Five reasons why AI algorithms can be difficult to implement in operational management

By: Evgenia MalinaHead of Business Strategy and Operations at food rocket.

According to a recent study by the McKinsey Global Institute, AI is poised to boost global economic output by $13 trillion by 2030.

However, this comes with its own challenges and unintended consequences. A part of the most frequently identified risks and challenges of AI implementation include issues of confidentiality, inability to generalize, and a general lack of trust.

Below are three AI-related challenges that specifically affect business operational management.

1. Poor data quality

The first thing a robust AI algorithm needs is data. To properly train an algorithm, you need to feed it large amounts of accurate, high-quality data. Unfortunately, it is not always easy to obtain this data, and a 2020 Gartner Report notes that poor data quality can cost your business around $13 million every year.

For example, some processes may not have a thumbprint at all when you start. There is no data to feed into an algorithm in these cases. Everything you give is just assumptions and educated guesses, which has two problems.

First, it introduces significant human bias in your process from the start. Second, it means that any results from the algorithm are simply an extension of your best guesses. Ultimately, this leaves you with a patchy data landscape and an unreliable and unstable decision-making process.

2. Navigate “Cold Starts” and Employee Engagement

Automation is great for streamlining existing processes, but the trade-off is “cold start”. This is when you need to start a process without historical data for the AI ​​to base its routine on. In any case, the AI ​​will have a hard time overcoming this obstacle.

According to Harvard Business Review, 80 percent or more of an IT team’s time is often spent trying to improve and refine inconsistent data for AI algorithms

It often takes a considerable investment of human effort to help the AI ​​overcome this “cold start” bump and resume operations smoothly.

In my experience, this can cause serious disruptions in supply management and it can also cost companies considerable revenue. We all know that AI is not yet mature enough to handle all aspects of an operational management system. This means that any AI solution used by your business will overlap with human decision-making processes.

While this can be a good thing, it can also lead to a disengagement from an employee’s sense of personal responsibility. In some cases, employees feel they can dissociate themselves from a decision because “the AI ​​did it”.

Additionally, it is common for the introduction of a new algorithm to coincide with a significant drop in quality metrics. In my experience, this paradox is the result of someone who was previously responsible for metrics feeling like they are just an unimportant link in an automated decision-making chain.

Managing this aspect of automation is critical because of how easily it can lead your team down a path of disinterest and reduced engagement. It also has the potential to harm your brand. If the AI ​​is left to make decisions on its own, it may unintentionally start to discriminate against customers in certain age, gender, or geographic brackets.

3. Challenges related to transparency and effective implementation

As every business owner knows, things can change in an instant. Companies don’t always have the time to create a complex AI solution for a new operation.

In fact, it’s far more common for companies to run out of time and be forced to fix a problem without the help of automation because setting up a new process is simply taking too long. Since there is usually no time to write complex models, one of two things happens.

First, the business may choose to implement a mostly complete process, but insert a manual intermediate step until the process can be refined. In this case, companies lose up to 80% of the calculated efficiency of the process.

Alternatively, the business can deploy SaaS to speed up the implementation process. Although the costs in time and money are lower, the problem of loss of efficiency remains. In this case, the implementation of SaaS algorithms not specifically adapted to the needs of the company can make a process less efficient than if it were done manually.

In addition to these issues, it’s important to understand that the transparency of the AI ​​process is incredibly difficult to communicate to management, even by experts. This is due to the complexity of the algorithms, but it can make your team hesitant to move to automated operations management.

Where do we go from here?

Some researchers suggest that the challenges we currently face will lead to new human roles in a business: Trainers, explainers and supporters.

Trainers will help optimize AI performance; explainers will be responsible for breaking down AI decisions for non-professionals, and supporters will work to make AI processes sustainable in the long term.

However, until then, companies and founders need to consider more than just the competitive advantage that AI can give. The benefit must be weighed against the ambiguity, time costs, and barriers to growth that come with operational AI.

Artificial intelligence still has a long way to go in terms of growth, development and implementation. It can undoubtedly make a huge difference in operational management, but we cannot yet fully rely on it to always be the best option.

Sharon D. Cole