document Management Systems Improve Metadata Creation
Despite the obvious benefits of having high-quality metadata, the task of adding it to countless files and records might seem like an incredibly time-consuming and tedious process. You may not have the manpower to comb through hundreds of thousands of text, image, and video files to select relevant keywords. Even if you do have the resources, that might not seem like the best use of your people. Advances in AI and machine learning can minimize human effort and produce excellent results and in that way, AI is finally living up to the hype.
How Does AI Work for Metadata Creation?
This list provides a basic overview of the types of AI technology that systems use to help with metadata creation:
Statistical learning: This technology relies upon statistical models to help divine important information from large sets of data.
Neural networks: This kind of tech finds patterns by sifting through information with neural networks that are designed to work a lot like organic, brain neurons.
Deep learning: These advanced systems can sift through layers of information to extract meaning, patterns, and comparisons.
AI Can Extract Metadata from All Kinds of File Types
In the past, people associated indexing mostly with text documents. Modern AI isn’t just limited to text files. Case in point… Las Vegas face recognition can identify known cheaters and card-counters in video images. You may have also seen examples of this on popular social networks. Like when Facebook knows that family reunion photo has Aunt Wanda in it and suggests a tag. Language processing can extract meaning from speech in audio files. Combining various techniques will also extra tags from video files.
Thus, you can use AI to help create and add metadata to text, graphics, and video files. For instance, today’s search engines can index and categorize .MP3 and .JPG files as well as .HTML and .PDF files. An intelligent information management system can do the same thing inside of your organization.
Consider some examples from CMSWire of using intelligent systems to categorize various types of files:
Images: The healthcare field has relied heavily on image recognition technology for all sorts of medical scans. Other industries can use this tech to help categorize scanned documents, including handwriting. If you have deposited a hand-written check in the ATM, you have probably seen this kind of image recognition at work.
Audio: Common examples of intelligent speech processing include Amazon Alexa and similar home systems. You have probably also used voice-to-text to compose text messages or request searches on your mobile phone. This same technology can find patterns in your company’s audio recordings.
Video: Analyzing video files combines the AI tech that’s used to process images, text, and audio. For example, you might tag everybody at a meeting by using facial recognition of a recording. Similarly, you may set time indexes of a video to make it easier to find the exact moment when a certain topic got discussed.
Humans Still Make AI-Assisted Metadata Creation Better
AI can help reduce effort and, in some cases, improve the quality of your metadata. Mostly, intelligent systems can make projects possible that you may lack the time or funds to accomplish quickly if you had to do them manually. Even better, these systems learn as they work, so they can provide increasingly better and more useful results over time. Since the machines never get tired or bored, they can also help minimize and eliminate the kinds of mistakes that people are prone to making.
Here’s a simple example: How fast could the fastest data worker look through 500 documents to find instances of social security numbers and then tag those documents as sensitive? Maybe a few days. Intelligent information management AI can do it in minutes, if not seconds. That’s the kind of power we’re talking about here.
You should still involve various stakeholders to determine which kinds of metadata you need, in order to create rules within the system and verify results. You can use these rules to help direct both the intelligent software and your quality control teams. Basically, the higher the risk of specific information, the more you may need to rely upon people to normalize the intelligence with governance rules and quality verification.
You might prioritize various kinds of information, so you can devote more time to the specific documents that carry the most value and associated risks. Also, you might start testing your smart systems with low-priority information, so both you and your AI system can learn to work together better.
See Intelligent Metadata Creation in Action
You don’t have to wait for future technology to involve intelligent computer systems in information management. Here at M-Files, we eager to offer you a free trial or a walk-through to answer your questions.
Artificial intelligence (AI) digital transformation refers to using machine intelligence to help solve problems and work more efficiently. According to Xerox, 46% of employees of small- and medium-sized businesses still waste time on inefficient, paper-related workflows, daily. That means that smart document management systems can do plenty to improve these processes.
