Explainable AI and the Role of BI in Analytics

Explainable AI and the role of BI in Analytics

Business Intelligence (BI) is at the heart of continuous improvement in business. It enables better decision making by collecting, storing, and analysing business data – from customer experience to quality controls – and converting this information into meaningful reports. With BI, progress is data-driven and easy to see.

However, like most things tech today, artificial intelligence (AI), and machine learning (ML) are changing existing practices; this will alter how we interpret and action data insight. While some think AI could overtake BI capability, I believe that - at least for now - the two are far more likely to work alongside each other. Here, I explore why.

How is AI different to BI?

AI is a computer system that thinks, acts, and learns with a human-like intelligence. It is part of the fabric of our everyday life – capable of visual perception, speech recognition, and decision-making. AI’s ability to learn from data and experience, and its ability to make judgements and predictions based on this understanding, means that AI can bring new layers of understanding to business processes. Where BI captures and analyses data trends, AI can use this insight to go one step further – identifying patterns, making predictions, and giving recommendations.

However, AI currently lacks the depth and well-rounded knowledge of BI. While it can flag anomalies and trends, plus make suggestions, it cannot tell us specifically why these exist. AI can tell us what to change, but not why. At a business level, it can be difficult to justify change when the reasons behind the recommendation are not clear.

Likewise, BI in isolation can be too rigid. BI often uses statistical methodologies to identify trends, lacking the flexibility to interpret variables that AI’s intuition allows. Meanwhile, AI can subtly adjust parameters to get the best business results – using algorithms to trigger changes that may go undetected by BI tools.

AI-enabled BI will solve problems holistically – considering past, present and future reality, linking causation to results, triggering alerts to new opportunities and determining the optimum result with full business context.

What’s next?

AI-enabled BI is one step closer to ‘Explainable AI’. To justify business decisions and recommendations, AI must be easy to interpret and clear-cut. For now, AI-enabled BI may bridge this gap, however I see the future with AI platforms that auto-generate an explanation of models in terms of accuracy, model statistics, and rationality.

Natural language processing (NPL) and VR/AR can also help to bring decisions to life and to make them more accessible and interpretable for the vast spectrum of stakeholders that BI has typically had to cater to.

Meanwhile, augmented analytics embedded in enterprise applications can identify patterns without human bias and trigger optimal business process changes that may go unnoticed by BI. According to Gartner, by 2020 40% of data science tasks will be automated. This will lend itself to continuous intelligence, where real-time analytics are integrated with a business operation to automatically make decisions based on a continuous stream of data, business rule management, and machine learning.

The result will be an AI platform that incorporates the best of BI - its clarity for stakeholders and contextual understanding of business - blended with the best of AI – machine learning, automated decision-making, and predictive analytics.

What does that mean for the job market?

According to IBM, demand for data scientists and advanced analysts will soar by 28% by 2020. Already, there is more demand than the market can keep up with. Machine learning, big data, and data science skills are among the most challenging to hire; but these are also the skills that can create the biggest market differentiation for companies.

To meet the talent shortage, Gartner predicts that this demand will see a rise in citizen data scientists – where individuals from a non-statistical or analytical background/role use AI-powered analytics to automate data science functionality, reducing organisational demand for skill short data skills. According to Gartner, through 2020 the number of data scientists will grow five times faster than professional data scientists. 

This means that data insight will be more easily and broadly leveraged across the business, with citizen data scientists using both BI and AI to make enterprise-wide, data-driven decisions possible.

So, if you’re interested in pursuing a career in data, now is the time to dive in. In addition to AI and machine learning, I recommend looking into MapReduce, Apache Pig, Apache Hive, Apache Hadoop, Big Data, Data Science, NoSQL, predictive analytics, and MongoDB – some of the most lucrative and in-demand analytics skills in the current market.


AI will transform the already vital role of BI in business. Companies that embrace both will have the potential to capture, interpret, understand, actionand predict both existing and future data trends. Meanwhile, companies that only adopt one may miss the bigger picture. To that end, companies that rely on BI to drive progress, will benefit from reviewing their existing BI practices to identify room for AI-based improvement. Meanwhile, those that work in BI would do well to read up on AI. And that goes for whatever business function you work in – we’re all citizen data scientists in the waiting.

I recruit exclusively to the BI and Data market in the North West. If you fancy a chat about your future – whether you’re recruiting in the industry, developing your career, or just want to find out more, please reach out to me via james.teasdale@onezeero.co.uk.

How useful did you find this article?
Thank you for your feedback!
5.0 / 5.0