By Alan Trefler | May 21, 2018
Alan Trefler is Founder and CEO of Pegasystems, and a visionary and expert in the field of customer relations for over 30 years
The industry is abuzz about artificial intelligence, or AI, and specifically the intersection of AI and CRM technologies.
Take the buzzwords away, however, and this is about using smart systems to better engage customers and to make it easier for people to get their work done. That is something I’ve been involved with since the early 80s. In fact, I began my career in some of the earliest days of AI, teaching computers how to play chess. When my company was first founded, we took that same approach—leveraging business rules and workflow to teach computers to act like skilled humans—to automate processes for banks. Over the years, the technology has evolved. Rather than static business rules, we use machine learning and predictive analytics. We now complement workflows with robotic process automation. And the massive amount of data and connectivity available in the digital world has increased our clients’ ability to build streamlined experiences and engage customers, even proactively.
After doing this for more than 30 years, we’ve learned some lessons. As you and your organization look to leverage the latest AI technologies, I’d suggest you ask five questions to make sure you get the most business value out of your AI investment:
1. How do I manage and optimize my AI with rules? AI—especially machine learning—can do incredible things to turn data into optimized customer engagement. But who is making sure the machine isn’t breaking laws? Am I optimizing my customer engagement to ensure retention? Grow share of wallet? Create more advocates? Reduce cost of service? Even the smartest “machine learning” system needs to be pointed at the right business outcomes and properly guided. The right business outcomes aren’t the same in every scenario. There are some things, like regulatory requirements or cultural expectations, you don’t want your machine to learn in real time. Do you have AI technology that provides ways for your business experts to blend predictive and machine learning algorithms with rules to get the right outcomes?
2. Can I build a centralized “brain” that works across all channels? To be most effective, your customer engagement “brain” should provide consistent experiences across marketing, sales, and service, and across all your communication channels. Many of the AI solutions hitting the market today (or promised for tomorrow) are cobbled together from a bunch of acquired technology, much of which was built as point solutions for individual channels. I’ve cautioned before about “Frakenstacks” -- collections of disparate software tools bundled together by mega software vendors. Today I’m seeing signs of a new “FrankEinstack” – a collection of AI tools glued together from multiple software vendors that won’t give you what you need and is likely many months if not several years from being real. It sounds good, but ask yourself: How will your “brain” ensure customers get the same treatment in all channels if each channel is running its own brain?
3. Can I tie insight directly to action? Using AI to make decisions and offer proactive outreach to clients is great, but decisions are made valuable only when you can take action on them. How will you tie AI to the ability to orchestrate outcomes, even those that must cross multiple legacy systems and organizational silos (such as opening a new account or processing a service request), and then use the results of that process to improve future outcomes? In our world, a brain without the muscle to get things done is a waste.
4. How do I feed my AI system with distributed data? Data is the lifeblood of AI, but if you are like most organizations, your data doesn’t live in one place, and it definitely doesn’t all live in the cloud. How smart will your AI be if it can’t see your core business systems and use that data or respond to those events? You need an AI platform that can run in the public cloud, but also in your private, on-premise cloud, where it can safely and securely access the legacy and transactional data it needs to make smart decisions.
5. Is it real? Or is it marketing? Focusing AI on solving the real problems of engaging customers to drive revenue and improve experience is hard to do. Lots of companies are talking about the promise of AI but they aren’t necessarily providing tangible, valuable, outcomes.
So, I say: “Danger Will Robinson!” The hype cycle machine is at its radioactive peak. Before you jump on board, ask a few fundamental questions. It could save you a whole lot of disappointment and trouble down the road.
Another version of this article first appeared on LinkedIn