HYLAND VISION LEANS HEAVILY ON AI AND CLOUD

A need for modernization

Hyland is a market leader in the content management space having grown from a start up in the mid-1990s to now generating over $1B in annual revenue. Hyland’s initial focus has been on the mid-market, but it has gradually moved upstream. Globally, Hyland now has more than 14,000 customers, the majority of which utilize its homegrown OnBase platform. The ISV, which, since 2007, has been majority-owned by the equity investment firm Thoma Bravo, has also made a number of acquisitions, including the formerly competitive repositories mentioned earlier.

However, none of Hyland’s major acquisitions had a SaaS-first model and as a result the vast majority of its customers have on-prem implementations or utilize the Hyland Cloud, which hosts individual instances of OnBase and not multi-tenant applications. Experience was first announced as a SaaS platform at CommunityLIVE 2020, but it is still in the early adopter phase.

Last April, Hyland’s goal of transitioning its business to the cloud was cited as a major factor in the layoff of 20% of its 5,000-person workflow. In August 2023, Hyland brought in a new Chief Product Officer, Leonard Kim, who had spent the seven previous years at ADP, where he helped implement a hybrid cloud strategy. Then, this past April Hyland appointed Ghai as president and CEO. Ghai also has experience transitioning a business to the cloud as the former EVP and Chief Product Officer at Informatica. Ghai replaced the retiring Bill Priemer, who had been at Hyland in an executive capacity since 1997 and was appointed CEO in 2013.

Clearly, Hyland’s board felt some changes were needed to push the company forward.

Hyland’s IDP Strategy

Partially due to an early focus on document imaging applications, Hyland owns a significant share of the North American Capture & IDP market, even as many of its reseller partners have historically offered third-party products to provide more advanced capabilities. Over the years, Hyland has made some attempts to evolve its Capture and IDP portfolio, acquiring AnyDoc and Brainware, but the application breadth of those offerings has proven limited. To introduce more flexibility, last year Hyland signed an OEM licensing contract with German ISV Skilja for its IDP software, which Hyland has brought to market as Hyland IDP or HIDP.

HIDP is currently being sold as an on-prem, private cloud, or Hyland Cloud offering, and it seems to fill an important niche. The HIDP sessions at CommunityLIVE were well attended with an engaged audience, asking questions and even commenting on the shortcomings of Hyland’s legacy capture offerings. Hyland executives indicated they have a rapidly growing pipeline for HIDP.

When Hyland introduced HIDP, it was presented as transitory offering designed to bridge the gap between Hyland’s on-prem legacy and its cloud future. This still seems to be the case. For instance, when Hyland demoed its Experience Automate SaaS offering, it incorporated document extraction functionality from AWS, not HIDP, although Hyland indicated that it has plans to enable Automate to leverage IDP in the future.

Of course, HIDP consists of more than just classification and extraction engines. There is also a GUI workflow designer for setting up production applications. Interestingly, Hyland demonstrated how this designer could be used to call third-party services from Microsoft, Amazon, ChatGPT and others to incorporate their document processing technology. Yes, it seems like Hyland’s long-term vision for IDP is definitely hybrid oriented.

Broad Ranging Potential for IDP

Let’s discuss how IDP is incorporated in both of Hyland’s current major Experience SaaS offerings. As I mentioned earlier in the case of Automate, it’s not necessarily HIDP that’s being leveraged, but, at least initially, Hyland Experience is leveraging more generic IDP capabilities powered by hyperscalers and LLMs.

Automate is an intelligent workflow platform. Hyland has always been strong with document-centric workflows, but has not really competed with the likes of Appian and Pega when it comes to automating processes incorporating multiple systems and input sources. At CommunityLIVE, Automate was demoed in a customer onboarding application, in which data needed was captured from some identification documents.

 Automate is being positioned as a low-code/no-code workflow designer that, when combined with “Content Intelligence,” will be able to make suggestions related to the next step that should be taken in a process. Ghai indicated that as the company moves deeper into process management, it still wants to leverage its ability to manage and store documents as a differentiator. There is also an “Application Intelligence” layer that helps determine how content fits into enterprise applications. This should facilitate low-code/no-code integration.

Hyland Insight, which was made available in July to existing Hyland customers, provides a way to apply to AI to better leverage unstructured content being stored in repositories. It consists of three building blocks: Curation, Discovery, and AI Agents.

Curation is the first step in that it enables users to better understand what they have. This is a multi-modal capability that can be applied to more than 140 file types including documents and rich media. (This new trend toward application of AI for understanding multi-modal content input is reminiscent of Infosource’s long-time vision of addressing multi-channel input in Capture & IDP.) In the case of documents, full-text recognition is applied and the text is then chunked and parsed, with semantics being utilized to show the relationships between the documents.

The Discovery step can then be used to train AI-driven processes on an organization’s curated documents, which facilitates Retrieval Augmented Generation (RAG). IDP applications could potentially be trained with this capability but it’s worth noting that curation is similar an IDP application itself. This should give you an idea of the wide-ranging potential for IDP technologies going forward. 

AI Agents can also be created leveraging curated content to perform specific tasks within an organization. Hyland indicated that it will make a library of these tasks available that can be fine tuned and adapted to meet customers’ specific needs. Presumably IDP functionality would be included this library.

Regarding the AI models used for Insight, Hyland was adamant that it will not incorporate any public cloud LLMs and it will only utilize AI running in their SaaS environment. For this they are currently working with Claude from Anthropic and other open source models.

Summary

Hyland has a long way to go to realize its vision of evolving into a SaaS-first AI-driven business, but they have at least laid out the framework for accomplishing it. There wasn’t a ton of detail revealed as to how this framework will be filled in, but, then again, Hyland’s customer base was probably not ready to digest too much yet, especially for a platform that, early adopters aside, still seems at least a couple years out.

Hyland continues to have a strong customer base to lean on and seems to have resources to invest in development. Despite a couple of false starts as it has tried to launch a next-generation platform, all does not appear lost for the ISV. The new management team has laid down a solid vision that they now need to execute on to keep the company thriving into the next decade.