Is artificial intelligence already here? Absolutely. Is it going to take over the world? Well, AI is certainly playing a growing role in making our own lives more convenient and more informed. Think spam filters, voice-activated personal assistants, and autonomous vehicles, to name just a few.
From a business perspective, AI has the potential to be a massive game-changer. In fact, it’s already disrupting almost every type of business process. Gartner estimates that by 2021, a staggering 70% of all organizations will be harnessing AI to improve employee productivity.
The rising stars of the AI world are machine learning (ML) and deep learning neural networks (DLNN). Both are based around the idea that systems can learn and improve autonomously as they are constantly fed new data. It’s not a new concept, but it has leaped forward in the last 10 years thanks to the arrival of computing hardware that can apply complex calculations to massive data sets.
Big data, big risks
This ‘big data’ comes from connected devices, social media, business apps, machine logs and more. Artificial intelligence and machine learning work best when large and highly varied amounts of this data are available. But it can also be an overwhelming task for enterprises to manage these often unorganized and siloed data sets effectively. In short, traditional storage methods just aren’t enough to keep up with the volume and variety of data available.
And as more countries introduce strict privacy laws, the costs of compliance can also put big data – and accordingly, AI-driven innovation – beyond the reach of all but big businesses. In fact, IDC Technology Spotlight found that the two biggest challenges of deploying successful AI workloads are data volume and quality (50%) and advanced data management (47%).
Another risk enterprises must consider is the use of AI and ML software that may still be in a nascent phase of development, or lacks the skilled professionals to leverage it properly. As with any new data technology, it exposes enterprises to the risk of misinterpretation of data and wrong decision-making. More than ever, enterprises need a way to both manage their big data, and comprehensively pilot and evaluate AI tools before deploying them to the field.
Enabling data-driven AI innovation
Sticking with internal siloed or synthetic data is one way to mitigate these risks, but this approach severely limits competitive insight. Consequently, the emerging world of data collaboration – by which companies in disparate industries (financial, legal, government and more) pool their data – has quickly become the hottest trend in big data. It means enterprises can now work with both governments and startups to share data insights and expertise that help drive artificial intelligence and machine learning innovation.
In the transition to a data-sharing model, some obstacles organizations will need to address include:
- Does the data contain personal, sensitive or proprietary information?
- Do we have current consent to use it for secondary purposes?
- What is the data governance process and chain of command?
- What are the InfoSec and technology considerations?
A secure data collaboration and governance platform, such as Data Republic, can alleviate these concerns. It enables organizations to safely share, match and license data across departments and with partner organizations. Using this model, the actual data analysis takes place in a quarantined environment where AI/ML solutions can be quickly deployed and piloted using real data. It empowers data leaders to execute their innovation agendas more quickly – and with less risk – than would be the case if they had to develop such capabilities on their own.
AI solutions in action: Hyper Anna
Recognizing a niche for AI-powered business assistants, an Australian startup developed ‘Hyper Anna’, a virtual data analyst that enables users to ask plain English questions about the key drivers of their business. A major Australian bank was interested in evaluating the AI capabilities of Hyper Anna – specifically, to allow its merchant customers to explore their sales data and obtain actionable insights.
The outcome of the trial was highly positive: a return of $308K per annum (6X ROI) for a group of 10 merchants, with opportunity to scale across 1000+ merchants. Prior to Hyper Anna, the bankers would spend about one week (40 hours) every quarter to prepare sales and customer insights for their merchant customers. With Anna, this went down to 30 minutes as the need to manipulate multiple systems and Excel files was removed. This is just one example of how AI technology can produce transformative outcomes when it is fed rich data.
Big data plays a critical role in the success of artificial intelligence and machine/deep learning deployments. As the volume and variety of data continues to grow, data collaboration between multiple parties will become even more important. Security, interoperability, and a standardized process for all stakeholders will be key to the success of these initiatives.