Companies that already employ sophisticated document management systems gain a competitive edge. The good news is that these smart tools may help most when organizations haven’t relied much upon tech in the past and need to catch up. Even better, smart document management systems have grown accessible and useful for all sizes and types of businesses.
5 Vital Efficiency Boosts from AI in Document Management
In contrast, consider just a few revolutionary ways that using AI in document management systems like M-Files can speed up and improve the integrity of common business practices:
Automate manual processes. After scanning or entering the document, the AI system can automate a lot of the drudgery while enforcing good standards. Examples include using smart systems to help categorize and tag entries, so they’re easy to find next time. This ability can speed up the initial digitalization efforts and help with future tasks. As employees use the system, efficient searches often rely on correct tagging and categorizing documents, but this is one area that’s often neglected at the time the information gets entered. Even if the system doesn’t automatically tag and categorize, it can enforce preset rules to ensure employees do this extra bit of work when submitting new ones.
Speed up business intelligence: Smart systems can speed up business intelligence. For instance, AI can spot patterns in very large sets of data much faster than people can. As one example, a hiring manager might search through resumes with an intelligent system that will find keywords associated with successful people hired in the past or based on preset rules. Another example could include looking for seasonal sales trends from past invoices or orders.
Applying structure to unstructured data. IT people are used to seeing data organized into formatted databases or reports. Nobody can expect that kind of organization in most documents. Instead, communication often arrives in the form of emails, text messages, letters, or another organization’s forms. In the past, people needed to comb through these kinds of communication to pull out and organize information. Today, intelligent data management can do a lot of heavy lifting. As an example, the Adobe blog mentions a superb use case:
“In a more dramatic example, some companies are using AI and machine learning to scour emails, texts, and other customer communications to understand words, semantics, and sentiments, and connect that data with billing and service history to predict who will buy what products and services. Most remarkably, these models regularly outperform models that use structured data only.”
Streamline document preparation. Instead of having to teach employees how to correctly format various documents, the AI can simply request the information for unique, new documents. For instance, the system can automatically generate an employment contract for new hires, based on the employee’s name and other information it already has from the recruiting process.
Enhance document security. AI’s ability to secure information may prove one of its biggest benefits. The document management system can automatically scan for sensitive information and flag any documents that contain it. Even better, the software can detect unusual requests for private data and prevent access or alert security teams. According to Market Watch many of today’s cyber threats come from inside jobs, either intentionally or unintentionally. AI can offer a powerful weapon against all sorts of threats to datasecurity.
Improve data quality. The smart document management system can help reduce data redundancy, catch input errors, and keep files from getting misplaced. Businesses need good data for efficient decision-making and business processing, and that’s another feature that AI can supply.
Document-Based Digital Transformation Can Streamline Most Businesses
From legal teams needing to parse discovery documents to retail stores that must predict sales trends, AI in document management can improve efficiency in revolutionary ways. It can bring order to chaos, deliver useful insights, and even help prevent such common occurrences as data entry errors. To make the most of your valuable data, you should consider getting some assistance from a smart document management system.
In the technology solutions space, the Gartner Magic Quadrant release is a very exciting time. As a leading analyst report, the 2019 Gartner Magic Quadrant for Content Services Platforms (CSP) offers established vendors the chance to see where they stack up against their competitors. It gives new entrants exposure. It gives stalwarts in the space some guidance for the future of the market.
But, let’s not kid ourselves. While the Gartner Magic Quadrant gives us, the vendors, key insights, Gartner doesn’t develop it for us. The Magic Quadrant is for you. It’s an indispensable tool for IT leaders and reviewers of CSPs to develop opinions and shortlists of solutions that may help their business. It helps people like you find the features and capabilities in a vendor that are best suited to your digital transformation”>digital transformation goals.
As we mentioned in a previous article, M-Files is proud to be the most Visionary solution in this year’s edition of the Magic Quadrant. It validates our vision and our mission to enable companies and their people to work smarter, work faster and work decisively by unlocking the power of information.
