Innovations must be grounded in real world process
There is a lot of discussion and interest in Artificial Intelligence these days. The promise of AI is one that uses data and automation to eliminate redundant and mundane tasks, allowing people to refocus their attention on activities that provide more value to the organization. But the trick is using AI in ways that make a real difference.
How can we make the most of AI without getting bogged down in the technical miasma of the approach? That’s what I asked David Jones, VP of Product Marketing at Nuxeo. Dave is a member of the board of directors at AIIM International and an expert in digital transformation systems and strategies. In this continuing series, we discussed innovations in intelligent information management and the opportunities that may have yet been realized. Click here for our prior discussion on “the Art of the Possible” with AI. Meanwhile, let’s continue your discussion and exploration of using AI to do real things.
Dave, you warn us that the fascination with AI can quickly fade unless the approach is used to ‘do real things.’ What do you mean by that?
I share the excitement surrounding Artificial Intelligence, but often the discussions are theoretical and based upon capabilities that are not always grounded in the day-to-day operations that organizations must manage and improve every day. AI can do fantastic and predictive things, but it can be hard to get your head around those possibilities. As a result, the potential promise of AI is wasted.
So what do you suggest as a good first step for organizations looking to move forward with AI in more practical ways?
My argument that there’s something in the middle between big transformative possibilities and small task oriented process improvements. These efforts are perhaps slightly more mundane but are important and beneficial use cases of AI. And it all starts with content discovery at the core.
Research tells us that 80% of the content that is commonly held in information repositories is redundant, obsolete or trivial. Let’s use AI to go find out which 80% is redundant, obsolete or trivial and either get rid of it altogether or apply low touch retention policies and practices to better manage that information. Instead of having people sit there and go through documents and figure out what should be kept as a record, let’s use AI to do it.
But assessing the status of your enterprise content can be more easily said than done. What is the approach that you and your team at Nuxeo suggest to make that that happen?
We’ve created an AI framework that performs in layers. The logic is that there are companies out there like Amazon, Google or Facebook who have built advanced AI engines with extended capabilities, but what those engines are difficult to use and you have to be highly skilled to use them. And even at that, they’re generic.
What we’ve built is a wrapper around those engines to make it easier to use, easier to train, very easy to apply business or domain specific datasets to those AI models. The result is a more pointed and practical approach that you can take advantage of in real world scenarios.
Can you explain what you mean?
One example is a scenario where you show an AI engine a picture of a truck. The system recognizes that the image is a truck; it’s got four wheels, it’s blue, and it’s a Ford that is parked in front by the shore of a lake. The AI will do a reasonable job of categorizing that and classifying it. This is great, but it’s really not that useful. If you’re Ford you what you want to know more Ford-centric specifics.
Exactly what model of truck is it? What is the exact type of alloy wheels are on that truck? What is the specific paint code of that blue?
This is the type of information needed for truly domain and business specific intelligence and automation.
So the framework allows you to plug into that business knowledge and train the AI model to come back with those specifics?
Yes. And then the result is that the data collected and the insights provided are more specific to your particular business and business process.
Read our other blog on this topic here:
Kevin Craine is a business writer, technology analyst and award-winning podcast producer. He was named the #1 Enterprise Content Management Influencer to follow on Twitter and has listeners and readers worldwide. Find him at CraineGroup.com
Follow Kevin Craine on Twitter