After the oil spill comes AI to fill in the gaps for enterprises

A puddle of oil : DaaS

We are way beyond the time when the phrase “data is the new oil” was a common adage in tech. Instead we are in the age of AI, generative AI and more specifically, large language models (LLMs). But behind all the ChatGPT hype remains the basic truth – AI can only be as good as the data it has to operate on. For AI to succeed it needs data and computation, both to train and execute AI models. So let’s look into this dynamic and the reasons why CxOs need to now have Data-as-a-Service (DaaS) rolling off their tongues.

Data, data, data…

Enterprises possess oil tankers of relevant data in the age of AI, so they are off to a good start. Let’s look at the different kinds of data that are relevant for AI, where they sit and how to leverage them.

graph of AI figures

Data-as-a-Service use cases

For CxOs to leverage DaaS, there needs to be a new approach to the enterprise data platform. This is because conventional data architectures have unfortunately not been able to facilitate the successful operation of DaaS use cases (see figure two).

CxOs must ensure their DaaS architecture approach begins with an inherent understanding that there is no finite estimate of the volume of data to be stored. It is crucially important to recognize from the start that DaaS not only enables exporting and monetizing data, but also allows enterprises to license and purchase data.

For DaaS to work effectively, additional capabilities running on top of the Infinite Insights layer are a prerequisite. Prominent among these are understanding who the customers and suppliers of data are, how the data is licensed and how data can be shared and made available to the various parties involved in a DaaS scenario. A strong monetization offering, with an understanding of the licensing and consumption of the underlying data, may be required as well.

As data becomes a strategic asset in a digital economy, the platforms for managing data also become strategic. When building a DaaS platform, the define-and-design approach should follow these architectural components (see figure 3).

The virtual data layer is the uber data management component. Although there are ambitions for the universal data lake, such a lake will never encompass all relevant data for the DaaS automation needs of an enterprise. There are just too many potential and relevant data sources to import and integrate all of them into a data lake, and too many potential delays in doing so – especially when those delays lead to delivery time lags and possible business penalties. Therefore, every DaaS architecture should have a federation of data, ideally through a virtual data layer, that is powered by an elastic data mesh.

Services and APIs are the exposure aspect of the platform. Direct access to data is always problematic as too many things can go wrong (data theft, performance issues and more). Instead, a modern DaaS platform allows access to data via API-enabled service layers that take care of access rights and privileges, log usage and consumption and security.

Data security is an inherent platform characteristic. Security needs to be “baked” into a modern DaaS platform. The potential risks of data breach, data theft and other malicious data operations are too high for any enterprise to not have a built-in security layer as part of its DaaS platform.

Consumers use only APIs to access data. DaaS consumers access data via APIs only, ensuring better policy adherence with respect to data access, usage and operation. A key success factor for DaaS platforms is the ability to provision and monitor these APIs.

DaaS verticals

A modern DaaS platform can facilitate excellent vertical use cases. The following are among the most prominent:

Data is critical for AI, and DaaS is the modern way to manage data.

Artificial intelligence. The DaaS foundation determines the success of an AI strategy, and that determines the potential of an enterprise to lead in the marketplace.

Advanced analytics. Traditional analytics only looked at in-house data. A modern, advanced analytics approach makes it possible to use third-party data to enrich and validate the in-house data. DaaS is a key enabler for procuring, licensing and operating third-party data in compliance with contractual and regulatory agreements.

Benchmarking. To benchmark performance in different areas of the enterprise, IT leaders require a platform that allows them to put in place agreements with benchmarking partners and to mask and neutralize data as specified in the benchmarking contracts, facilitating the continuous export/import of data. DaaS is the platform to power benchmarking for enterprises that have so far been able to look only at their in-house data.

Customer data hub. Customer data is one of the most valuable assets of an enterprise, yet it’s highly fragmented and often is enriched with third-party data. With an effective customer data hub enabled by a DaaS platform, IT leaders can aggregate all of the customer data within the enterprise, enrich it via third-party data licensing and satisfy regulatory requirements.

Data brokering. Data is valuable, and enterprises may not only license and sell it but also may broker it against other data they are missing or want to enhance. Enterprises may also broker third-party data to their customers and suppliers. A DaaS platform makes this possible.

Data marketplace. To be able to offer data at scale internally, data marketplaces are the right strategy to pursue. A data marketplace empowers enterprise users to not only understand what data they may need to acquire or license, but also to continuously inject third-party data into their own data. On the other hand, by documenting second and third-party demand for their own data, a DaaS-powered data marketplace can help organizations discover what value their internal data may have, if monetized.

Data science. Data scientists cannot rely only on in-house data as the base for their artificial intelligence/machine learning models but also need access to other data they can use to validate, cleanse and enhance the existing data. DaaS platforms in combination with data marketplaces are key enablers for data scientists.

Fraud. Globally, enterprises lose hundreds of billions of dollars every year due to fraud. DaaS can help IT leaders bring in a far better capability for flagging and reducing fraud, and for mitigating its intensity. Fraud-associated use cases will gain further momentum as enterprises embrace DaaS platforms.

People data. Information about people is highly fragmented, highly regulated and often beyond the control of a single enterprise. When enterprises want to validate employment history or work references, they either need to work in a cumbersome manual fashion, or they can use a DaaS platform to gain better transparency into people data.

The takeaways

Enterprises cannot afford to not have an AI strategy in 2023. And while there may still be some time to implement AI correctly, its success will stand – and fall – with data.

Fixing and creating anything related to data is always a lengthy process. So it is crucial for enterprises to start looking at their data platforms now, as fixing them when the competition is making life hard with the power of AI might take too long and may well be too late for an enterprise to stay in business.

As in any case when moving and fixing something in IT, it is well worth it to look for the latest thing, in this case DaaS, to help manage an enterprise’s data platform. In other words, it’s no time to be wet behind the ears when it comes to data.