Explainable Decisions are the Key to AI Adoption
One of the greatest challenges facing business leaders considering implementation of AI is organizational resistance towards the adoption of AI-driven business tools.
Technology changes the world constantly and there is a long list of previously successful businesses that failed because they didn’t evolve as technology moved the world forward. Today, AI is transforming the way business is being done across industries in a way not seen since the industrial revolution.
AI-driven business tools are propelling future-focused companies to success and to outperform their competition. However, when it comes to adopting AI, the primary reason that many organizations struggle is down to trust.
Employees who have until now fulfilled their roles successfully, making decisions based on their tribal knowledge and experience, don’t trust in the decisions and recommendations that AI-driven tools provide and have resistance to handing those responsibilities over to an algorithm.
What Are the Leading Reasons for Mistrust in AI Systems?
Assuming that the organization is implementing the right technology in the first place, there are two primary reasons for this lack of trust:
- There is limited understanding of AI technology within the organization
- The AI, and the technology, is a black box
In AI, as with many advanced analytics systems, a black box is a system or program that only permits visibility to the input and the output but provides no understanding of the mechanisms in between the two. Most commercial sellers of AI systems will allow their customers a look inside the workings of the system. But even with this, companies often don’t have the expertise on staff nor the organizational capacity to gain any deep understanding of AI.
How Do Explainable Decisions Build Trust in AI Systems?
This is why suppliers of AI technology are turning to the explainable decisions approach, and it is working. Rather than providing a masterclass in the inner workings of AI technology, the emphasis is placed on providing a clear understanding of the reasons behind the decisions and recommendations that the AI is making and combining that with aligned success metrics.
It is the same approach that organizations already take to measure the performance of employees: asking them to explain decisions and evaluating their performance in reaching their goals. We don’t look in to how their brain technically functions.
How Explainable Decisions Function in Retail Merchandise Planning
There remains some education required at the rudimentary level of the technology, but it is not the focus. As an example, let’s look in the area of retail and the AI-driven Merchant planning tools that are being adopted by retailers today. Retail is ideal for the employment of AI due to the millions of variables that have to be considered weekly, daily, and even hourly.
Let’s suppose that the AI recommends that hotdogs are put on sale at a very low price, perhaps even a price where the retailer will lose money on each sale. In and of itself, this recommendation seems illogical. However, when the AI decision is explained (this will increase traffic as more customers will come in for this offer - these customers will then automatically buy ketchup, mustard, pop, and buns at a slightly increased price), it is clear that the associated sales more than make up for the lost margin on the hotdog promotion. With each decision clearly explained and justified, the retailer begins to trust and understand the AI technology.
Obviously, the decisions made across hundreds of thousands of SKUs and price points being sold through multiple channels and being advertised across dozens of media channels can’t all be reviewed. There simply isn’t enough time in the day.
But the Merchant can question a large sample of them, each time gaining a greater understanding and, more importantly, building trust that the millions of other decisions are being made correctly by an order of magnitude greater than humans can.
Explaining the logic behind the decisions of the AI, rather than focusing on the complicated technology involved, quickly builds trust within the organization, allowing for swift adoption of this technology – vital to succeeding in this pivotal moment of transformation for business.