AI should be used to augment the abilities of data analysts, rather than replace them entirely, according to Francois Ajenstat, an ‘assistive intelligence’ advocate and the Chief Product Officer with data visualisation platform Tableau.
He spoke to Dynamic Business (DB) about the value businesses can derive from adopting AI-powered tools.
DB: Investment in AI is predicted to triple – what’s fueling this?
Ajenstat: The hype and excitement surrounding AI, which encompasses machine learning and deep learning, has surpassed that of big data. In fact, by 2021, Gartner projects that 40% of new enterprise applications implemented by service providers will include AI technologies.
Companies have started to yield financial gains from AI investments, which helps explain the growing investment in developing AI capabilities. Examples of successful AI investments include profitable consumer-facing applications such as image recognition in photographs, spam filtering and those that help to better target advertisements to web surfers.
Two further developments which rely on AI, esecially machine learning and robotics, are self-driving cars and virtual personal assistants who can understand and execute complex tasks – these have been heavily invested in by tech firms, both in terms of hard dollars and talent hired.
DB: How is AI helping businesses to unlock efficiencies?
Ajenstat: It’s improving productivity by, for example, automating routine work tasks, enhancing employees’ capabilities and otherwise freeing them up to focus on higher value-adding work. The American Bar Association journal reported a case study, which showed 1,000 legal documents can be reviewed in a matter of days with the aid of a virtual assistant, instead of taking three people six months to complete.
We believe the term assistive intelligence is a more appropriate phrase for the AI acronym (as well as being far more palatable for analysts who view automation as a threat). The concept of assistive intelligence, where an analyst or business user’s skills are augmented by embedded advanced analytic capabilities and machine learning algorithms, is being adopted by a growing number of organisations in the market today to enable them to make smarter and better data-driven decisions.
The notion of AI completely replacing and automating manual analytical tasks done by humans today is far from reality in most real-world use cases. In fact, full automation of analytical workflows should not even be considered the final goal — now or in the future. Due to its smart capabilities, AI has proven useful in assisting with data preparation and integration, as well as analytical processes such as the detection of patterns, correlations, outliers and anomalies in data. However, organisations should continue to leverage their collective expertise to ensure users use the right data, and eliminate unnecessary data prep work by using data that’s already been prepared, and avoid the creation of data silos that introduce inconsistency and unnecessary risk to an organisation’s data.
DB: Which industries have the most to gain from embracing AI?
Ajenstat: Companies in just about every industry can benefit. Organisations are becoming increasingly data-driven and to help them derive insights from huge amounts of information, they are turning to platforms that can assist them to improve productivity.
And this extends to every line of business. If we’ve seen any trend, it’s that industries that generate a lot of data – such as government, services, banking and financial services, telecommunications and manufacturing – tend to be more proactive about using data and AI. But thanks to ongoing price reductions in technology, even industries that have been resistant are getting on-board with data. They are starting to recognise not only is it becoming more affordable but the benefits are becoming too significant to pass up.
DB: What does it mean, in 2017, to be a business that’s not data-driven?
Ajenstat: In the past, data was available but there was very little of it and a significant amount of effort was needed to derive insights from this data. Therefore, businesses relied on opinions which were derived from experience to make decisions. As data and business intelligence (BI) technologies progressed, data was being collected but the frequency was very low and new data for newer insights was still very difficult to come by, and hence businesses continued to rely heavily on opinions.
Today, in the age of big data we have access to volumes of data, coming at high velocity from a variety of data sources. Therefore, the need to derive insights from huge chunks of data continues to increase, leading to the demand for self-service big data analytics.
Every line of business can be optimised by implementing insights derived by BI solutions. Using technologies such as predictive analysis, trend monitoring, real-time data visualisations and dashboards, big data converts every action of a customer or business function into quantifiable insights. These insights may include consumer behaviour, sales effectiveness, revenue management, supply chain management, marketing campaign efficiency etc. that will help empower businesses to make insights-driven decisions.
People in just about every industry and level of the business will need to become familiar with BI and data analytics to do their jobs well. Note that at the most basic level, data is a simple collection of rows and columns of information. This information can be related to any vertical and you can surely derive insights from this information.