a neon sign that says "Data has a better idea" with window view of city overhead
PHOTO: Franki Chamaki

Artificial intelligence has already been integrated into our daily lives in more ways than imaginable. Businesses are also starting to implement AI to make faster, smarter and more complex decisions. This not only enhances the overall customer experience for customers, but also increases the efficiency in all parts of the business, and ultimately drives greater profitability. However, enterprises are adopting AI at a much slower pace than expected.

In 2020, we’ll see enterprises begin to overcome this challenge. We will also see humans become even more involved in generating the data for training and refinement of AI. In the coming year, we can expect enterprises to make strides towards addressing human and AI collaboration challenges from the planning process to the end customer reaction.

Emotional Factors Are Key to Enterprise AI Success

While many predicted that AI would have the potential to automate our work and make businesses more efficient, adoption has been slower than expected. What many fail to consider is that human emotional factors play a larger role in enterprise adoption than they do in everyday consumer use.

For enterprises, the cost and risk of a wrong decision is much higher compared to consumer use cases. This makes workers and business leaders more reluctant to use AI in case its implementation goes wrong. For example, if a consumer asks Alexa or voice assistant about the weather forecast and it doesn’t understand the question, the risk is low — the consumer can choose to repeat or rephrase the question until the AI understands, or simply check the weather manually. On the other hand, if a business AI solution recommends increasing the price of an item to $120 instead of its usual $110, there is a much bigger risk involved. If the price jump is too high, the business may lose customers, and if too low, the business may be leaving millions of dollars on the table. Either way, this will result in revenue loss. With this risk and consequences comes a larger human emotional factor — fear — that plays into this aversion.

In 2020 and beyond, AI adoption in enterprises will become a focal point of interest. Previously, we only had to design products that worked. Now, we have to go a step further when designing AI-based products, and we must consider the human emotional factors in order to make these products more adoptable.

The AI industry has already explored many adoption strategies. One example is “explainability.” Today, AI faces the “black box” problem. While we are able to see the results and outputs that AI helps produce, it is often unclear how AI makes certain conclusions. In the future, we’ll see a greater attempt to make this “black box” more transparent, giving us a more explainable AI that is easier to adopt. Other examples include more empathic AI, more accountable AI, and more ethical AI — all trends that will continue to develop to make AI more adoptable in the workplace.

Related Article: The Next Frontier for IT: AI Ethics

Tap AI's Potential With Business Process Changes

Once human emotional factors are addressed, enterprise AI adoption will accelerate. In addition to seeing “augmented intelligence” evolve — or the joint partnership of humans and AI working together — we’ll begin to see a larger shift in overall business processes as AI becomes more advanced.

AI systems need constant feedback for learning. So the more data and input humans can provide as feedback, the more human-like their decisions will be. While augmented intelligence has always occurred, this form of human supervision can now happen closer to real-time. Because more feedback data can be collected and processed in near real-time, more feedback can be fed into the machine learning engines to help AI learn faster. As a result, AI will become smarter and will be able to automate more human tasks. 

In the past, AI served as a backup tool that helped humans automate simple tasks, meaning many organizations were not utilizing the full potential of AI systems. In many instances, humans still make up to 80% of decision-making with AI helping humans for the remaining 20%, often on the simplest tasks. In the future, this will reverse. AI will cover 80% of decision-making with humans handling the remaining 20% that involves more difficult reasoning, decisions with higher stakes involved, or simply new and never before seen situations.

In addition to the increase in the passive and indiscriminate collection of feedback data, processes will also start to develop where we will see AI learn from select data curated by humans. Due to the scarcity of feedback, AI previously had to learn from all human feedback data. As more feedback data is collected, we will be able to pick and choose the type of data we want the AI to learn from. We are essentially teaching the AI to do and act how we want it to, much like how we teach an intern. New processes that enables quick and frequent human interactions and curations will further accelerate AI learning.

Related Article: Why AI Isn't Mainstream in the Digital Workplace, Yet

The Next Stage for AI

In the upcoming year, we can expect enterprises to make a clear drive to better address the human emotions and interactions as a way to improve AI adoption in business. We will also see humans work more closely with AI systems than before, which will result in more human-like AI that is more explainable, empathic, accountable and ethical. If enterprises want to tap the full potential of AI, they will need to embrace these changes. In addition to driving business efficiency, these shifts will ultimately provide better customer experiences.