But what many may not realize is that there is a companion report that underpins the Magic Quadrant – the Critical Capabilities report. It acts as the scorecard, quantifying performance in feature sets and use cases. That intel can be especially powerful to companies reviewing CSPs, seeking a well-balanced solution or a mix of certain capabilities.
The 2019 Gartner Critical Capabilities for CSPs Report outlines the fifteen critical capabilities with which Gartner graded all eighteen vendors:
One key takeaway from this year’s report is the parity between vendor scores. Aside from a few outliers, most scores tend to be pretty close, indicating a mature market and not one with tons of emerging technologies. Indeed, M-Files lives in a competitive landscape where the difference between solutions lies in the minutiae, in the details.
In examining, M-Files scored very well on the critical capabilities. In three of those, M-Files landed at the top of the heap:
Analytics and Reporting
M-Files scored the top spot for analytics and reporting, presumably on the heels of the investment in artificial intelligence and machine learning as an embedded capability of the solution – working to automatically identify and extract valuable data from documents.
An excerpt from Gartner’s description of this capability:
“These features enable users to discover insights about the content and data stored in the CSP. Artificial intelligence in the form of machine learning capabilities has become a major component for the delivery of these capabilities. Analytics and reporting deliver insights to end users. These insights can be extracted from content in the form of text, videos or images. They can also be delivered from the data that is inherent in many CSPs, including metadata and task-/workflow-based tracking. Greater emphasis has been placed in this report on the ability to utilize more advanced analytical capabilities using AI-driven machine learning to provide actionable dashboards.”
Metadata and Classification
M-Files scored the second spot with a 4.7 out of 5 for metadata and classification. We’ve always exclaimed that metadata is the most important part of a CSP, as it remains the force that drives almost every other capability. If an enterprise excels at classifying their information, a whole world of productivity and efficiency becomes reality, as metadata underlies workflows, permissions, searchability, creating relationships between documents and ideas, and so much more.
An excerpt from Gartner’s description of this capability:
“Metadata and classification defines the features that are used to associate metadata with content in a CSP. The ability to create and manage unstructured tags is fundamental here. More advanced features enable users to define structured metadata patterns, thereby enhancing findability and the extraction of insights.”
M-Files took the third spot with a 4.5 out of 5 in content security. Data security is a top priority for businesses. There’s so much more at stake than just fines and penalties for a data breach. The organization’s reputation is at stake, and thus lost business. According to an IBM report, the global average cost of a data breach in the professional services industry is .5 million. The biggest contributor to this cost is lost business. M-Files makes content security a priority, protecting your data from internal and external misuse.
An excerpt from Gartner’s description of this capability:
“Content security capabilities enforce controls that relate directly to the protection of content. Content security is a key capability that is essential for organizations which prioritize privacy and security of the content they store in the CSP.”
Let’s get real. What office worker hasn’t heard of Office 365? Unless you still use a typewriter, you probably spend a great portion of your day in Microsoft Office applications. There are, after all, about 155 million active business users of Office 365, making it the most widely used cloud service by user count.
55% of businesses say that the biggest ongoing issue for Office 365 is persuading users to manage and share content in Office 365 and not elsewhere.
Another key issue is how to connect all relevant data, be it structured business data, or unstructured documents.
The Sensitive Document Problem
According to the AIIM 2018 Industry Watch Office 365 Revolution Impact on Governance and Process Automation, 46% of an organization’s sensitive documents are contained in Office documents. The unstructured documents make it hard to manage business critical and sensitive data.
And, to make matters worse, it is estimated that as much as 90% of all data is unstructured, and that unstructured data is stored in at least four different content repositories.
M-Files to the Rescue
According to feedback from our customers, there are five main challenges they face with Office 365 implementations.
Permissions are role-based and easily managed with metadata. Archiving is automated with AI-driven tools. Content outside of Office 365 can still be managed from the familiar Office interface, e.g. Teams. Best practices are enhanced with automated workflows and processes. And M-Files can help you gain access and edit content from different places in one common view.
