Heretto: Using AI tools to bring structured technical content to a whole new level

Photo by Andrea De Santis on Unsplash

Almost no one would dispute the fact that Artificial Intelligence is expanding and evolving on a daily basis. AI is already changing how content developers approach their tasks, and is creating a lot of concern about the future of human technical content developers.There are lots of technical writers worried that AI is going to replace them altogether.

At the same time, for many organizations producing technical content, the process is still remarkably full of tediously manual tasks, incessant copying and pasting, and constant struggles to optimize content for search engines and potential reuse. The advent of structured authoring seemed to be the answer, but over the last few years, adoption of the standard for technical structured content, Darwin Information Typing Architecture (DITA), has been slow–generally limited to larger, more sophisticated content teams. Many organizations have shied away from implementing DITA due to the perceived learning curve and potential costs associated with DITA solutions.

IBM, the original developer of DITA defines it as follows:

The Darwin Information Typing Architecture (DITA) is an XML-based, end-to-end architecture for authoring, producing, and delivering technical information. This architecture consists of a set of design principles for creating “information-typed” modules at a topic level and for using that content in delivery modes such as online help and product support portals on the Web.

Sounds like a simple, commonsense concept that should be particularly well suited to content development right? Write a piece of content once, tag it in a way that describes what it is within a standard structure. In fact, structured content that relies on DITA XML has already proven to provide organizations with four key benefits:

  • Enhanced efficiency
  • Improved accuracy
  • Greater collaboration among team members
  • Production of content  that is machine readable for integration with multiple tools and technologies

This can be tough for any organization, especially smaller businesses with fewer resources, because:

  1. Developing a structure that can accommodate unique needs inherent in any organization’s product or service offering requires a significant investment of time 
  2. Most organizations have a vast amount of unstructured content already stored in a variety of systems. Inventorying, consolidating and converting that content is a daunting challenge. 
  3. Investment in a platform and a set of tools to manage structured content, requires more time and money than many companies have allocated.
  4. Many content teams find the tools to create DITA content hard to understand, and more difficult to master. In an environment where everything needs to be faster and perfect, this is hard to accomplish.

Some in the industry recognized the potential of DITA, and the potential of AI to take on some of the difficulties associated with it, such as analyzing and tagging existing content and making recommendations for usability and search engine optimization.

Heretto, one of the leaders in structured content development platforms, has beta launched its first generative AI tool, Etto the Heretto Copilot. Their goal: make structured content implementation approachable while boosting productivity and delivering accurate and compelling content, leading to better customer experiences. 

Heretto believes that content developers need not fear for the future jobs, because AI tools can enhance their roles at providing best in class content experiences that organizations need to remain competitive. 

The team at Heretto developed Etto to work with the Heretto CCMS solution, and its forthcoming HelpAI product to support the Herretto Portal  solution. Both are based on the premise that AI using large language models (LLM) relies on content to learn, and that structured content improves the process enormously. 

At a recent presentation at Lavacon, CEO Patrick Bosek and Senior Solutions Engineer Jarod Sickler noted that structured content is all about “metadata and connections.” Etto, generative AI trained in DITA and the rules governing structured content, can assist content developers, both experienced and beginners, by using metadata and making connections to automate time-consuming tasks and guide developers as they create content by answering questions, providing best practices as well as creating, revising, and refactoring content.

For example, this Etto demo walks users through common tasks such as:

  • Create a populated table
  • Insert a short description
  • Convert a paragraph into a list
  • Make a paragraph more concise
  • Adjust the tone of your content
  • Fix a validation error
  • Find and replace within a doc
  • Create a new table or list from pasted content

In addition, it can suggest additional content and answer product questions to jumpstart developers’ research tasks and address more complex issues.

screenshot of Heretto workspace and Etto Copilot
Etto performs complex content analysis, identifies inconsistencies, and recommends solutions.

Bosek and Sickler recognize that AI tools such as Etto are not fool-proof, because “AI is still in its infancy–it’s a black box.” Which is why they remain optimistic about the future of AI in creating technical content … the human element is needed to review the content and identify accuracies or inaccuracies to train the tool to improve not only the content, but also the guidance it provides.

Etto is currently in limited beta testing, and organizations interested in exploring the Heretto CCMS and Etto Copilot can join a waitlst, and schedule a demo.

As part of its plans to provide an end-to-end, AI supported structured content solution, Heretto plans to launch the beta of HelpAI, an intelligent agent that answers questions for end users based on content within Deploy API, in early 2024. Integrated with the Deploy API, HelpAI will be able to identify content in the system that is personalized for specific users and return only answers applicable to the user. Its features will include a confidence score for how well it believes it answered the question as well as sources to what information it used.

Producing technical content that is always accurate, maximized for reuse, and compliant with both the rules for structured content, and the industry and government required regulations, remains a challenge. As users consume ever increasing amounts of content, delivering an exceptional experience for them AND for the developers of that content remains paramount. Heretto’s new tools, used on its existing DITA platform show a lot of promise.


Heretto is an advertising partner of TechWhirl. This article was written under TechWhirl’s Editorial standards.

Subscribe to TechWhirl via Email