Building artificial intelligence (AI) that performs on a par with human beings is difficult. But technology companies are beginning to have success delivering AI solution stacks for specific industry verticals.
Vertical AI is about to change the competitive landscape
While some of tech’s biggest players are pursuing general AI, other companies are focusing on using machine learning algorithms, industry-based subject matter expertise, and stores of proprietary data to reduce costs, increase productivity, and improve accuracy. This shift to vertical AI promises to further automate time-intensive tasks and free up human capital for higher-value activities.
Al Carmona, an executive vice president with Mars & Co, a global strategy consulting firm, says vertical AI solutions are part of a second wave of digital transformation. This wave will further disrupt current business markets by allowing nimbler and more data-driven companies to innovate and out-compete entrenched market leaders.
“The adoption rate for vertical AI will start to accelerate within a year or two,” Carmona says. “Ultimately, it is going to cause a huge reshuffling of the deck in the corporate world.”
Vertical AI delivers value by addressing specific problems
Unlike horizontal AI, vertical AI is intended to solve isolated problems where human participation is a significant part of the data workflow, such as an issue in demand planning. Implementing a vertical AI stack typically involves assembling a mix of technical skills and domain knowledge.
Combining different skill sets enables vertical AI teams to leverage domain knowledge and pick out the best opportunities for improving value chains. Teams can use this approach to map out how planned AIs might touch and improve particular workflows.
Teams also can use this strategy to create end-state visions and set expectations for key stakeholders. It also allows them to communicate realistic goals to business executives who may have only a high-level understanding of the technology.
“The idea of using this technology is enormous,” says Dr. Mohamed Aly, founder and CEO of Seeloz Inc., an emerging leader in using vertical AI for process automation. “We need to be clear about what we are addressing. With vertical AI, we are tailoring solutions to the needs of a given department or business user.”
Vertical AI engagements should follow a crawl, walk, run strategy. Even focused AI efforts require assembling massive volumes of data and iterating algorithms until they produce desired results. Stakeholders must explore critical questions, such as: Does the team have the right data structure? Is the team trying to solve the right problems? Are there discontinuities in the data stream that will impede results?
“What you want to do is attach the technology to a specific outcome or result,” says Geoffrey L. Glaser, the senior director for North America healthcare at SAP America. “If you pick a beginning point that can positively affect an organization’s revenue cycle, then you are going to get faster adoption.”
A methodical, iterative process is essential for capturing needed proprietary data and designing an effective interface. It enables developers to build models that drive high-value functionality in a virtuous cycle. Incremental problem-solving is the key to achieving broader business transformation.
If the AI experts solve a big problem with a lot of variables, they become heroes on that specific topic, Carmona says. This “win” builds credibility for the technology and creates organizational momentum for tackling other issues.
Vertical AI ultimately will be disruptive
As vertical AI is adopted, it will challenge established ideas about how entrenched processes should unfold. These changes will be disruptive—particularly since they will often require employees to trust that the AI is making good decisions even as its actions depart from established practices.
“Vertical AI is going to create a lot of change-management challenges,” Glaser says. “People are not used to having their thinking or behavior challenged by software.”
Successful AI engagements will increase corporate and organizational accountability by pulling back the veil on extremely complex processes. These insights will help decision-makers gain confidence in AI solutions and their ability to deliver actionable intelligence in real time.
Ultimately, vertical AI will deliver numerous opportunities to manage and optimize complex workflows without routine interventions from human operators. AI experts suggest that it is easy to know how this change process will begin. What is less clear is how it will end.
Want to learn more? Listen to “Vertical AI: The Path to Enterprise Transformation” on SAP Radio and check out @SAPradio on Twitter.