An effective artificial intelligence strategy uses the right tools to solve the right problems, an analyst says.
Dean Lacheca, Senior Director Analyst for Gartner, told delegates at the Gartner IT Symposium/Xpo in the Gold Coast on Monday that they need to stop seeing AI as a futuristic piece of technology.
“AI, in reality, is more than just a tool. It’s a whole range of tools with different variations in complexities, costs of ownerships, consequences and opportunities,” he said.
A good strategy allows organisations to manage risks, address and mitigate concerns and accelerate the role of AI within an organisation, Mr Lacheca says.
“An effective AI strategy is (making sure) that you’re using the right tool to solve the right problem at the right time,” he said.
“It will help you manage the expectation of your organisation and it will help you put some structure around how you improve and evolve the use of AI in your organisation.”
Five questions to answer
Mr Lacheca outlined five questions that an effective AI strategy should address including use cases, skills, data, technology and organisation.
“AI is a game-changing technology, but you’ll need a roadmap to see how you grow its potential within your organisation.”
There’s a broad range of use cases that come into play from a government perspective, Mr Lacheca said, and they fall into three categories: insights and decisions, user experience and process improvement.
“As you’re thinking about how you’re going to put some strategies around the use and implementation of AI, think broadly to these three categories and how your organisation will prioritise each of these areas,” he said.
A range of skills are required to build an effective strategy, and most of these skills already exist in organisations.
“As part of your AI strategy, you need to understand how you’re going to foster and develop not just the specialist skills that you need for AI adoption but how you’re going to develop the capabilities across all domains of your organisation so they can readily adopt the change required for AI,” he said.
This begins by looking internally first and not neglecting existing employees within an organisation.
“They may not be professional data scientists, they may not be experts in machine learning but they’re a resource that you’ll have available to ensure that you’ll have a strategic approach,” he said.
Data for AI comes in all shapes and sizes, and links back to use cases. It impacts the approach taken with use cases and affects how data is used within an organisation.
When it comes to AI, there’s an incorrect assumption that a lot of data is required, Mr Lacheca said, but the first challenge is to identify those data sources.
“Many of the approaches that you’re going to adopt in your early stages of AI are not requiring huge volumes of data but they require a high quality and availability of data,” he said.
No right or wrong answer
In regards to technology, Mr Lacheca said there are a range of technologies for AI, however, not all of these need to be used.
“It’s about evolving and focusing on the technologies that you need to meet your requirements for today,” he said.
Organisations should not be concerned about using the wrong technology because there is no such thing, according to Mr Lacheca.
“There is no wrong technology in AI, there’s just a mismatch between where you are from a maturity level, and the use cases that you can afford to address at this point,” he said.
From an organisational point of view, an effective AI strategy must examine where expertise should reside within an organisation, which can pose a particular challenge for government, Mr Lacheca said.
“Like other technologies of the past…if AI doesn’t come under some sort of structured approach of how we’re going to manage it as an overall government, then we’re going to see duplication, costs competition for similar resourcing, and inconsistencies in the way that we evolve our technology going forward,” he said.
“An effective AI strategy is (making sure) that you’re using the right tool to solve the right problem at the right time.”
More than a ‘set and forget’ strategy
An effective AI strategy is not a ‘set and forget’ strategy; it is one that needs to be revised over time, Mr Lacheca says.
“You need to build an AI strategy that is going to evolve with you as the maturity increases over time,” he said.
“AI is a game-changing technology, but you’ll need a roadmap to see how you grow its potential within your organisation,” he said.
Comment below to have your say on this story.
If you have a news story or tip-off, get in touch at firstname.lastname@example.org.
Sign up to the Government News newsletter