IT leaders need to understand some hard truths of Artificial Intelligence tools in order to shape AI strategy. Consider these key questions to discuss with your developers
5 ways Artificial Intelligence (AI) is reshaping IT
How will AI change your IT organization? From the help desk to the front lines of data analysis, new tools, tactics, and relationships are emerging
As Gartner’s most recent hype cycle report for artificial intelligence (AI) points out, AI ranks high on the CIO’s agenda for the next five years as a source of potentially transformational business impact. However, for many IT organizations, AI is not just on the IT leader’s radar as a business enabler: It’s having fundamental impacts on the function itself – from automating some longstanding functions to demanding greater involvement and newer approaches from IT teams.
[ Do you understand the main types of AI? Read also: 5 artificial intelligence (AI) types, defined. ]
AI is beginning to reshape IT in a number of ways that forward-looking IT leaders will want to follow. Let’s consider five worth watching:
1. IT becomes a major AI consumer
Tools to automate traditional break-fix and other IT service desk processes are not new, but they’re getting significant traction these days, says Wayne Butterfield, director of cognitive automation and innovation at ISG. “An IT Service Desk is as prone to repetition (and therefore automation) as a customer service operation,” he says.
That’s not the only area of hyper AI-enabled automation coming for the IT function. “IT has quickly become not just a partner but a consumer as well, leveraging AI for security and system management to automate processes and move at the speed of an AI-driven enterprise,” says Shawn Rogers, vice president of analytic strategy at TIBCO.
2. Shadow IT could expand
IT activities taking place outside the tech core are proliferating, sometimes as a result of AI. From self-service data science and analytics tools to the adoption of robotic process automation (RPA) for functions across the enterprise to business-bred machine learning models, the power of perceived shadow IT functions in the enterprise, is expanding, according to ISG’s Butterfield. The definition of, and the line between, “self-service” and “shadow IT” depends on your culture, of course.
[ Get our quick-scan primer on 10 key artificial intelligence terms for IT and business leaders: Cheat sheet: AI glossary. ]
3. Data science demands deeper collaboration with IT
Some mainstream enterprise applications (think CRM, for example) are baking in more AI and automation. But for more advanced applications of AI, the need for greater partnerships between IT and data science is becoming evident. “The early days of having a data scientist tucked away in the organization are over. Today, data science takes a village and IT is part of that team,” says TIBCO’s Rogers.
As companies prepare to scale their AI and analytics usage, they need deeper access to the systems, data, and applications that IT knows. “Building AI-led solutions requires intense collaboration between the data scientists and engineers,” says George Mathew, client partner for technology services at Fractal Analytics. “While each of these is a deep area by itself, successful teams have enabled these two groups to work together, and in many cases overlap across areas, in order to productionize AI solutions.”
4. IT and data science need shared tools and tactics
The partnership between IT and data science demands that each group adopt each other’s technologies and techniques, says Mathew, “at least for the sake of familiarity, if not for expertise.”
Engineers will need to be able to read source code that pulls data from local data sets, understand exploratory data analysis and feature engineering, and become fluent in algorithms such as Bayesian techniques to generate insights. “They need to have this knowledge in order to refactor and modularize this code so that these can then be operationalized on enterprise IT systems,” Mathew says.
Conversely, data scientists will need to learn how to ingest data via database connectors or APIs, store and process the data in structured stores, and write modular code that can be containerized for downstream consumption.
“We have seen that this familiarity and appreciation of common challenges lead to increased collaboration between data scientists and engineers,” says Mathew, noting that several of his organization’s teams have delivered complex AI-related solutions on the strength of these stronger partnerships.
5. AI governance takes center stage
As organizations implement increased AI-enabled automation and processes, regulatory and reputational risks increase. Gartner points out the importance of creating policies to fight potential AI-related bias, discrimination, and other issues. (Read also: AI bias: 9 questions leaders should ask.)
Again, this is an area where data and IT leaders can join forces, and Gartner suggests focusing on three points: trust in data sources and AI outcomes; data and algorithm transparency requirements; and diversity of data, algorithms, and viewpoints to underpin AI ethics and accuracy.
[ How can automation free up more staff time for innovation? Get the free eBook: Managing IT with Automation. ]