IBM Watson: On leveraging AI to improve productivity

Thoughts from IBM on getting AI to work for Productivity

Jay Migliaccio, IBM Watson: On leveraging AI to improve productivity

IBM has been refining its AI solutions for decades and knows a thing or two about helping businesses leverage the technology to improve productivity.

In 1997, IBM’s Deep Blue supercomputer was used to beat World Chess Champion Garry Kasparov. At the time, all too familiar headlines suggested that computers would soon replace humans. Over two decades later, AI has proven to be an assistive tool that benefits us every day.

IBM Watson’s first commercial application was announced a little over a decade ago in February 2013 for utilisation management decisions in lung cancer treatment. In the years since, we’ve seen it used to deliver game-changing advancements in healthcare, weather forecasting, education, science, and much more.

AI News caught up with Jay Migliaccio, Senior Product Manager for Watson Orchestrate, to learn how IBM is now using its vast experience to help businesses with their digital transformations.

AI News: So, Jay, can you tell me how IBM is helping businesses to improve the productivity of their workforces?

Jay Migliaccio: Yes, Ryan. Thanks so much for the invite and for asking me here.

IBM is expanding its suite of offerings in the area of digital labour. Digital labour leverages AI and automation to help workers become more productive. And, much like human labour, digital labour performs work on business systems through “skills”.

Digital labour skills enable digital labour to interact with business applications, much like you and I would interact with a system of record or system of engagement. We can do this now through digital labour. And, what’s new and unique, is that digital labour leverages the human-centric interaction style.

So, we’ve introduced natural language and we’ve also introduced intelligent orchestration to be able to execute not just single skills, but actually multiple skills to be able to achieve higher-level tasks.

AN: Generative AI is a hot topic in the market at the moment. Do you see that being used practically in the workplace and what risks should businesses be aware of?

JM: Yeah, great question. I actually do use it myself in the workplace, I occasionally have to develop software tools and simple scripts and I have had it generate a number of scripts for me successfully. So I’m impressed not just with its ability to generate verbal and written content, but also code content. I for sure believe it will become increasingly useful in the workplace.

Current generative AI platforms have been trained on the internet, so remember your results may vary. I know anytime I Google or search for things on the internet I take the results with a grain of salt.

I believe that enterprises, as they go to look and adopt generative AI systems, they’ll lean more towards AI that they can trust. Therefore, we need to work on being able to create that trust element in generative AI solutions.

AN: What is the value of Watson Orchestrate for developers?

JM: When we talk about developers, I’m talking about automation developers. And that’s by and large developers that are integrating apps and business apps and business systems to work together.

For the most part, those developers have been integrating business systems to business systems. Now what we can do with Watson Orchestrate is we can introduce the human into the loop.

These automation developers now have a platform where they can build and integrate their automation workflows and they can bring a human experience into these automation workflows for everyday human workers.

Watson Orchestrate provides a platform for creating human-centric workflow automation, designed to interact with humans in our native communication style which is spoken or written word.

AN: How does Watson Orchestrate learn from user interactions?

JM: There are a couple of ways Watson Orchestrate is monitoring the behaviour of humans and learning from them. 

Perhaps most important is its ability to interpret the natural language through which humans are communicating. Today it’s the written word, but in the future spoken word. Watson Orchestrate can not just do a pattern match based on existing known sentences, but it can actually understand the intent of those utterances. 

Additionally, it can extract entities from those utterances. So, when you use proper nouns in a sentence, it can understand that’s an entity that it would use as part of an automation. It can match the intent of the utterance to existing skills that it has and can react accordingly. It can understand the intent of the utterance and then take action on those skills. Increasingly, it can sequence multiple skills together.

Also, we are working on systems for empowering Watson Orchestrate to monitor the user’s behaviour. And, just like any modern SaaS application today that has recommendations based on your behaviour, we’re working on recommendation engines to recommend to employees how they can use Watson Orchestrate to be more productive in the future.

AN: Talking about AI more generally, what new ways of working are today’s advancements enabling?

JM: As I just alluded to, we’re increasingly empowering systems to understand human natural language to a much more complex and sophisticated extent. Natural language interpretation has grown way beyond the basic pre-programmed bot experience.

I’m sure everybody has had an experience on a website where there’s a bot responding to your basic questions. What we’re trying to do is make that bot much more intelligent. The current generation of digital labour can understand your intent, extract entities from your utterances, and, most importantly, take action on your behalf.... '