“Less is more.”
This turn of phrase, one that feels as old as time, actually first appeared 168 years ago in the poem Andrea del Sarto by Robert Browning. About thirty years later, it was notably used once again by architect Ludwig Mies van der Rohe when referring to the desirability of less visual clutter when building homes.
But hey, you’re reading Future of Work. Why Browning and van der Rohe? And what does the etymology of a random phrase have to do with any of this?
Trust me, it all ties together.
At this year’s ITEXPO at the Broward County Convention Center in Fort Lauderdale, FL, (taking place from February 14-16), a session I was able to attend yesterday was titled “Doing More with Less: Integrating AI and Automation into Modern Workflows.”
So, these two statements aren’t mutually exclusive: Less is more, and we can do more with less. (When we play it smartly.)
Enter AI and how we can automate smarter workflows. This Enterprise Solutions session was moderated by Akshay Sharma, and the speakers were David Hartmann and Roger Wallman. Hartmann is the Founder and CEO of The SilverLogic, a client-focused and value-driven custom software development company, and Wallman is Senior Manager of Solutions Marketing at Avaya, a customer experience innovator.
Sharma presented several intriguing inquiries to the speakers.
“Let’s jump to the chase,” Sharma said. “Will AI become part of workflow logic?”
“Yeah, certainly,” Hartmann said. “I mean, from an experimental point of view, it’s already playing a role. ChatGPT and other Generative/Conversational AI models, albeit in early stages, are doing this. I know of a lawyer who recently used ChatGPT to write, believe it or not, an expert witness statement. That’s a glimpse of where AI is at, workflow-wise. We’ll just want to make sure we don’t bench original human thought.”
“ChatGPT already fits into the workflows in my world too,” Wallman said. “The practical applications show great deals of promise. How about reading through a fourteen-page transcript from a 45-minute meeting? ChatGPT can provide a summarization of that meeting’s major points. Or how about certain work items identified during that meeting? You can tap AI to note and prioritize items, and then automatically build a task list on the fly.”
The speakers also explained that businesses require architecture-ready states to support AI. From ChatGPT to CRM automations, to noise cancellation and customer sentiment analysis, having the right architecture in place will allow businesses to more freely pick and choose which AI models best fit their use cases.
“For example,” Wallman said, “Let’s say you’re hosting a conference, much like ITEXPO. Someone, somewhere, will be performing transcription duties. A colleague of this transcriber might be doing translation, as well. And of course, a moderator gauges the space; are attendees paying attention? Are they taking notes? Are they enthused and engaged? Through AI’s capabilities, these work outputs can be assembled into a single augmented workflow to remove manual impediments.”
Switching gears, Sharma asked about voice and video. The speakers agreed that AI can take in-person meetings and transition them to remote meetings; the world has seen a phase of this with work-from-home technologies in light of the pandemic. With AI, it can be taken a step further; meetings can be simulated as if all participants are actually in the same room, but aren’t. Speakers stuck in traffic could attend and appear as if in a conference room chair. With AI, individual video streams could be overlaid on top of presentation content. “We have enough power now,” Hartmann said, “with GPUs and the cloud and whatnot, that we can do this in very real time.”
In our workflows, AI (with defined operating procedures) can liven customer service interactions and automate multi-platform orchestrations, action VoIP delay decreases, interview recording analyses, generate photos of workplace operations, and process metadata. Not without fault, but it can elevate what we do and how efficiently we do it.
Edited by
Alex Passett
