Artificial intelligence is on the top of everybody’s agenda, but few companies have a comprehensive strategy in place.  An InfoSys survey found that 86% of organizations have middle- or late-stage AI deployments in play now and they view AI as a critical to the future of their business. 53% said their industry has already experienced significant disruption due to AI. By AI, these leaders are really referring to a collection of related technologies, including machine learning, decision support, process automation, natural language processing (NLP), virtual service agents, chatbots, deep learning, image analytics, big data visualizations, biometrics, cyberdefence and pattern recognition. 

Over the next few years, AI will change every aspect of modern life, more dramatically than the mobile revolution did. From voice assistants and tools that listen and learn, to more advanced machine learning that can find patterns in large volumes of data and help us make discoveries that would otherwise be impossible, to applications that can improve back end processes and the customer experience.

Cognitive Maps

Our own suite of tools, including the InnoSurvey® and the ideaton360 platform, depend on AI to find patterns and explore the data more deeply. The right application of AI tools can help company leaders challenge their biases or see possibilities that lie completely outside their own cognitive maps. Some applications have even begun to tackle the complexity of understanding human sentiment and emotional nuance.

Intelligent interfaces are learning how to interpret your emotional state based on the tone on your voice, or gestures, expressions, and visual cues. Analyzing these inputs, smart systems can respond accordingly, and even predict how people will behave. This technology will transform many different spheres of economic and social life, including retail, entertainment, and even politics.  

Fast Followers Are Not Fast Enough

This is not happening in some distant future. AI is already impacting every business, either indirectly or directly, because AI-based customization define customer expectations and set a new baseline for relevant services.

Unlike other technologies, however, fast followers may not reap the benefits of waiting for market maturity. A study in Harvard Business Review concluded, “By the time a late adopter has done all the necessary preparation, earlier adopters will have taken considerable market share — they’ll be able to operate at substantially lower costs with better performance. In short, the winners may take all and late adopters may never catch up.”

The main reason why waiting may put you on the wrong side of a digital divide is that it takes time to integrate AI into your workflow. Off-the-shelf AI programs really don’t add a great deal of value to the organization compared to a customized, tailored solution. Companies that integrate AI across the organization gain the greatest advantage, but it takes time to configure the software, set up interoperable channels, and find the right APIs to make information transfer seamless. For those exploring how to generate next-generation customer experiences using machine learning, your implementation team will need a great deal of data spread out over time to train the AI on your customer base. Competitors who have already started collecting that data are at a distinct advantage.

Another reason why time is a critical factor is that your workforce needs training to manage and collaborate with AI effectively. This can be another time-intensive project all its own. Either hiring the talent you need and integrating them into your company culture, or training the skills you need into your current workforce will push the time to peak efficiency further down the road.

The Dangers of Delearning

The third type of obstacle in optimizing AI has to do with governance. Data rot is a serious problem because algorithms based on consumer behavior start to decay unless you assign resources to keep data up to date and decommission data that is no longer valid. Data monitoring addresses problems related to fraud and “delearning,” a growing problem where malicious actors deliberately feed in bad data to compromise an algorithm’s accuracy.

Although that might sound far-fetched, delearning is already happening. The journal Science recently published an account of researchers who were able to successfully use “adversarial noise” to convince a medical AI algorithm to diagnose benign moles as malignant tumors, with 100% confidence in their diagnosis. By carefully examining the evidence, they found examples of adversarial data for nearly every type of machine-learning model and across a wide range of data types, including images, audio, text, and other inputs. A strong governance model is the key to monitoring and correcting for corrupt data, whether that data is intentionally misleading or simply out of date.

These three represent just a few of the most important reasons why it may be difficult to keep up with those that have already started implementing AI projects are pulling ahead in their respective markets. There remains one way to make up for lost time, however. An intelligent deployment strategy, built to enhance your company’s market strengths, can help erase some of the advantages that early adopters have gained.

Innovation360’s CEO Magnus Penker advised, “An AI strategy is unique to each company. They must analyze their own organization’s prerequisites, needs and readiness, among many other things. AI should start through experimentation (test hypothesis and learning in a small scale), then you can scale it up to a full implementation once you have tried and learned about it.”

A Meeting of Minds

For companies where AI has become more prevalent, leaders tend to put a higher priority on investing in and motivating employees, according to a new study by Microsoft. The more tasks that are automated, the more companies can clearly see the value of skills that are uniquely human. Those skills include creative problem-solving, empathy, the ability to establish trust with customers and collaborators. Skills like these are becoming even more valuable in a business climate defined by rapid change and uncertainty. The study also found that high growth companies are more than twice as likely to rely on AI augmentation for help with direction-setting and strategic decisions.

The place where physical and digital brains come together to become more than either on their own is the true frontier of innovation. Strategic assessment of your innovation capabilities related to AI can help you stay ahead of competitors, but time is short for closing the gap.