Dael Williamson, CTO EMEA at Databricks
As the rise of AI continues, businesses are rushing to invest in the technology – but are they prioritising effectively?
Whilst competition continues, enterprises are keen to deploy AI as fast as possible, but it can actually be more beneficial for them to focus on ‘big impact’ AI – technology that will help them the most in the long run.
With this in mind, we speak with Databricks’ Dael Williamson, Chief Technology Officer EMEA at Databricks, about the importance of strategic AI investments and how companies can best stay ahead in the digital transformation race.
AI is booming. How are companies looking to keep up with the competition?
The past few years have seen companies aggressively seeking investments in AI tools. This is set to continue with every single company surveyed in a recent MIT report saying that they will be increasing AI spending over the next year, with half reporting that the increase will be over 25%.
However, company resources are finite, which means teams have to think very carefully about how they can keep pace with the rate of innovation, whilst remaining conscious of their bottom line. Simply reducing spending is not really an option, as that could lead to falling behind the competition, so the answer instead is to be ruthless with AI prioritisation.
Critical to this is to avoid a scattergun approach to AI investment and not go after every opportunity and product out there. Instead, companies must develop an implementation strategy that will offer the biggest impact to their business – either by increasing efficiencies and reducing costs, or by driving revenue.
What role does data management play in developing an effective AI strategy?
Data management is key to everything and ensuring that any AI effort is built on top of solid data foundations is step one of any AI implementation project. If these are not in place, then companies will struggle to cope with the vast amount of data processing demands or the quality of data governance that is needed to facilitate rapid AI innovation.
Investing in a fit-for-purpose data architecture, such as a data intelligence platform, is a good start. Data intelligence platforms revolutionise data management by using AI models to automatically analyse all data across an enterprise and how it is used. This ultimately empowers every employee, even those in non-technical roles, to use data and AI.
40% of CDOs, CTOs and CIOs report that the biggest problem they face with their data is training and upskilling staff, so the importance of this approach can’t be overstated.
With strong data intelligence in place, companies can not only identify the correct AI products to invest in, but they can also ensure that they are implemented in the most effective way possible.
What investment approach should companies take when it comes to AI?
The first thing to think about is whether to go after quick wins, or take the time to pursue more complicated tasks that might reap more reward down the road. In this regard a balance has to be struck depending on the specific circumstances of each company. However, another key factor is whether to go after ‘offensive’ or ‘defensive’ use cases.
Much of the talk around AI focuses on ‘defensive’ use cases, which look to improve efficiencies and reduce costs – for example the automation of repetitive tasks that typically can drain resources or be prone to human error.
However, those who do not also think about ‘offensive’ use cases, those that drive revenue, will miss a trick. In fact, from the same survey, 70% of those asked said that it was just as important for AI products to drive revenue as it was for them to save costs. The number of use cases for driving revenue are steadily increasing. These include identifying new markets and untapped opportunities by using AI to analyse vast amounts of data; detecting subtle changes in customer behaviour through AI programmes to bolster retention; and analysing patterns and trends in forecasting data in order to optimise stock management or fine-tune production schedules.
How can companies know which AI products to pursue?
This will depend on which industry or even which geographies the company is in. For example, companies must assess whether the potential costs of noncompliance or a data breach are more of a concern than the potential benefit of onboarding a new customer to drive revenue – and this will be very different for a heavily regulated industry compared to a less regulated one.
To help determine this, scorecards can be developed by teams to assist in working out what should be focused on. These scorecards can be broken down into three main categories: strategic importance, feasibility and tangible ROI.
These will give a score that takes into account a balance between how important a use case is to meeting business goals, how achievable it is given current infrastructure and how easy it is to measure success. By including all areas of the business, this scorecard can give a strong indication as to where money and time should be spent.
Content Courtesy – AI Magazine
