Where to start?
To get the conversation underway, one member raised the point of data professionals being “unable to talk about technology and data without talking about the cloud strategy.” Over the last five years, it has become commonplace to experience multiple cloud strategies within the same organisation, whereas a decade ago it was the norm to have just one. This is partially due to the growth in cloud platform options, rather than the monopoly that was available when the technology first emerged, but also because of the way businesses have begun to utilise cloud technologies.
This then leads succinctly into the next issue that was raised by the participants: if an organisation has multiple big source systems or clouds, how is access to all of them approached? A participant explored buzzwords being used, such as data fabrics, data mesh and lake houses which can blur the line for accessibility for non-data professionals, which in turn slows the pace of investment in technology as new developments arise.
Elsewhere, there is the topic of business intelligence (BI) and the different tools that are now available, plus the focus on artificial intelligence (AI) in BI. How are these rapidly developing tools being embraced, and are they being utilised efficiently by organisations, or just added to the toolkit because they are the hot new thing? One member noted that they have been approached more recently that ever by other areas of the business to discuss AI’s use in platform strategies, automation, implementation in all business areas, cost takeouts and end-to-end.
Risk
There is undoubtedly a high degree of risk whenever a new technology in unleased, but the level of risk appears to be more evident with the development of generative AI tools as it is an uncharted world for most businesses.
Education is a core part of managing risk and, as one member mentioned, risk is a multi-departmental issue that can often cover multiple tools and storage locations. This can be costly and needs buy-in from multiple decision makers, which means storytelling as to why investing in a risk is necessary. As one member explained, “The reality is now we need to frame it in a price level, rather than ‘let us do a little pilot project on the side’.” As with most things in business, money talks, and data is not cheap – there is even a cost of extracting data from storage – so data teams must assess the risks and then assess the costs in order to make progress. “We are starting to have those conversations about how to fund it properly, how to set it up, and deciding where it lives,” said one participant. “There is a natural correlation between technology data, and then whatever the use case is.” This focus on correlation and use cases is key to receiving investment in risky new technologies and keeping pace with developing tools.
However, as one government-related member highlighted, trust in the tools needs to be examined and assessed. “The next thing I would I think the challenge will still come down to, can you trust whatever analysis and results that the tool is giving you?” they said. “In that situation, if you have got a good understanding of the prediction, you can assess and state that it is interesting.” The problems and further risks arise when these predictions are taken at face value and stated as an inherent truth. One of the founding principles of data is accuracy, so it is essential that this must remain, even in a time where machine learning has accelerated to a point of making predictions. It was agreed by the group that they do need see a time in the near future where AI-based predictions will not be further assessed by humans.
An interesting notion raised by one member from a media-based business was that would the general public be happy subscribing or paying to receive content and entertainment that had been created by AI? There is still a long way to go before generative AI can write a comprehensive script for a television show, but would viewers be happy receiving content not created by humans?
There is not necessarily an issue with data tech keeping pace with the developments of other technologies, but a lack of education and guidance around new technologies such as generative AI. The attendees agreed that a slow pace must be taken with implementing generative AI tools, which would run the risk of data technology not keeping pace with developments of the outside.
Opportunities
As tools such as ChatGPT enter the public eye, the uses of such technology can be far reaching. Some DataIQ members have already begun experimenting with different possibilities of the tool, including one retailer. The retailer described to the group how they asked ChatGPT to predict the number of units of a seasonal item sold in its stores, after checking the numbers that were sold in reality for that season, the prediction by ChatGPT was within 1% of the total. The benefits of this type of prediction are clear to see and can have some considerable advantages on finances and wastage for retailers, food-based businesses and those with complex supply chains.
Other opportunities include being able to assess the efficiency and lifespan of certain infrastructures. These types of items are often taken for granted in the day-to-day operations of a business, but they require maintenance and often have a shelf-life. Generative AI tools can help maintain these infrastructures and assess when and where upgrades, replacements or further maintenance is required. The cost savings could be enormous as businesses are able to replace tools at the optimum moment reducing downtime and extracting every opportunity from each tool.
There are countless opportunities that have not even been considered yet as businesses are, rightly, taking a steady and measured approach to new technologies. A major consideration is what happens to ownership of the data input into the tool – data ownership and intellectual property are increasingly important for businesses and teams. There is also the need for regulations and guides to be created in order to protect consumers, businesses and data, but these will take time to be agreed upon and implemented, which means data technology runs the risk of not keeping pace with new technologies.
To take part in an upcoming roundtable, view the topics of discussion here.