Experts speaking at the inaugural AMR Technology Forum offered a picture of the road to deploying AI in enterprise which is not unlike the early days of Formula One and other motor racing series.
The core message from Aston Martin Aramco technology partners is the best way to employ AI is to use it.
It is an ethos which is arguably the very basis on which Formula One racing is built. Want to try a different valve configuration on your engine? Do it. Is the car better with the power unit behind the driver than in front? Let’s find out. Are six wheels better than four and so on.
AI strategy
Speakers from the AMR Network of technology partners essentially said the strategy for deploying AI is no different to those shifts and trials in car design. Devise a clear plan about what you want to achieve and just do it. Do not be fearful, go all in and make sure what you do helps the people with the real expertise to do their jobs better.
David Ingham, digital partner in media and entertainment with Cognizant (pictured, third from right), said any AI deployment needs “to be solving a problem”, so it makes sense the first step would be to identify what that is, then the rest “falls in line”.
Ingham explained Cognizant’s approach is born of its position as a company which brings “all the layers of the AI stack” together. He referred to chips, infrastructure and software as the elements in mind, explaining the primary goal is to combine them into “a coherent workflow”, a “process that adds value”.
He argued Aston Martin Aramco’s racing cars are a great example, because individual components “don’t add value, they don’t win a race” until they are brought together.
An obvious use for AI in any organisation is coping with vast amounts of data, an element Ryan Lewis, head of Cohere’s operation in the UK and Ireland (pictured, second from right), said is another in which Formula One is a fitting example due to the volume of information teams generate and rely on.
He pointed to telemetry, diagnostic and simulation data as examples from the sport which emphasise the challenges any organisation faces in making the most of their information.
Data is “sometimes really hard to access” due to being stored in different systems, modalities or even languages.
Lewis believes “making sense of the data wherever it’s sitting” to enable rapid decision making “in high-stakes environments” presents a “really exciting opportunity for AI just to take that off the table”.
He envisions a set-up where individual AI agents communicate directly to synthesise the insights, “reports and responses” staff can use to work more effectively.
For Lewis, the watchword of deploying AI in a business is empowerment.
Cohere is not inherently seeking to use AI to “replace the things that are high value in organisations”, instead it is driving to “give the people the power to do the things that are mundane and routine” more easily, in turn accelerating the decision-making process.
Teamwork
Aston Martin Aramco technology ambassador Eric Ernst (pictured, second from left) elaborated on the empowerment theme, explaining it is the experienced staff within the team which drive the value of AI usage.
He noted machine learning is well-established, meaning AI deployment strategies today are more about how best to employ generative iterations.
Aston Martin Aramco is deriving benefits by using AI to create and assess several possible strategies, which knowledgeable engineers can then pick from based on their real-world experience.
Ernst’s advice is humans must remain the foremost element in use of AI, interpreting recommendations and conclusions the technology produces to derive the smartest outcomes.
For a Formula One team like Aston Martin Aramco, this could be as simple as when drivers should go flat out and when they should be more conservative during a race.
The point about the human element was picked up in a later session by Jonathan Armstrong, senior director of Brand and Creative at Arm, who noted AI is good at deconstructing “complex problems” but argued the full benefits of the technology can only be reaped when it is combined with the vision, ingenuity and creativity of people.
“When you have technology and AI and human insight altogether, that’s a powerful weapon.”
He said Formula One teams are working to optimise the combination of AI and human abilities, an “exciting” period for the engineering sector.
Simon Cox, chief transformation officer at ServiceNow (pictured, far right), urged organisations not to be paralysed by worries over the accuracy of their operational data. He explained the companies making the most of their AI strategies “are the ones that embrace” the chaos of current information sources and adopt a “fix it as we go” mindset.
He argued the iterations of AI involved in the latest dawn of the technology deliver the capabilities to correct data anomalies on the fly and so make a success of the very approach which embodies the ethos of Formula One racing.
The technology today can “find those gaps and see through the noise” in a company’s data. The volume of information and challenges of employing it should not hinder companies, instead Cox believes the secret to success is to “just start, get going”.
Cox’s advice to take the plunge was echoed by Ben Richardson, head of CoreWeave International (pictured, third from left), who urged companies not to wait for the perfect AI or deployment strategy.
“Like anything, if you really want to use it you’ve got to take risks, you can’t sit there and you can’t just wait”, because rivals are not.
“People have to be brave and you have to be able to say this is happening, it is here, how can I start to capture it”?
Richardson’s view is cracking on and using AI holds the potential to create a virtuous circle, where the technology becomes more valuable based on how much it is employed. Or, as he more succinctly put it: “the more people that use it the more you get out of it”.
Restrictions
With the experts’ broadly agreeing the best strategy is to learn about AI by using it, what do they think is holding most organisations back?
Lewis believes many large organisations are simply not clear enough about what they want to achieve and so are struggling to develop a “cohesive enterprise AI strategy”.
Companies must consider the result of their AI deployments and doing so requires “a top-down perspective”, one which ties every element together.
Rushing headlong into deploying any technology is, of course, inherently risky. Richardson acknowledged as much and advised companies to build protections into their strategy “if that’s what you need” to deliver the confidence required to reap the AI benefits he outlined.
Cox noted governance is a decidedly unsexy element in any discussion of AI, but explained it is still necessary.
Experience from using data, cloud computing and more has taught ServiceNow the importance of installing controls “from the beginning”.
“Resilience by design can give you the control tower for your AI and then you’re able to fly.”
Governance “sounds dull” in any discussion of enterprise AI strategy best-practice, but is an element Cox argues is absolutely necessary.
The post Feature: Experts offer top tips for enterprise AI deployments appeared first on Mobile World Live.
Login