Bridging the Gap between Structured Business Data and Unstructured Documents
Our unique approach manages information based on what it is rather than where it lives. M-Files connects several business systems and external repositories and allows you to manage information from any repository.
You can easily get a 360-degree view of your business without the pain of migration. Rather, you can gradually migrate the most critical data to M-Files and manage it there. And additionally, allow people to access it through their daily tools.
3 Steps to Increased Efficiency and Convenience
M-Files for Office 365 approaches the challenges of information management on three levels, making office work both more efficient for the business, and more convenient for employees.
Free IT from the pain of migration and integrations. There’s no need for migration as information can be found, accessed, and managed regardless of its origin.
Let people use their familiar daily tools even for document management. Employees can use their daily tools — like Teams or Outlook — to access and manage all documents and data.
Let people focus on value-adding tasks. Employees do not need to waste time looking for information or tagging it as AI can eliminate routine tasks.
It seems we may still be a few years away from feeling the real impact of AI — and a more realistic assessment of the value of AI. Forrester Research VP Srividya Sridharan says: “We believe 2020 will be the year when companies become laser-focused on AI value, leap out of experimentation mode, and ground themselves in reality to accelerate adoption.”
There are several AI trends on the horizon this year for IT leaders — and business leaders, at-large — to follow.
AI will be the sexy, new career path for IT professionals
There will be 133 million new jobs created by AI by 2022, according to the World Economic Forum’s (WEF) 2018 Future of Jobs report. The share of jobs requiring AI skills has grown 4.5 times since 2013. And for those worried about automation and it’s net effect on job displacement, by 2020, AI will eliminate 1.8 million jobs and create 2.3 million, according to VentureHarbor. Hiring for artificial intelligence pros of various titles has increased 74% annually over the last four years, according to LinkedIn.
All of these statistics speak to a new demand in the marketplace for AI and machine learning (ML) professionals — a demand that will spur up-and-coming IT practitioners to focus on AI as a main course of study.
Privacy and governance will come front-and-center in the AI conversation
As far as data governance, 2020 will be about operationalizing AI, making governance a top priority. Forrester states in its 2020 AI predictions report that the problem for AI lies in “sourcing data from a complex portfolio of applications and convincing various data gatekeepers to play along.”
The focus will shift from simply having AI to measuring the impact of AI
Having AI for AI’s sake is passé now. Companies have made the investments and now they will use 2020 to figure out if that investment will bear fruit. “’When you do AI right, it generates value and ROI for the enterprise’ is an excellent premise, however the full potential of AI hasn’t been attained,” says AJ Abdallat, CEO of Beyond Limits. “Many conventional AI systems are merely machine learning, or neural networks, or deep learning. They’re good at handling large sets of data but lack situational awareness or the ability to navigate around missing or incomplete data. They get stuck.”
AI research will reduce speed
No one thinks that the research will come to a screeching halt, but the question is: Can it continue at the steep pace it has been? The demands that AI places on data and processing power may be too much to scale AI objectives, in some use cases — think self-driving cars. We’re still a way off. In an interview with Wired, last month, Facebook AI VP Jerome Pesenti said deep learning research may “hit the wall.”
Pesenti stated, “Clearly the rate of progress is not sustainable. If you look at top experiments, each year the cost is going up 10-fold. Right now, an experiment might be in seven figures, but it’s not going to go to nine or ten figures, it’s not possible, nobody can afford that.”
M-Files has pushed a lot of chips into the middle when it comes to AI in information management. We can verify that AI has risen to the hype, at least in our sector. It will remain a prominent focus of the IT community for years to come, no doubt. IDC figures that by 2025, embedded AI functionality will be incorporated in at least 90% of new enterprise application releases. To temper a bit though, IDC says that comes with a caveat: Truly disruptive AI-led applications will be only about 10% of this total.
AI ain’t going anywhere. But we’re still in a nascent stage, with lots of exciting innovation and healthy disruption to come.