The Company

Microsoft is one of the world’s largest software companies. Founded in 1975, the company is headquartered in Redmond, Washington, with revenues of $198 billion in 2022 and 221,000 employees. Microsoft Office, one of the firm’s core product offerings, was launched in 1990 and has gone through many evolutions, including the 2011 release of its cloud version, Office 365 (later renamed Microsoft 365). SharePoint (released in 2001) is another critical offering from Microsoft and provides document management functionality. In 2019, Microsoft announced Project Cortex, an artificial intelligence (AI) initiative to enhance Microsoft 365. In late September 2020, the first releases from Project Cortex were announced, including SharePoint Syntex. In Oct 2022, SharePoint Syntex was renamed “Syntex” and at the same time its functionality was expanded significantly. That latest release is the focus of this report.

The Technology

Though labeled as a product, Syntex is actually many capabilities bundled under one umbrella. In our analysis, these capabilities are subdivided into four groups, each somewhat overlapping.

• Document processing
• Search
• Governance
• “Sorting”

The eagle-eyed will note that “sorting” is not an oft used or recognized term in information management, and we will return to that later. First, let’s look at the document processing technology in Syntex, as this functionality will continue to garner most of the headlines. In short, Syntex provides no-code intelligent document processing (IDP) that is easy to configure and manage (see Figure 1). Many customers will undoubtedly use the IDP functions both for simple workflows, such as automated receipt tracking, and more complex common workflows, at times in conjunction with Microsoft Power Automate (its BPM product). Syntex also provides integrated e-signature functionality and support for third-party e-signature tools such as Adobe and DocuSign. In fact, what is in Syntex today is good, AI-driven OCR and NLP, along with no-code workflow design and automation. It’s certainly on par with many third-party document processing systems.

At its launch, Syntex also comes with two pre-built models, one for invoices and another for receipt processing. It would be good to have many more pre-built models for generic use cases, and we expect to see this in the coming year. In addition to the two pre-built models come three “custom models” for unstructured, freeform, and structured document processing. These give the buyer the basic building blocks to design their own models. Again, these are on par with much of the competition. Note that there is a learning curve for those who want to create, train, and manage new custom models. This is to be expected of any such system, but a more comprehensive range of pre-built models, once available, will undoubtedly accelerate and stimulate the adoption and creation of enterprise-specific models.

Syntex also promises much-improved, impressive search functionality in 2023. For example, Microsoft tells us that natural language, machine reading comprehension and semantic search will be offered, along with the ability to search by concepts. In fact, Microsoft leverages its transformer-based natural language generative model (Microsoft Project Turing) and can generate answers and summarizations to search-based questions. In other words, the search functionality will go beyond keyword searching. The Syntex AI & Knowledge Graph tools analyze content at a generic and individual level to create pre-built answers, making for much better and more accurate connections between related content items and adding contextual elements such as dates, locations, and people to improve the search experience.

These are much-needed improvements to the SharePoint stack, as few current users are satisfied with the quality of the inbuilt search. For us, this is one of the essential upgrades Syntex brings to SharePoint. Still, as noted, this is not yet available but is expected later this year. As with any search engine, its accuracy will depend on the quality of the underlying data and data structures it searches across.
Though SharePoint has long been used for records management and general governance requirements, this has not traditionally been its strongest point. On the surface, Syntex does not appear to offer much new in terms of boosting enterprise governance. Still, looking a bit harder, we think Microsoft should make more noise about the pre-defined retention labels introduced a year or two ago in Project Cortex. Retention labels are used to “retain” the data and documents you need and to eliminate those you no longer need to keep. It’s a shockingly simple concept that is not new, but few documents have retention labels applied to them today. With Syntex, you can apply a retention label to a model and automate the entire process. So, if a model (a custom contract process, for example) has been applied to a library, you can apply the retention labeling to that library and ensure that everything that needs to be retained for X years is retained. Documents not fitting the model are not retained. This is simple stuff and not unique to Syntex, but if used well it will improve governance and compliance situations and dramatically reduce storage costs.

In our four sub-groups, we listed “sorting”; it’s not a technical term, yet at its heart, Syntex depends and runs on the accurate sorting of documents and data points. Like the sorting hat in the Harry Potter books, Syntex automates the analysis, sorting, and tagging of incoming and historical data. Once a document or data point has been analyzed and understood, Syntex utilizes a Knowledge Graph to create and visualize connections between different data points. It is the core of Syntex and was designed and built based on decades of knowledge and experience regarding the use and misuse of SharePoint.
Like other document management systems, SharePoint relies heavily on users to proactively tag, manage, and govern content – work that nobody wants to do, so it seldom gets done. Automating these unloved but foundational tasks using AI is critical not just to Syntex but to the future of SharePoint, as far too many implementations are little more than unmanaged and unmanageable repositories. Using Knowledge Graphs takes this automation to another level and, frankly, one that other competitors will have to follow at some point. The traditional use of relational databases in document management systems focuses on storing data, with analysis a secondary affair. Knowledge Graphs focus on the connections between data. We are seeing that happen in the broader KM and Enterprise Search world. However, it’s still a novel yet powerful approach to document management. Syntex automatically sorts and connects data, which is the standout element that everything builds around for us.

Beyond these four groups, Syntex brings a wealth of new functionality, from document annotation and image tagging to automatic summarization and translation into many different languages. These tools allow users to automate the creation of standard documents such as statements of work or contracts. All good new features, but the one that got our attention was document assembly. It works by analyzing an existing document sample and, using that as a template, automatically generating new content and documents following those guidelines. Again, this is a Syntex element that has a lot of promise for creating internal documents or in the bulk creation of customer documentation. We say it holds promise, as in this release of Syntex the assembly template is limited to one document library and cannot be easily transported or used elsewhere.

Finally, even though we are clearly impressed with what is a massive upgrade and new functionality for the SharePoint stack, Syntex’s challenge is a universal one in the industry: the lack of skilled resources. Even in low- or no-code environments such as this, generic information, process, and business analysis skills and awareness are required to take full advantage of something like Syntex. There is a shortage of these skills today, though Microsoft does provide extensive end-user and partner training for Syntex. It’s the broader dearth of good information management professionals that is challenging. That’s not Microsoft’s fault, but it is probably the single biggest challenge it will face to gaining adoption.