The days of enterprise AI are over

According to McKinsey’s global survey on the state of AI in 2021, more than a quarter (27%) of respondents reported that at least 5% of their earnings before tax and benefits were because of artificial intelligence (AI), and more than a third (33%) saw their costs decrease by 20% or more in 2020 through the use of AI.

Companies investing in AI are reaping its benefits, and many in the early stages of AI adoption have seen quick results. As organisations mature, AI-driven business value compounds. Eager to see further results, organisations may try to expand their AI capabilities by hiring data scientists and experts.

However, a trend has emerged. Companies that rely on data and AI experts alone for AI transformation notice diminishing returns anywhere between the 10–30 projects mark. In addition, the current skills shortage plaguing the Australian labour market has made it more challenging than ever before to hire data and AI experts.

This is where ‘everyday AI’ comes in. Everyday AI refers to AI that is so intrinsic and entwined with the day-to-day workings of a business, not just being used or developed by a central team. Here are three reasons businesses should adopt an everyday AI approach.

1. Everyday AI enables organisations to undertake more projects

According to an analysis of the local talent pool by Intelligen, a leading data and AI consultancy, there are approximately 1700 data scientists across Australia, up just 7% from a year ago. Additionally, a third of these data scientists have changed positions in the last 12 months.

This is not enough talent to keep up with demand from organisations that are looking to hire these specialists to drive AI results. Legacy organisations have an even more difficult time, as they tend to be unable to offer the perks and prestige of big-name tech companies.

There is a solution. Companies are better positioned to unlock AI’s potential when they empower all employees, not just those in the data science teams, with the skills and infrastructure to use AI daily. Establishing AI upskilling programs that cater to employees’ levels of technical expertise, coupled with an investment in AI infrastructure and commercially available tools, can democratise AI across the board. This investment in training and resources helps employees identify AI use cases in their business units and enables them to run and, for some who have an affinity, develop AI models of their own. With this approach, the entire organisation can use AI and scale out to thousands of projects, undertaking far more than organisations that centralise AI in the hands of data science teams.

2. Everyday AI generates better-quality products over time

In a 2021 Harvard Business Review article, David De Cremer, professor in management and organisations at the National University of Singapore Business School, and Garry Kasparov, chair of the Human Rights Foundation and world chess champion, made a case for AI augmenting human intelligence. The authors argue that it’s a fallacy to believe that it’s a zero-sum game between AI and humans. AI is good at ingesting and understanding data at scale, but humans can imagine, anticipate, feel and judge changing situations, allowing them to more easily balance short-term and long-term concerns. Together, humans and AI systems have worked hand in hand in a variety of cases, from winning chess tournaments to improving semiconductor manufacturing.

But what does this mean for businesses in practice? Encouraging people across the organisation to work in collaboration with AI allows for a quicker and better understanding of stakeholders’ needs and a more efficient delivery of new products and services.

For example, Dataiku customer Etihad Airways built a tool that provides recommendations to airport managers about how many counters to open for customers to check in, based on data around passenger arrival times. Managers onsite can accept the recommendations or make adjustments as deemed fit. The relationship between the manager and AI is a virtuous loop. The AI helps managers more easily prepare ground staff for future check-in loads and the managers improve the AI through their feedback.

This example illustrates how everyday AI leads to stronger products and processes by allowing teams to use data insights along with human intelligence to drive AI-led changes.

3. Everyday AI improves business continuity

AI projects don’t end when they move to production. Instead, they are living tasks that require constant monitoring and evaluation.

For example, an organisation that creates a model to understand fraud must continuously update it. Fraudsters change their methods and patterns when identified schemes no longer work. Unless models change to adapt to new methods, the culprits will be able to continue causing damage. It is important to make sure AI projects are being updated to ensure they are working well and can adapt to any changes in data or business needs.

When data and AI are concentrated only in the hands of experts, other members of the organisation may be left unable to understand the projects. This is where organisations run into continuity issues. Non-experts being unable to run, edit or understand AI projects or experts leaving the organisation without documentation creates substantial risk. Empowering employees across the board with the knowledge and the tools to understand these models eliminates this problem.

But there’s a way to go for most organisations

A recent study from IDC (commissioned by Dataiku) found that just 39% of organisations in APAC have already invested in AI. This insight highlights that many businesses have a long way to go before they can reap the benefits of the technology. Regardless of stage, businesses should look beyond their data teams and empower employees across the organisation to leverage AI. This will allow them to take on more projects, develop more efficient products and establish business continuity.

Image credit: iStock.com/monsitj