On Enterprise AI-ready platforms. Q&A with Drew Wanczowski

Q1. What are the new challenges enterprise face when using AI and Gen AI?  

Today’s enterprises are savvy when it comes to AI. There are initiatives going on in many departments and lines of business. They are creating applications that leverage AI in a multitude of ways. Some examples are scientific research assistance, content creation such as summaries or translations, and validating tax laws and processes.  

When we engage with our customers, we see challenges around data silos, lack of internal context and issues with data readiness. Securely sharing and auditing data is also a major challenge.  The intellectual property and information that these enterprises manage are highly valuable. Whether it is content being produced by media outlets, new product development in pharmaceuticals and manufacturing, financial data, or classified intelligence information.  It is high stakes if any of this information is leaked or produces wrong answers due to hallucinations or data biases.  

Q2. How would you define an AI-ready platform?  

At Progress we look at AI across the whole stack. We have been introducing AI products within our infrastructure, data management, and user experience products for some time now. 

Focusing on the data management side it is paramount to have a secure and auditable workflows that prepares content for AI. AI is not perfect, so it is important to still have a human in the loop to validate that the processes and insert their domain expertise.  

An AI ready platform should have processes to connect to data silos integrate and prepare data, add context and metadata to capture knowledge, ensure business rules are being met, all while recording its lifecycle and keeping the data secure and auditable.  

Q3. Looking back at the past year, what were the most significant advancements and key milestones for the Progress Data Platform?  

Over the past year the Progress team has been working hard to provide our customers with features to enhance data management in flexible ways while ensuring domain knowledge is captured.  

• Progress MarkLogic has introduced features for native Vector Search, Virtualized Views for ad-hoc data retrieval, increased scalability and enhanced security posture. 

• Progress Semaphore has introduced tooling for better Knowledge Graph and Modeling including additional metadata standards support, AI assistance for creating Knowledge Graphs, and scalable semantic integration features.  

• Progress Agentic RAG makes use of structured and unstructured data to ensure business context is included with your AI integrations.  

• Progress OpenEdge has also introduced features to utilize your trusted transactional data with AI Agents.  

• Progress Corticon offers AI assistance for building complex business rules. 

• Progress DataDirect has continued to offer a vast library of drivers to provide data to these AI pipelines. 

Q4. Could you elaborate on how your philosophy of use-case driven innovation and constant adaptation is shaping the platform into a more unified, multi-model solution?  

One size does not fit all in many cases. Enterprises have data in various shapes, sizes, and quality. Utilizing a single model and storage paradigm becomes difficult and unmanageable. With the flexibility of Progress Data Platform, we focus on use cases and business outcomes first. Being a unified platform, we allow our customers to focus on the task at hand. This allows them to navigate projects in a more efficient and timely manner.  

A recent customer has over a hundred years of scientific research data in hundreds of formats. This is a challenge. By focusing on the use case of developing scientific research assistance we were able to unify this data, provide intelligent context that their subject matter experts have encoded in the system, and finally provide AI ready data to answer their research questions. As part of this journey, we partnered with this customer to ensure that their feedback has made it onto the roadmap.  

This would not have been possible without a unified multi-model solution. Gathering and organizing all this data and preparing it for AI would have been a multi-year project with traditional tooling.  

Q5. In your experience how organizations use multi-model databases?  

Most of the information produced by the enterprise is unstructured. 

IDC, a global market intelligence firm, states that 80% of all information generated is unstructured. 

Unsurprisingly, Forrester says 80% of new data pipelines are built for ingesting, processing and storing unstructured data. Traditional systems struggle with unstructured and multi-model content. The ability to load data as is and handle data in varying formats, schemas and sizes is key.   

Organizations use multi-model databases to support items that don’t traditionally fit into rows and columns. Records such as XML, JSON, PDFs and Office documents are just a few formats that are managed in the Progress Data Platform. 

In addition to capturing data, it’s equally important to capture domain knowledge, which is often achieved through the use of knowledge graphs. Knowledge graphs are comprised of interconnected objects and facts. For example, the statement High-Fructose Corn Syrup is a Sweetening Agent expresses the relationship between a subject and an object. A knowledge graph also reflects the world around you. It can mimic business processes, real-life physical objects or even domains of study.   

Adding a semantic layer to your system enhances information classification, organization, and discovery. When systems can understand your business concepts and terminology, the impact is significant. For instance, an R&D team may refer to a product by its code name, while the marketing department uses its commercial name. A knowledge graph unifies these perspectives by linking facts and harmonizing terms, creating a common foundation for collaboration and insight. 

Having the flexibility of these models allows you to use data in a way that is more natural to its origin.  

Q6. How multi-model databases leverage the inherent structure of data being stored?  

The Progress Data Platform ensures that you can consume your data in the shape and application you want. Whether it is through Google-like search, a traditional BI tool, ad-hoc SQL queries, REST APIs, graph visualizations, or data replication.  

This is achieved by leveraging a multi-model platform. Progress Data Platform allows you to load data in as-is and indexes the content as well as the structure. Because this happens within a single transaction, all views of the data remain consistent. This unified approach allows you to run both analytical and operational workloads on the same system. 

Additionally, with having both structure and data indexed you can discover information across whole records or select data with specificity.  

Q7. What strategies do you suggest for a multi-model database to fit into an existing architecture?  

Progress Data Platform is not meant to be a rip and replace solution. We suggest taking a use-case driven approach that involves bringing in data and incrementally building features. Given the multi-model nature of the platform, you can do this without starting over, thus delivering continuous value to your organization. We have had many success stories with this approach, including customers in verticals such as Healthcare, Insurance, Finance, Research & Development, Manufacturing, and Publishing to name a few.  

Additionally, you can leverage the Progress Data Platform to enhance existing applications. Knowledge graphs that include your domain expertise can be used to auto classify and harmonize data in other systems. You can also include this data in other applications through APIs or widgets.  

To support this connectivity is important. We offer tooling that make this seamless. Progress Data Platform includes data connectivity drivers, data movement utilities, and rich APIs & SDKs. 

Qx. Anything else you wish to add?  

A key takeaway would be that AI is a journey. It takes refinement and flexibility to get to desired results. In many cases adding AI tooling to applications and products reflects on your organization. You want to ensure you have the best foot forward with clean, governed, and traceable workflows. In highly regulated industries you will need to prove when and how you arrived at an answer. Giving the tools to your users and having a human in the loop is important, and should be considered when starting any project. 

………………………………………………

Drew Wanczowski is a Senior Principal Solutions Engineer at MarkLogic in North America. He has worked on leading industry solutions in Media & Entertainment, Publishing, and Research. His primary specialties surround Content Management, Metadata Standards, and Search Applications.

Sponsored by Progress.