Last week, we sponsored Computing’s annual IT Leaders’ Summit which took place at Carlton House, London. The summit provided an opportunity for senior IT executives across all industries to discuss how they can drive digital transformation in areas such as the cloud and Artificial Intelligence.
Richard Simmons, our Head of European AI & IoT Practice presented to a packed house and hosted a round table; here’s a quick recap of our day.
The hype of AI
Artificial Intelligence has become the latest focal point in the conversation around data insights. Yet, contrary to the volume of noise surrounding it, according to industry studies only 4% of organisations have actually deployed AI*, with most of these businesses still in the early phases of AI adoption and facing unexpected challenges. Richard took to the stage to bust some myths and to provide the facts about AI that have been overlooked in the hype.
“AI is not the answer to everything. It needs more data, it requires more resources, you need the right foundations in place and your infrastructure has to be ready, otherwise there is a huge amount of inefficiency” he said. “You cannot underestimate the time it will take to develop and build. It can take weeks, even months, to get an AI strategy up and running and a massive 80% of an AI project’s time is spent in the data preparation phase.”
Richard also highlighted the extremely long training times required. This is partly due to the different sets of skills needed to fine-tune and deploy it – the skill set of the person managing an AI project will be vastly different to the skill set of the person building it.
“After you have worked on a business strategy, then spent a long time preparing the data, you have to experience all that pain again. Because the more data you give an AI project, the more accurately it performs. If you want to really drive value from the data you have, the project is never ending. AI is not a quick fix.”
Opening the discussion
Following Richard’s presentation, our over-subscribed (it’s almost like it’s a hot topic!) IT leaders round table began, where we discussed the key factors and approaches to be considered to deliver a scalable AI strategy.
The discussion started with the question ‘who on this table has deployed an AI project in their company?’. Out of 29 people, only five IT leaders said yes. Those who responded with yes reinforced that their projects were in the early phases and were starting small before they scaled up.
“It can take a lot of time with very little return at the start of an AI strategy. So, it can be hard to encourage the rest of the business to support the project when they can’t see the rewards. This is why starting modest and breaking it down into smaller projects can help. You don’t want to bite off more than you can chew” agreed Richard.
Sharing is caring
We are currently seeing a great push for data sharing across businesses, a concern that was raised during the discussion. As Richard said, AI operates most efficiently when it has been fed a lot of data therefore it would make sense for businesses to share already processed and interpreted data with others in a similar sector. So why aren’t businesses doing this?
“There isn’t always a desire to share what you’ve worked so hard on. If you do share your data, there is a huge risk that the person you’ve shared it with will implement it and use the data better. If you share your work, get ready for the competition to begin.”
Who owns the data?
One topic that dominated the round table was the ownership of data within a business. Many of those who sat at the table expressed the desire to be a data-run business, but getting to that stage wasn’t as easy as they hoped. According to Richard, if a company wants to be data led and to use AI efficiently, there needs to be a “new culture built internally. Every part of the business needs to work with data in mind, not just the IT department and those involved with the AI process.”
This is where data ownership is a necessity. It was mentioned that whilst employees may be really interested in deploying AI, no one was excited about the management, governance and upkeep of the data needed for the AI to work efficiently! To combat this, one person said that their company recently wrote a data strategy – from compliance, to governance, to how the company values and uses data – in order for every employee to be on the same page.
As a final note, Richard said “getting every part of the business on board is vital, but it will take work. This culture change to a data driven enterprise will not happen overnight, it should be on going just like the AI project itself.”
Is your business ready for AI? Is your infrastructure? Find out how Logicalis UK and IBM can help you overcome AI infrastructure limitations and access an IDC expert infrastructure readiness report – Finance or Retail and Manufacturing.
Logicalis UK would like to thank Computing for hosting us at the IT Leaders’ Summit and to those who joined us at our panel session and the round table.