Breakthrough Barriers for AI Breakthroughs

As technology develops, business must develop with it. Companies that want to keep up with society must find ways to effectively integrate new technologies that will benefit them and allow them to continue to grow. AI technology is relatively new to both business and the technology field. Successful adoption of Artificial Intelligence and Machine Learning (AI/ML) technologies beyond bleeding-edge tech firms is unique and may not have as much to do with the technology itself as one might believe.

Changing Your Mind(set)

In addition to the obvious technological prowess required, companies must examine—and possibly change—their cultural norms. In order to fully take advantage of AI, a company must address the changes they need to make in management styles. Management needs to be aware that they cannot micromanage the process. Trust is essential in AI integration. Company leaders need to trust the system, and their teams. This ties into another struggle which is that AI integration does not yield immediate results. AI makes it possible for machines to learn from experience, adjust to new inputs and perform human-like tasks and relies heavily on deep learning and natural language processing. Using these technologies, computers can be trained to accomplish specific tasks by processing large amounts of data and recognizing patterns. Advanced algorithms are being developed and combined in new ways to analyze more data faster and on many levels. This intelligent processing is key to identifying and predicting scenarios/outputs and takes time. Realistic expectations must be set and communicated to avoid the making premature adjustments, which can be detrimental. AI systems have a give-and-take relationship with the people who operate them. Having only one team dictate necessary changes for integration will stunt the growth of the AI at the core of the implementation.

For a company to best prepare for full integration, management needs to allow room for trial and error. The teams setting up AI systems will make mistakes, as will the system itself. This is part of the process, and an important part. Every company is different, and so are their customers and employees. AI works with these differences to create a comprehensive understanding of everyone involved in a company’s process. This is messy work and takes time and corrections. We expect tech to automatically adapt to our needs, and AI is not as simple as that.

AI adoption cannot grow out of only one progressive department. For AI to be integrated successfully, all departments need to be involved. AI’s purpose is to recognize patterns in human behavior and build an understanding of the consumer and/or company’s method. Because of this, diversity is critical to a properly functioning AI system, and a siloed approach will not provide that.  Historically, this has been a major problem with AI effectiveness. The field itself lacks diversity , Black researchers represent less than 12% of the US AI workforce, and women make up less than 15%. This makes it almost impossible for the system to be inclusive. If the Data Scientists have inherent biases, the system will too. Nobody is implying that data scientists as a group are personally bigoted. That said, the occurrence problems like misclassification of non-White individuals through image recognition, or the radicalization of chatbots, or the failure of technologies to recognize darker skin tones must be addressed. There needs to be focused energy in developing diversity in a company’s AI system in order for it to actually serve its users. 

Adaptation and Evolution

Agile, responsive, and adaptive cultures are rare, yet they are key to a successful adoption of AI/ML. Properly utilizing AI requires a change in thinking and practice on almost every level. Management is arguably the most important level to adapt, as their managing style ripples out and guides their direct and indirect reports both. It’s important for management to allow transparency, mistakes, and to focus on data-driven decisions. Alibaba, the largest e-commerce platform in the world, attributes no small part of their success to their AI system, called Small Smart Selection. Its algorithm is based off of “deep learning and natural language processing that helps in recommending products to the customers who shop and then communicates to the retailers to increase inventory to keep up with the demand of the products and services.” Alibaba used AI to take over certain roles, whether it was product flow management, customer service, deliveries, or coordinating with both customers and retailers. Because they no longer had to spend energy and resources focusing on those parts of their business, they were able to focus on growth and expand. You must work with the system in order for it to work with you.

Cooperate to Integrate

A company must be prepared to make mistakes, which is why transparency is so important. The entire company must work together and trust both the system and themselves. According to some research, “only 8% of firms engage in core practices that support widespread adoption. Most firms have run only ad hoc pilots or are applying AI in just a single business process.” This hesitation can stem from the fear of making those inevitable missteps. However, to not fully integrate across systems, departments, and culture is the best chance to actually lead to failure. Errors should not be a deterrent; rather, they are an indication to make an adjustment. It is vital a company trusts the data AI is providing, otherwise the system is useless. A frequent problem management runs into during integration is their desire to make decisions based off their own knowledge base. Management should exercise appropriate caution while still seeking to avoid internalized prejudices about human capacities. We are more comfortable believing that we can make a better ethical decision than a computer system. This is not necessarily the case. AI compiles such diverse data, from diverse groups (when applied properly), which provides greater objectivity and identifies underlying patterns of behavior.

Besides adaptability, the most common piece of advice given by companies who have successfully integrated AI, is to ensure the process is customer focused. AI serves companies, but it’s important for business to remember that customers are part of the company culture as well. Just as AI needs to be integrated across all departments, it needs to be allowed to be aware of those who are taking part in a company’s services or products. An AI system allows a much faster flow of information between the customer and the company. A business is then able to see what the customer needs and wants clearly and quickly.

