The pandemic was a key driver in artificial intelligence adoption – 52% of businesses accelerated their AI projects in response to the crisis. Fast forward to 2022, and 90% of businesses are using AI or plan to. But even as businesses scale their AI rollouts, only around a quarter of commercially-built AI models are actually being deployed; and even fewer are generating value.
Part of the problem is the isolation of data practitioners from the business users they support. The second issue is a tendency to give technical teams free rein, setting them off on ‘data science projects’ simply to see what they yield.
To be successful with AI, businesses need to turn the working method for data projects on its head. Instead of starting with data and seeing what can be achieved, the trick to successfully deploying AI projects that drive real value for your business is to start with a commercial objective that needs to be achieved and work back from there.
This end-to-outcome approach has been championed by AI platform Peak. The company pioneered the Decision Intelligence (DI) category – DI being the application of AI to decision-making – and provides a platform on which businesses can build and quickly deploy DI applications that deliver on business needs.
Caption: Peak
Peak’s platform provides one environment where technical teams can manage data, model, evaluate, orchestrate, and deploy DI apps, and another where commercial teams can visualize, share and interact with outputs.
Uniting both technical and business users on one platform might sound like common sense, but it’s an atypical approach from an AI/ML platform. Yet, it’s one that addresses the main issues actively hindering the commercial deployment of AI. Very few data scientists have the business experience to build a solution that has real utility for end-users. And end-users, in turn, are typically skeptical of mysterious, ‘black box’ models.
A collaborative, cross-functional approach to developing and delivering AI projects is the key to ensuring they deliver meaningful value to the business. Technical teams gain the business understanding from their commercial colleagues to ensure model outputs are useful and business users are bought into the project from the start. They know what data is fed into the model, the guardrails in place and what it’s designed to deliver – because they had a hand in deciding each aspect. The entire team is invested in the success of the project, and once the model is deployed, they have an established relationship and can easily provide feedback. In that way, the app can continue to be iterated and improved.
Source: Unsplash
It’s an approach that increases return on investment (ROI) and reduces time to value at a business level. But at a people level it is equally beneficial. Successful digital transformation is as much about culture as it is about tech, and a fear that AI will eventually replace personnel stops many end-users from fully engaging with the technology. By bringing people into the conversation from the start, businesses can create a culture that engages with and embraces the tech, so it’s not only deploying AI that meets a real business need but has a workforce incentivized to actually use it.
To learn more about Peak and how an end-to-outcome approach can help your business leverage AI, book a platform demo.
10 April 2024
