Comment: Debunking the three biggest AI myths

As it stands, there are many companies that struggle converting business data into business impact. Beyond this, there are several commonly held myths and misconceptions around AI that only add fuel to the fire. The reality is that, to build an AI strategy based on valuable business use cases, future-proofing technologies, and inclusive, sustainable processes, these myths need to be debunked, says Gregory Herbert, senior vice president and general manager, Dataiku.

(Image: AdobeStock)

AI Myth #1: AI is only for enterprise companies with hundreds of data scientists

Organisations often don’t know where to start when it comes to incorporating data and AI into their business model. Or sometimes, they do know where to start, but they are faced with overwhelming challenges that impede its adoption.

One of the top challenges is finding and retaining the right data talent, and this often comes down to the fierce competition in the market for data scientists. However, there are alternatives to perfect, ready-made data science talent.

Instead of looking for a data scientist unicorn, companies should focus on isolating the precise skills needed and finding the data talent that has some of these skills. Over time, organisations can build out a data team while fostering collaboration and knowledge sharing. This has the added benefit of encouraging an open data culture and improving employee retention.

It can also help to ensure a technical infrastructure that can support the necessary quantity of data or processing an organisation requires: data architecture design needs to be sustainable enough to support a variety of users’ needs but agile enough to scale as the organisation grows.

To scale successfully, organisations need to infuse AI and analytics everywhere and amongst everyone, so that processes and technology are so deeply ingrained that every part of the organisation thrives.

AI Myth #2: AI technology alone will fast-track success

AI technology can undoubtedly be transformative, but it is one piece of a bigger picture. Without the right people, processes, and data, technology alone won’t be able to achieve its full potential.

Succeeding at AI initiatives requires fostering a culture of data creativity at every level. However, organisations also need to find a way to provide employees across the board with the autonomy to make more informed decisions with data to achieve collective company progress and purpose.

Not only does this encourage more people to collaborate around data, but it also ultimately serves the company’s collective purpose — solving their business goals and becoming more agile and productive.

Finding the right use cases for your business is also critical for success with AI. Every industry has its own unique use cases where applying AI has the capacity to introduce new efficiencies, cost savings, and even revenue increases. But it’s the use cases that are very quick to return value to the business’s operations that will become the most important and given the most support by leadership.

AI Myth #3: AI will automatically deliver ROI 

AI is not a magic bullet that is going to drive organisational change immediately. But, a thoughtful approach that combines people, processes, technology, and data can solve high-value business challenges.

There are many stages that go into recognising value from AI. It may start with looking at data sources and understanding what teams need from data to produce value. It may involve stitching together those data sources and identifying the skills available to build out new data and analytics capabilities. Regardless of whether an organisation is working on optimising processes, or more advanced machine learning, its operating model should always be built to maximise its ROI.

Use cases vary widely by industry and are best selected in close conjunction with business teams. A research study by ESI Thoughtlabs shows that delivering ROI on AI can be difficult at the beginning, and slow going even for the most experienced organisations.

It is not until companies’ scale AI more widely across their enterprises and become more established leaders that the ROI rises. This can often come down to high upfront costs in data preparation, technology adoption, and people development and typically requires a time frame of more than two years to generate significant returns.

When it comes to AI, tangible value often comes from the very first use cases — the straightforward, often mundane ones that are used to gain executive buy-in and as proof points for something bigger that an organisation wants to achieve.

At the end of the day, many organisations are feeling the pressure to transform their businesses through AI, but very few firms are far advanced in their AI journeys and almost all can see room for improvement.

There are many decisions and challenges to make as companies navigate their AI roadmaps, and many companies may still feel they need to choose between empowering their business analysts and empowering their data scientists.

However, if companies can look to systemise the use of AI and data by including more people in analytics processes, they should be off to a solid start to experiencing the benefits of AI.

Gregory Herbert, senior vice president and general manager, Dataiku