LNETM Digest: State of the “Union” — Knowledge Graphs in the Enterprise
Michael Atkin is the Director of the Enterprise Knowledge Graph Foundation
Solving the data integration challenge emerged as one of the key themes from the London Enterprise Technology Meetup virtual session on semantic technology and knowledge graph. The event, hosted by Oli Bage, Global Head of Business and Data Architecture at Refinitiv, went a long way toward illustrating why this capability should be included in the organizational toolbox for every knowledge worker.
The feature presentation was by Ruben Verborgh, Professor of Decentralized Web Technology at Ghent University in Belgium. Ruben is supporting the Solid ecosystem — a new project started by Sir Tim Berners-Lee — to pioneer ways of giving people control over their own personal data. Solid maintains that people should be free to share “whatever they want with whomever they want.” The key to this mechanism of control is a personal data “Pod” based on semantic technology and open data standards. Think of the Pod as a form of data vault where the individual, rather than a platform like Facebook or LinkedIn, has control over the content.
Ruben explained that control over the data is the key to both identity protection and business innovation. In today’s world of data silos, there is “a single market for applications based on who owns the data.” Those who win this data harvesting game are not incentivized to innovate because they don’t have to. They are building functionality for the average user that doesn’t exist — and preventing new companies from entering the market solely because they don’t have the data. Solid is changing that paradigm; they are seeking to create a data marketplace that facilitates competition between different data providers — all based on open standards. A solid idea to be sure.
The case for knowledge graph is not just about privacy protection. The focus for Dow Jones was to make it easy for customers to integrate news and business information into their internal environments while also providing a contextual resource for their journalists. According to Dylan Roy, Senior Engineering Manager — the Dow Jones knowledge graph was the mechanism used to implement their CEO’s goal of “providing readers with the facts they need in this age of misinformation.” Dylan went on to conclude … “this is the holy grail of a publishing enterprise.”
Ben Gardner, Solution Architect at AstraZeneca, expressed a similar view. Their initial goal was simply better data integration. They are emphasizing the adoption of FAIR data principles to ensure that data is findable, accessible, interoperable and reusable so that their R&D users can leverage artificial intelligence and machine learning to drive insights. Ben sees this as an opportunity to move away from the application-centric way of operating to help their scientists navigate content and explore the concepts they need for pharmaceutical discovery. According to Ben, there is significant value in using semantic technology to help users “escape from the tyranny of extracting information from many sources and overcome the struggle to stitch it back together in meaningful ways.”
Marit Rødevand, CEO and co-founder of strise.ai has a mission to fundamentally change the way business executives think about enterprise software. She points out that companies around the world spend hundreds of billions of dollars every year in consultants and manpower just to integrate standard CRM and business intelligence software into their environments — and that most of these integration projects fail because the relational data model is so outdated. Strise has constructed a massive knowledge graph of information (40 million entities and 100 million nodes) about how the world is connected that helps banks, legal institutions and consultancies manage risk and comply with complex regulatory obligations. The underlying message, however, was one of simplicity because of the “beautiful new paradigm” of knowledge graph — structuring data and relationships in a way that lets users ask conceptual questions based on how they think. According to Marit, “semantic technology is future proof” because there is no need to restructure data, shift through large numbers of searches or hire armies of consultants to provide integration customization.
Jeremy Posner, a principal at the knowledge graph consulting firm agnos.ai has been my partner in the Enterprise Knowledge Graph Foundation since the beginning of this year. As a senior data practitioner in banks over the past two decades, he offered some practical advice for those just starting their knowledge graph journey with the goal of helping companies get over the “semantic hump” that can inhibit adoption. A key success factor is to bring together the business user owning the problem, the data champion who understands how knowledge graph can help solve the organization’s challenge, and the technologist(s) that would own the platform or represent the innovation group — as each has different perspectives and requirements. He has been an advocate of the enterprise knowledge graph (EKG) as “operational infrastructure” — a connected inventory of the business, data, technology and organizational assets. He is right. The inventory of what runs your business is a baseline requirement that leads to many solutions including risk mitigation, resiliency planning, and identifying cost-saving opportunities.
Jeremy explained that the development of the EKG maturity model (EKG/MM) defines the requirements for sustainable knowledge graph initiatives and embodies key design principles. This helps to define where you are, your roadmap and next actions for your EKG platform. The bottom line is that there is a tremendous amount of activity raising knowledge graphs up into the enterprise technology landscape, helping us all to finally manage data as a true business asset.