This is part 1 of a 4 part series, sponsored by Alfresco.
We all know that artificial intelligence (AI) is very clever, sometimes a bit too clever for its own good at times. But many enterprises we speak to, don’t think that AI is for them, that it is too complex, too expensive or that they don’t have the in-house skills to work with it. Most struggle even to figure out where to start or, how to map the potential benefits of AI to their current business processes and activities. This slow start is understandable, as AI is not just another feature or function, nor does it fit the mold of traditional IT investments. It is not really an addition to, nor is it a replacement for a CRM, CEM or CMS system, but even AI does have the power to transform. It’s no wonder that so many organizations are reluctant to dip their toes in the AI waters or struggle to know where to start. But the fact is, getting started is relatively easy if you’re willing to rethink the conventional approach to IT.
Alfresco Founder John Newton, has been rethinking approaches to AI and shared his thoughts in a recent discussion.John views AI as being transformative. So rather than simply using AI as a function to look read and extract a document's content, AI can transform a document's content into actionable insights and triggers for further processing. The view at Alfresco is that insight is only as good as the information you've got. One area that John is seeking to explore when it comes to innovation, is tying process and content to AI/ML, for example, using services such as Reckognition™ and Comprehend™ to update the metadata that users rarely, if ever take the time to complete.
The key to figuring out where to start with AI, is to understand that AI allows you to fundamentally reimagine how you work. That doesn’t mean you should immediately embark on a massive digital transformation project or that you should rip and replace your existing systems. In fact, you can, and should, start small. But whether you have grand or modest ambitions, you do have to think differently. Knowing how to think differently, is the key to unlocking AI’s potential.
Look at it this way, whenever you buy, implement or use a business application like a CRM system (for example) you first configure it to meet your business needs. Every business application will be pre-configured, out of the box, to meet an assumed set of industry best practices. That is the way we have long approached automating business processes and tasks with business applications. You will continue to use these same applications in the future (no rip and replace), however with AI they become much more flexible and adaptable to change.
With AI, you can focus on how users or customers genuinely undertake tasks and activities in the real world, rather than working with your best guess process design. Similarly, you can move from a pattern of focusing your efforts on perfecting regular, repeatable activities to, instead, letting the AI focus on the exceptions, understand those exceptions, and learn and evolve by example. AI will work with your existing platforms, frameworks, processes and applications through analyzing content and data through a series of stages.The CTO of Alfresco, John Newton, explained this well in a recent interview:
“First you need to take and enrich content through auto-classification, translation, summarization etc. Second you use machine learning, evolving new information from the information you just brought in. Third, the new information feeds into business processes. Fourth, you collect more information regarding the context of what is happening within the business process. From that staged approach you will be able to anticipate many of the actions that a user might take, and over time the AI system will learn and improve both automating the process and improving speed and accuracy.”
Such an AI enabled system will then also be primed and ready to interact with the task in hand. The AI will understand you or your customers’ intent and act accordingly. Imagine how this could transform something like case management, or sophisticated customer experience activities, to move them from being pre-configured into a learned and intent based system. Typical AI enabled use cases to get you to that point, include things like: content recognition and auto classification, voice & video translation, the detection of personally identifiable information (PII), and document comparison.
We have spent decades analyzing business problems, and then trying to hard code processes and solutions to those problems. Be that our planned responses to customer interactions or in configuring an ERP system. That approach works, up to a point. But stuff happens, and things change even in seemingly simple and static situations. Our customers taste and preferences change, suppliers and partners change. Our initially slick, hard coded solutions start to encounter more and more exceptions until ultimately, they become redundant.
AI works and thinks differently and demands that we do too. AI moves us away from pre-defined functions and features toward an end state where the system reads the situation and responds to the intent. It embraces the state of flux and organically adjusts to change. Or to put it another way, it learns and improves rather than deteriorates over time, and never becomes redundant.
AI is for you, it’s for every organization, though it is not for every situation. It is not rocket science, it is not overly expensive and it doesn’t require you to rip and replace or transform your entire organization overnight. But AI is transformational, and it does require you to think and plan differently, and focus more on the relationships between content, process and good governance. Nevertheless, AI improves and grows in strength over time, and that means that you should start now, start small with maybe something like auto-classification and summarization, and expect, over time, big returns on those early investments.