In an age when new AI offerings and partnerships are announced every day, no business wants to be left behind. There’s genuine excitement about AI’s potential to unlock new opportunities, efficiencies, and growth – and rightly so. But success isn’t guaranteed.
New research by MIT shows that 95% of AI projects are failing. And a separate study shows that AI is being used most by those who understand it least. For businesses anxious to get ahead, rushing into AI investment without laying solid foundations could spell trouble. Instead, putting AI realism at the heart of strategies could save companies from costly mistakes and help ensure any changes drive genuine benefits for colleagues and customers alike.
AI realism shouldn’t be conflated with technophobia or luddite tendencies. Rather, it means prioritising accuracy over speed and focusing on impact over adopting AI for AI’s sake. Realism – and even a healthy dose of scepticism – can be a strategic advantage. It’s what separates companies that are able to guide AI strategies to success from those that rush headlong into projects without goals or guardrails.
How blind spots steer bullish AI users off-course
We know that, for many businesses, AI strategies are yet to deliver measurable results that impact the bottom line. But where exactly are they going wrong?
AI strategies that prioritise speed and experimentation above all else typically have good intentions. This approach certainly has the potential to unlock new efficiencies and opportunities for a business. But these less well-informed AI approaches tend to be so focused on the technology’s potential that they can overlook its limitations or risks or fail to ensure systems are designed with the relevant end users in mind.
Duolingo, Klarna and Taco Bell are just some of the companies who very publicly and very quickly went all-in on AI; announcing bold plans to run with ‘AI-first’ approaches and replace swathes of workers with the technology. But their sweeping, hastily rolled-out approaches fell hard and fast. All three companies were forced to U-turn on much touted elements of their AI strategies, sometimes within weeks.
And it’s not just pride that’s lost when businesses are forced to backpedal. Firms that overhaul entire operations only to bail out of AI projects risk wasting significant resources and eroding the trust of employees and customers – setbacks far greater than if they had taken a more measured approach.
In contrast, AI realists know that you have to define what needs to change and how value will be measured before ramping up to full speed. They also know that, unless you have the right data, use cases, tools, oversight and systems to measure value in place, it’s hard for AI to deliver true impact for the teams who’ll be using it.
How realists set AI strategies up for success
Effective AI projects invest time in identifying the right use cases – and the tools that will support them – before jumping into implementation. MIT researchers identified tools that aren’t fit for purpose, or that don’t easily integrate and adapt with existing workflows, as a common feature of failing AI projects. The researchers also noted that tools supplied by specialised vendors (those with ‘been there, done that’ credibility) had a higher success rate when implemented on the ground.
For any AI project to be successful, establishing high-quality data sources is also essential. AI outputs are only ever as good as the information that models ingest. But that doesn’t mean it’s necessary to overhaul entire data estates. Instead, effective strategies focus on refining, preparing and cleaning the datasets that AI tools will need for specific jobs.
Training is also a critical cornerstone of successful AI strategies. With AI literacy rates typically varying significantly across workforces, it’s important that the staff who will actually be using AI tools learn to do so safely, effectively and in a way that makes their jobs easier. AI realists are typically well placed to identify potential gaps in knowledge, spot the risks of leaving them unfilled and develop the processes needed to get colleagues ready for the AI transition. Because there’s no value in AI rollouts that leave the team behind.
Crucially, this doesn’t mean sidelining those with a passion for AI who want to jump into strategies with both feet. That zeal, when effectively harnessed, can be great fuel. Instead of shunting keen but potentially overly optimistic colleagues to one side, leaders should focus on finding the right role for them in the wider strategy and providing the infrastructure that can channel that passion in a targeted way. In well-structured strategies, there should be scope to embrace AI aficionados and digital doubters alike.
The path forward
AI is evolving rapidly but this doesn’t mean businesses should succumb to pressure to run head-first into projects. Instead, companies should lay the foundations for long-term success; putting strategy, tooling and training at the heart of their plans and focusing on innovation that makes a genuinely positive difference to colleagues and customers.
When AI realism is given a place at the top table, companies can feel confident that they won’t get sidetracked by shiny gimmicks or pulled down expensive rabbit holes. Instead, a healthy dose of pragmatism can set up businesses for success – ready to unlock the true power of AI in the enterprise.
Steve Salvin is CEO and founder of Aiimi, an AI software company.
In an age when new AI offerings and partnerships are announced every day, no business wants to be left behind. There’s genuine excitement about AI’s potential to unlock new opportunities, efficiencies, and growth – and rightly so. But success isn’t guaranteed.
New research by MIT shows that 95% of AI projects are failing. And a separate study shows that AI is being used most by those who understand it least. For businesses anxious to get ahead, rushing into AI investment without laying solid foundations could spell trouble. Instead, putting AI realism at the heart of strategies could save companies from costly mistakes and help ensure any changes drive genuine benefits for colleagues and customers alike.
AI realism shouldn’t be conflated with technophobia or luddite tendencies. Rather, it means prioritising accuracy over speed and focusing on impact over adopting AI for AI’s sake. Realism – and even a healthy dose of scepticism – can be a strategic advantage. It’s what separates companies that are able to guide AI strategies to success from those that rush headlong into projects without goals or guardrails.
