CIO Insider

CIOInsider India Magazine

Separator

Cutting Through The AI Noise

Separator
Don Schuerman, CTO & Vice President - Product Marketing, Pegasystems

Don Schuerman is an industry professional with over 20 years of experience in Digital Process Automation, Robotics, Low-code, CRM, Case Management, Rules and Decisioning technologies.

The term artificial intelligence was first used in 1956 by John McCarthy, a computer scientist who’s known today as the father of AI. But today, AI has grown to mean many things to many people. It’s been called “one of the most elusive topics in computer science” because of its scope and complexity. Given the rapid pace at which technology progresses, the AI category is a bit of a moving target.

For many of us, much of what we think we know about AI comes from Hollywood – the heartless HAL in 2001: A Space Odyssey, Tony Stark’s Javis, the lovable Wall-E, the enigmatic voice in Her. But autonomous robots are in many ways an edge-case for AI. Most AI doesn’t live in the Hollywood world of sentient – or at least semi-sentient – beings. Most AI is focused on much more specific tasks and is so entrenched in our lives that most of us use it every day without even realizing it.

AI is Everywhere
AI is the brains behind our GPS, our smart phones, and the games we play. It recommends relevant products and articles we might have interest in. It translates menus we can’t read and alerts us when the car we’re driving drifts outside of our lane. Whether we like it or not, it learns from our purchasing patterns and browsing history. It identifies suspicious spending patterns and sends us warnings. When you really think about it, AI is almost as ubiquitous as oxygen.

We often talk about AI like it’s one technology, but it’s actually a family of different technologies that are used in various ways. At the most general level, we can define AI as software technology that simulates human intelligence and reasoning with algorithms – a fancy word for processes and formulas that do calculations, make decisions, and

solve problems. Most AI algorithms are developed from either business rules – things humans know that they then teach a computer – or are learned from finding patterns in sometimes massive sets of data. The math and theory behind machine learning algorithms were first developed in the 1800s, with machine learning computers like Alan Turing’s code-breaking machine featured in The Imitation Game dating back 50-60 years. Newer technologies like deep learning exploit the massive data sets and cloud-based computing power of the Internet to build high-level data abstractions with multiple processing layers. There’s also speech recognition, biometrics, natural language processing (NLP) and generation, virtual agents and chatbots, AI modeling, and automation technologies like robotic process automation (RPA) and digital process automation (DPA). The AI landscape is wide and varied.

Newer technologies like deep learning exploit the massive data sets and cloud-based computing power of the internet


AI is on everyone’s radar these days and businesses everywhere want it because they know it can help improve customer experiences and boost profitability. But there is a huge amount of confusion – and fear – around this category, and many people aren’t clear on what they really want. Asking about a company’s AI strategy is like asking what their software strategy is – there’s no one-size-fits-all solution.

AI isn’t its own Special Bubble
It’s time to stop treating AI like it’s in its own separate bubble. We should start treating it like it’s a core part of all the software we’re building, because it is. My colleague, Rob Walker, talks about how AI can help engage and build long-lasting relationships with customers by infusing AI with human-powered empathy. When it comes to automating menial tasks, processing data, and impacting efficiency, AI is invaluable. Why should humans sort through emails when AI can do it for us?

Some AI is becoming commoditized, so it’s not necessary for enterprises to reinvent the wheel. For instance, if I’m an executive at a bank, I don’t need to leverage my proprietary data to build speech-to-text functionality. I should be more than happy to go to Google or Amazon or someone who has massive amounts of data and buy their already trained AI models for speech to text or image recognition.

But then there’s a set of AI capabilities I think should stay proprietary – in a good way – to each enterprise. To stay with my banking example, I don’t want to use a commoditized algorithm that gives my customers exactly the same offers other banks do. Instead, I want AI that lets me define my business goals, customer needs, and brand values and combine that with learning from the customer data that I manage to build a unique experience for each of my customers.

AI is a huge, growing category, and the way we talk about it needs to expand as well. The AI that powers Alexa is not the same AI that powers customized, contextual interactions with customers. And you can’t lump them all in the same giant bucket. I doubt Hollywood is planning a blockbuster about powerful AI algorithms that recognize images, automate emails, or improve customer experiences, but that is where the real value of AI for today’s enterprises is found.

Current Issue
Cloud Migration In Manufacturing Industry