AI Strategy
AI Business Translation
Data science vs content science
Since the rise of Generative AI tools such as ChatGPT, Bard and Midjourney, the threshold towards using AI has been lowered. An evolution we evidently applaud, since it is our aim to make Generative AI more accessible for companies and, in doing so, help them use Generative AI in optimizing their corporate processes. With the increased access to these tools, we see a paradigm shift occur. In ‘Radar’ a Nexxworks podcast with host Steven Van Belleghem, technology entrepreneur Peter Hinssen addresses this paradigm shift from data science to content science. That’s why we zoom in on what this means and how MbarQ responds to it in this blog.
Data science vs content science
First things first, if we want to zoom in on the paradigm shift from data science to content science we need to explain both concepts first in order to understand what’s happening.
- Data science
“Data science is the study of data to extract meaningful insights for business. It is a multidisciplinary approach that combines principles and practices from the fields of mathematics, statistics, artificial intelligence, and computer engineering to analyze large amounts of data.” Bron: AWS
Roughly translated to a general workplace context: data and information that has been structured or organized in Excel files, some sort of Content Management System or other methods of collecting information such as e-mail subscriptions, lists of orders, invoices, … . The key takeaway is that it concerns data that is already organized in some way. All the data is organized in rows and columns with defined labels.
- Content science
On the other hand, content science or content analysis deals with the presence of certain words, themes, or concepts within some given qualitative data such as free texts, PDF’s or images. Applied to the workplace, this involves unhandled content such as written emails, unprocessed invoices, Microsoft Teams snippets, … Or in other words, bulks of unorganized data.
With regard to AI and Generative AI, the main difference is that conventional AI often was applied to organized data, whereas Generative AI based tools such as chatGPT, Midjourney or others have the ability to handle unstructured data. Ask ChatGPT to give you the three main topics in a plain text and he will be able to do so.
Peter Hinssen and the 20% – 80% gap
In Radar, a Nexxworks podcast hosted by Steven Van Belleghem, Peter Hinssen addresses the paradigm shift from data science to content science. To explain this, he uses a number of examples.
- AWS Healthscribe:
Peter Hinssen starts off with a little story: a lot of the professional time spent by doctors and health professionals goes to administration. After medical maintenance they need to update the patient’s file, write medical notes and fulfill other administrative tasks that keep them from doing their core business: providing health care to patients. If you look at this from a data perspective, you have a lot of unorganized data that needs to be handled. AWS developed a tool, based on Generative AI, that tackles exactly this problem, enabling medical staff to be more focused on their core business.
- Hollywood writers guild strike
A second example Hinssen highlights is the strike by the Hollywood Writers Guild, a guild consisting of 11,500 writers of TV and film. They are on strike, first of all because they are dissatisfied with the way they are paid and the major let-offs there have been, but secondly because they fear that the new box office hits will be generated by Generative AI tools such as ChatGPT. Why or why not people should fear the impact of Gen AI is something we’ll address later on, but this story is another perfect example of the way Generative AI is forcing a way into the workplace.
Whether you’re talking about the conversation between doctor and patient, or an entirely new script for Spider Man 3920, all these examples contain written content, aka unorganized data. This data, as opposed to organized data, entails 80% of a company’s data, with technological innovation providing generative AI tools to handle this information.
MbarQ and the data paradigm shift
However, this shift from data to content science provides a number of problems as well. More and more companies acknowledge that other or third parties have data from them in terms of emails, powerpoints or documents that are – as of recently – able to be organized and used. In other words, information and knowledge management will be really crucial and become more crucial in the future too, raising the need for players in the world of content. Peter Hinssen refers to Collibra, for instance, as one of the first companies to acknowledge this and create a market for it.
The main question will be: how do we manage content sources within a company and how do we use Large Language Models (f.e. ChatGPT) to organize this data. Aka, content science.
And this is exactly our core business. We sincerely believe that every knowledge process within a company and with regard to the 80% unorganized data can be augmented or semi-automated by Generative AI. The example of the Hollywood Writers Guild shows people are scared of the immense power of Generative AI. Whereas AWS Healthscribe suggests the opposite and shows the possibility of focusing more on your core business.
We’re out to reach exactly the latter: how can we help you optimize your corporate processes with ChatGPT, making your company operate more efficiently and focus more on your core business.
Use cases
Content science, or the shift towards content science, implies a new field of experience and therefore, new business possibilities as well.
- Example #1: Instruction manuals
Imagine a production company that has installers almost worldwide. They often need to look up certain handlings in instruction manuals of the products they sell. Usually, they call the support center for this. But that’s not scalable. For the more basic questions, they want a chatbot integrated in their corporate app where you can speak or type your question, and it automatically searches for the correct answer in all those manuals, and even provides the right tools and photos or whatever is needed.
- Example #2: Invoices
A travel agency employs 120 travel agents who arrange B2B trips. In doing so, they have a ton of emails, invoices and input to … one mailbox. That’s roughly 1,200 invoices a month that need to be paid and correctly booked. The company has several full time employees dedicated to this task daily. For less complex quotations, they now have a new mailbox that uses a GPT system. The system automates 80% of that mailbox.
We add to the paradigm shift
Our initial purpose was to make AI more tangible and feasible for companies by exploiting its enormous potential to make your business run more efficiently and effectively.The recent rise of Generative AI entirely changed the approach to AI. It is currently possible to use this 80% unorganized data. And that’s exactly what we do. That’s why we want companies to face this paradigm shift and acknowledge that AI isn’t merely for the technological innovators or early adopters.
It’s here. Use it.
And whether you apply a buy (f.e. MS co-pilot) or build approach in doing so, is something we’ll tackle later.
Blogs
AI Implementation
Smart Automation
AI Strategy
AI Implementation