As the system augments, AI is able to take over some of the responsibilities of ‘human workers.’ There is a common fear as technology takes over human roles that jobs will be eliminated. However, this automation of tasks allows humans to develop new skill sets and focus on growth with themselves, and then the company.

Artificial Intelligence and Company Culture

It is always difficult to generalize; thinking about adopting AI systems into your business certainly militates against generalities. The cultural preparations necessary for the successful adoption of AI will obviously vary depending on the specific company and industry. There are, however, at least a few constants we can discern as well as some considerations for how AI can itself be part of the cultural assessment of any organization.

Think of how difficult it is to manage a change of email services from something like a Google-based system to a system like MS Outlook. Even something this seemingly simple cannot be accomplished without thoughtful and forward-looking change management practices. Adopting AI solutions will require even more forethought and will likely require greater cross-team collaboration in order to ensure an effective rollout. In the case of changing email servers, the onus falls almost solely on the IT department to ensure a smooth transition. In the adoption of AI solutions, the entire organization needs to be involved.

Preparation for the adoption of AI systems may induce both structural and cultural changes to any organization. Our colleagues at Walking the Talk have investigated the impact of AI on company culture in a recent white paper. One of the more surprising insights they share is that AI may indeed be able to help measure and improve company culture. AI’s ease at processing large amounts of data quickly may enable leadership to respond to employee concerns with greater agility. While innovative and tech-forward companies may be the most obvious beneficiaries of AI solutions, even mission driven organizations outside the technology sector may also be positively impacted by the deployment of AI in the measurement, assessment, and optimization of company culture.

With appropriate preparation and expectation setting, AI solutions can optimize business outcomes. This is true not only of operational efficiency where one might expect to see such improvement; it is also true that AI can help quantify and analyze difficult-to-parse data sets around employee satisfaction and expectation. Companies who can successfully prepare for and deploy AI optimizations will need to enact holistic change management practices, as AI is more transformative than simply additive.

Artificial Intelligence, Organic Talent

Unsurprisingly, AI/ML talent is in high demand, and that often leads to upward pressure on salary. IEEE has tracked software engineering salaries across the past few years. 2021 appears to have been a high-water mark across AI/ML and natural language processing salaries, but machine learning experts are still seeing approximately an 8.5% increase since 2018. Natural language processing has seen approximately a 12.4% increase in the same timeframe, while artificial intelligence engineers are roughly flat, having peaked in 2021 along with the others. Of course, these numbers are related specifically to the engineers and not to their managers. It stands to reason, however, if the engineers who are overseen by a manager and or a C-suite executive have seen compensation increase, those managers and executives would at least have seen concomitant increases.

Whether you believe Clive Humby that “data is the new oil” or Peter Sondergaard that “information is the oil of the 21st century, and analytics is the combustion engine”, it is clear that data remains a hot commodity. Here we do not mean one’s personal data gathered by unscrupulous means from internet traffic. Rather, we mean the use of data to orchestrate and operationalize the value of a company’s AI initiatives whether it comes from an unstructured or a traditional source. Managing that data certainly requires the right technology, but it also continues to require top talent as well.

These trends are not limited to the software and technology industries, either. The innovative disruptors are being targeted by traditionally non-innovative industries to ratchet up innovation themselves. Industries such as financial services or traditional manufacturing are all looking to disruptive technology companies for talent that can think outside of the traditional industry paradigms that have stifled innovation in the past.

At the same time, disruptive technology driven by AI brings about regulatory, compliance, privacy, and security needs that require talent who can be the “adult in the room” in disruptive ventures. Many emerging, AI-driven technology disruptors in the financial services space are looking for people who have the necessary technology interest and skill set but also bring a traditional bank security, privacy, and/or compliance mindset to ensure new products and offerings can be built to scale without running afoul of regulatory bodies.

The line between technology and product leaders can become more porous with AI and ML adoption. As has been the case with prior disruptive technologies, AI/ML has forced a revision to the purview of the traditional product leader. Product leaders must now become more technologically and data savvy in order to lead the innovation side of product, while still retaining the ability to run the commercial side. We have further seen companies begin to separate the Data Product person from the true Chief Product executive, forming both a partnership but potentially an interesting contention between to two as the organization evolves.

Meet the Authors

Let's Move your Organization, Business, or Career Forward

Get in Touch with a Consultant About any of our Talent Solutions

Sign Up
to Receive
Our Newsletter

Our email newsletter delivers unique insights and inspiring anecdotes that can help you take your business to the next level. Subscribe today.

Be a Part of Our Global Talent Network

Fill in the form below to become a part of our talented candidate pool.