For CIOs who feel behind the eight ball on artificial intelligence, here’s a bit of good news: You’re probably not.
According to a survey of 83 Gartner clients, 60% of respondents reported to be in an AI “knowledge-gathering phase,” 25% said they are piloting an AI solution and a mere 5% of respondents said they have implemented an AI solution.
The sample size is small, but the message is clear to Gartner analyst Whit Andrews. “If you’re in a place where you have multiple implementations of AI, you can absolutely stop and congratulate yourself,” he said during a webinar that provided a glimpse into material he will present at the upcoming Gartner Symposium. AI laggards can breathe a sigh of relief. “If you have none, you’re not in a place where everyone is ahead of you.”
That doesn’t mean CIOs should sit idle. Developing an AI strategy is imperative because some aspects of AI are delivering value right now, Andrews said. The harder question for CIOs to figure out is where to focus their efforts. Andrews had a few ideas about that, laying out three trends CIOs should keep in mind when crafting an AI strategy.
Trend 1: Natural and contextual interfaces
New interfaces will dramatically change the way consumers and employees access computing resources, Andrews said. Specifically, the new wave of interfaces relies on natural language processing and generation, visual analytics and gesture interpretation — technologies powered by AI. They will play an increasingly important role in enterprise interfaces.
Gartner predicts that by 2022, more than 80% of enterprise IoT projects will have an AI component. That’s up from less than 10% today.
In a client example Andrews titled the “Warehouse of Babel,” artificial intelligence is bridging a language barrier for a European-based warehouse. The warehouse is now using a natural language interpretation system so that employees, who come from all corners of Eastern Europe, don’t have to speak the same language to communicate or access applications “in a comparatively unified way,” Andrews said.
These natural and contextual interfaces are being embedded into consumer products right now — Amazon Echo and Google Glass are two examples. As these offerings proliferate, they will raise employee expectations for workplace tech, Andrews said, with good reason. “[These new interfaces provide] a more effective way of allowing people to use AI as a means of interacting with content, with social situations and systems, with business applications, with data and with documents,” he said.
Trend 2: Smarter IoT and better application integration
AI capabilities are being embedded into the internet of things (IoT) devices that operate on the computing edge, but those capabilities will be limited. Model building with AI will happen elsewhere, but runtime analysis and “interaction into action models” that provide, say, visual analysis can live on an edge device, Andrews said.
For example, Raspberry Pi, an inexpensive microcomputer, can use basic visual analytics to recognize colors and shapes. And in the Warehouse of Babel example, Andrews said the client is looking to invest in a visual analytics application to more efficiently load shipping containers and reduce unused space. “I think of this as being like Tetris,” he said.
Indeed, Gartner predicts that by 2022, more than 80% of enterprise IoT projects will have an AI component. That’s up from less than 10% today.
Companies are also beginning to integrate AI into existing applications to create a more fluid experience, predominately for customer-engagement applications and call center service and support applications, according to a Gartner survey.
The net effect? It’s the beginning of “an AI ecosystem that is intriguingly complex,” Andrews said. He expects the pattern to continue, especially for early adopters who, according to a Gartner prediction, will use four virtual personal assistants on average by 2022.
Trend 3: Rise of AI-enabled computing ecosystems
AI-powered applications will be able to tell each other what they need to meet a goal without human interaction. But to create this kind of commonplace AI, application diversity is crucial. “In any ecosystem, strength comes from that diversity and from multiple perspectives,” he said. “It’s something extraordinary that allows a system to be resilient.”
He said the models for AI-powered application-to-application interaction will mature in a predictable fashion. Initially, vendors will build a “branded AI, so that two applications from the same vendor will work together better than two applications from different vendors,” he said.
That will eventually give way to a system clearinghouse, which will connect separate and independent AI-powered applications, according to Andrews.
The integration capability, in turn, will make way for complicated business scenarios and relationships. Andrews provided an example: In an effort to gain visibility into its supply chain, a large manufacturer has placed an AI module into its supplier’s environment to collect data.
“When you get to this level of integration, you get to very interesting questions around firewalling of data, around transparency, around the nature of the data gathered, around the publicness or non-publicness, around the alerts being generated from that system and within that system. I think all of these are very important questions,” Andrews said.
PlatoAi. Web3 Reimagined. Data Intelligence Amplified.
