There is widespread panic about a perceived shortage of data scientists in the financial services sector. A recent Financial Times article warned this talent gap is holding banks back. Marketing Week has highlighted the frustrations of chief marketing officers in trying to recruit for this role. The skills gap is challenging HR leaders in a similar way.
But what if data scientists aren’t who organizations should be looking for – or at least not them alone? A growing number of AI experts argue that what the business world really needs now is more knowledge scientists.
You may not be familiar with this fast-growing but still niche discipline. Nevertheless, the likely future importance of knowledge scientists – and the consequent need to recruit them – makes it vital that business leaders get to know them and what they do. With organizations increasingly betting big on generative artificial intelligence (GenAI), knowledge scientists could represent your best chance to get the most out of new technologies while staying on top of governance
Knowledge scientists are skilled at creating knowledge graphs. These are datasets stored in a graph structure, rather than a conventional format.
In simple terms, knowledge scientists prepare and structure data in ways that make it interpretable by AI systems. They also optimize datasets for the use of these models. And, in doing so, they provide a transparent route to AI – an increasingly important consideration with stakeholders worried about hallucinations, bias and other flaws.
Knowledge scientists are skilled at creating knowledge graphs. These are datasets stored in a graph structure, rather than a conventional format such as a table. They allow computers to manipulate the information based on its meaning. Knowledge graphs make it much easier to visualize connections between multiple data points. These points are usually referred to as “nodes” and the connections between them as “edges.”
Several industries have already begun to use knowledge graphs in their data management and analysis, with the potential to identify edges considered an essential aspect of generating insights.
In finance, for example, knowledge graphs are an effective way to discover hidden relationships between accounts and transactions. This is a vital aspect of anti-money laundering work. In retail, they can help e-commerce businesses identify customer preferences, allowing them to deliver more personalized sales and marketing content. Media platforms use the same techniques to make tailored viewing recommendations. The healthcare sector uses knowledge graphs to attain a better view of disconnected patient information and drug properties.
“Knowledge graphs can help AI to interpret the intent of content, rather than focusing on the precise words being used.”
The knowledge graph approach has many advantages. It provides a fuller picture of an organization’s data, making it easier to integrate multiple sources. It is also much more flexible and scalable than conventional formats.
These benefits can also provide a superior experience when used to harness the AI tools many organizations are adopting. Knowledge graphs can help AI to interpret the intent of content, rather than focusing on the precise words being used. For example, it could infer what someone conducting an internet search might be looking for, rather than just making a narrow reading of the search terms they’ve entered.
Combining knowledge graphs with other types of datasets could also support AI adoption in other ways, too. This may be a particularly useful approach early in the development of a new model, when there is relatively little information to work with. It could also support the reuse of training data in future projects, saving time andmoney.
Importantly, this discipline also lends itself to transparency. One of the concerns most often expressed about AI – and particularly GenAI – is that it relies on a “black box” approach. If we can’t see how a GenAI tool has arrived at a particular result, or even which data it has considered, how can we be confident that the result is trustworthy?
Knowledge graphs, by contrast, can help us interpret what’s going on inside the black box. They shine a light on connections to explain how and why the system has made classifications and reached specific conclusions. They can also be augmented with new nodes to address concerns about data governance – the imperative to comply with new AI regulation, for example.
None of this is to suggest that we don’t need data science or scientists anymore. Rather, the roles of data and knowledge scientists will be collaborative and complementary.
Data scientists have developed rich data management skills and are adept at resolving problems ranging from lack of data to the build-up of information silos. They have already helped organizations extract valuable insights. Their expertise has been well suited to stewarding the recent development of AI-powered analytics.
Knowledge scientists will be able to build on that experience and expertise, working closely with functions across the organization to understand specific requirements. The knowledge graphs they build will provide the data required to train and refine models for today’s AI tools.
Some will argue that the distinction between data and knowledge scientists is superficial. To some extent, that’s fair comment, with a degree of overlap between the two roles as they continue to evolve.
But the fundamental point is that AI technologies are advancing at pace. There is huge excitement around this but, to get the best results from the use of these technologies, organizations need people with the skills to fully exploit them.
Enter the knowledge scientists.
IMD’s TONOMUS Global Center for Digital and AI Transformation provides proprietary analysis and executive education focused on helping organizations enhance their digital and AI maturity. To discover more about the AI Maturity Index methodology or to assess your organization’s current capabilities, visit our information page here.
All views expressed herein are those of the author and have been specifically developed and published in accordance with the principles of academic freedom. As such, such views are not necessarily held or endorsed by TONOMUS or its affiliates.
Customer centricity often steals the limelight, but what about your employees? Anne Kawalerski, Michelle Lynn, and Elisabeth D Oak explore ways to ensure your organization attracts and retains the best talent in...
How do organizations recruit and retain staff amid a fierce global battle for the best people? Novartis’s Rob Kowalski explains how organizational culture can provide the winning edge in a competitive talent...
Breakthroughs that reshape business models can feel sudden, but they are rarely random. Innovation happens when you reframe challenges and opportunities, says Stefan Michel. The good news? You can learn to reframe....
Natalia Wallenberg of Ahold Delhaize describes how one of the largest retailers in Europe is tackling the upskilling dilemma....
Explore first person business intelligence from top minds curated for a global executive audience