How to do AI analytics correctly

Focus needs to turn to doing AI analytics right as adoption soars

Digital transformation and adoption of artificial intelligence (AI) have soared around the world since the onset of the Covid-19 pandemic, making it increasingly important for organisations to focus on the tools, models and ecosystems in which AI will be deployed.

This emerged during an executive webinar on AI Powered Analytics for Digital Acceleration hosted by iOCO and IBM, in collaboration with ITWeb this week.

Anthony Marshall, senior research director at the IBM Institute for Business Value (IBV), said the IBV’s research showed that the pandemic had catalysed a dramatic increase in digital transformation, and the adoption of technologies such as AI and the cloud.

“At the beginning of 2020, organisations were focused on advancing incremental transformation, building platform business models, improving business agility and resilience and closing the customer engagement gap. Then the pandemic happened. The IBV noted that immediately, organisations pivoted, protected and rescaled, and digital transformation projects sped up dramatically," said Marshall.

Data is a team sport and this requires integration and collaboration.

Varsha Ramesar, iOCO.

"What we also saw was an increase in investments across process automation, remote working, workforce resource management, workforce skills and training and cyber security. 60% of organisations accelerated their investments in digital technologies. Public cloud technologies started declining in parts of the world, but hybrid cloud investments increased. Contrary to expectations that investment in IT would decline, we saw CapEx spiking,” he said.

“In terms of AI, we’ve seen a dramatic increase in adoption over the past two years. Right now, we’re tracking that 78% of organisations globally have adopted AI in some form,” Marshall said. “Organisations are only now beginning to understand the full impact that AI can have. But there are key issues we need to be aware of, and AI bias and ethics are among the important topics to focus on now.”

Varsha Ramesar, iOCO, data and analytics cluster executive, said: “Ethics has been slow to the table in analytics discussions. There are two perspectives often quoted in the data world– if you torture the data enough it will tell the story you want it to tell; but on the other hand, if you come without data all you have is an opinion. Somewhere in the middle is what you want.”

She said overcoming bias required access to more diverse data sets and training of the analytical models.

“Integration is also critical," she added. "There is a tendency for organisations to assume data analysts will rescue everything, but data is a team sport and this requires integration and collaboration.”

Lisa-Giane Fisher of the IBV in South Africa said another priority for organisations moving to harness AI-powered analytics in hybrid cloud environments has to be security.

“The number one priority is security. Not all organisations were ready for the mass move to cloud. One of the biggest challenges they now face is securing their own resources and people, as well as their interactions and supply chain networks," she said.

"Our recent papers on cloud security and Zero Trust security found that across the world, in 2020, 90% of cyber security incidents originated in cloud-based environments. Given the fact that the cloud is foundational for business and platform business models, organisations need to be able to maintain a consistent security posture across cloud environments. Skills is another priority – organisations need people who can actually bring the transformation to business, and implement and utilise the technologies.”

Marshall noted: “IBM has been a leader in terms of defining the ‘new collar’ skills needed in this changing environment, in which organisations are competing for cloud, analytics and AI skills. You can compete with everyone for a limited talent pool, but what is largely untapped is the talent pool that may not have formal qualifications, but which could be skilled up with robust training and education.

Dave Albert, global business partner segment lead at IBM Data and AI Apps said: “The pattern is for companies to move towards a digital world, leveraging hybrid and multicloud. What is important now is for them to leverage tools that make it easy for their employees to access and use the data, and enable better self-service to give the average user the ability to use the software.”

Link to original article...