The data and AI asset class
1. The data and AI asset class
I've set up the explanation of how we monetize data and AI. Why is it different from other technologies?2. Shifting paradigms
If you look at code, the digital or software paradigm, you write code, and the digital product supports a single implementation. Digital products can sometimes be bent to a couple of different use cases, but most are built to support a specific workflow and a particular set of use cases. Digital product functionality is very narrow and can only be monetized once. A few digital products can be adapted and monetized once or twice more. Data is entirely different and supports multiple implementations.3. Data is not used up
Think about a data set. If I use it for a report or a visualization, the data set is still there. It's not been used up. Data is still good after you've used it in one report, 5 reports, or even 500 reports. You can use it to visualize as many different concepts as the data is relevant to. You can train as many models with that data set as you want.4. Data is a novel asset class
The critical concept is that data and AI are a novel asset class, so we must think about them differently to monetize them properly. We know the code and algorithm aren’t what's monetized. It is the data itself, and the data set’s value is connected to this construct that the business can monetize it repeatedly.5. What makes data valuable?
High-value data combines high-value use cases and the total number of use cases the data set can support. Again, we must return to the workflow to define high value. Suppose the workflow, as defined through the value stream, delivers a significant amount of value to customers or creates a significant amount of value. In that case, the data sets it generates will be high value. Some data supports multiple workflows. Several sales and marketing workflows have overlaps that even extend to finance. The domain knowledge concept is an easy way to frame this. Sales, marketing, and finance have domain knowledge overlaps that allow people to run those parts of the business or complete workflows critical to their roles. As a result, data gathered from one workflow will have applications in several other workflows. Businesses have data sets that multiple workflows can leverage to create or deliver value to customers in new or optimized ways. The use cases are connected by common domain knowledge. We’re also creating a connection between opportunities and implementations.6. Data is never tapped out
Product-first AI starts with the opportunities business leaders see and customer needs. Those use cases connect business value, workflows, and opportunities to implementations focusing on outcomes. What outcomes can we drive with this data set? That's where monetization gets really interesting because a data set is never tapped out. If we can figure out another way to monetize it or another model that could leverage it to become more functional or reliable, then we have more opportunities to explore. We can monetize that data set again, and again, and again, and again.7. Understanding ChatGPT's success
Think about OpenAI’s GPT. Why is ChatGPT such a valuable tool that’s used by so many different types of people? It can support a massive range of workflows. The model can be repeatedly monetized. Why? It was trained on data that covers a massive range of workflows. All that information, or domain knowledge, is locked inside that data until we use a model architecture like Generative AI, in this case GPT, to unlock it. The LLM supports several different workflows, and as a result, it can be monetized repeatedly by appealing to a massive range of customer segments.8. Upcoming success
That’s why Microsoft will make a ton of money with Copilot. Microsoft can put Copilot into so many different parts of its business. OpenAI can monetize ChatGPT repeatedly. People keep finding new ways to use it to create and deliver value to customers or to improve their lives. It's amazing how many times you can monetize models that are built on high-value data. Remember, it isn't the code. It isn't the model. ‘We have no moat,’ was only partially true. It's the data. It is this understanding that monetization connects through value streams to workflows. By gathering data with that connection, business or customer context, businesses are better positioned to monetize data and AI.9. Let's practice!
Create Your Free Account
or
By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.