In the first part of this blog series I discussed how modern data processing systems and human experts can together help you to optimise your mobile processes. In the second part I then looked at recent developments in artificial intelligence that mean computers increasingly develop more human-like understanding of complex situations. In the final part of this mini series I now want to look at what the challenges are in harnessing the power of the new AI technologies and how we overcome them.
So what are the challenges on the way towards a future where AI optimises your mobile operations? Most advanced AI technologies are black boxes, you feed in some data and you get back a result, but it is very difficult to verify on what basis that result was derived. In part 1 of this blog series I mentioned how important it is to establish confidence in results. If I tell you a device is likely to fail within the next month, but I cannot explain to you how I came to that conclusion, or give you any evidence why that might be the case, you are likely not to put much trust in my statement.
Establishing trust becomes an ever bigger challenge the further we move from simple observable metrics to abstract predictions, and this problem gets hyper-charged with AI systems. Not even the person who builds a deep neural network model will be able to fully explain how that model comes to a conclusion, just as a parent might not be able to fully understand their child’s behaviour. In addition, if we want to utilise all the power of the AI to come up with new creative solutions, we need to give it the freedom to think outside the box.
As a standard our mobile analytics product Elemez collects many basic operational metrics of devices, such as battery health or network connectivity. Based on those metrics you might want your batteries to last long and your networks to have a high signal strength. But these are only indicators of your business running smoothly, you do not actually want to optimise your processes around maximising battery performance. Instead what you might be interested in is the number and duration of support calls to your help desk or the number of parcels handled by a worker in a warehouse.
With the introduction of AI, our role at B2M changes. In the past we were the experts designing a solution, we were making the call that battery health is important for the smooth running of a mobile deployment. With advanced AI solutions we are becoming more like a teacher, enabling the system itself to figure out the most important predictors for optimal performance. A teacher chooses the right inputs to create a conducive learning environment, a teacher gives its student the right tools to solve a problem, and he might also have to keep oversight of the students such that they cause no harm to themselves or others in the process.
That does not mean our expertise in large mission-critical mobile device deployments becomes irrelevant. But it does mean that we rely even more on the trust of our customers and partners. Beyond the device metrics our client application collects from mobile devices, Elemez is set up to integrate additional data about business processes and support tickets, both on the devices as well as from third-party back-end systems.
To harness the full potential of AI we ask you to trust us with the most valuable thing you have – your data. Only when we integrate the data on device health collected by default by the Elemez device client with data on the business process you want to optimize and the key performance indicators you measure success with, can we train AI models to find the best solution for you.
And only if we build those solutions together in trusted partnerships, we will be able to verify their value and build the confidence in the results that is necessary to make an impact.
That is why we seek to build close partnerships with our customers. We are here to answer your questions, to work through your individual requirements, and to earn your trust every step of the way. Because we know, only as trusted partners we will be in a position to enable you to get more done and deliver on the both the strategic and tactical business goals set for your enterprise mobility deployment.To learn more about our bespoke, customized solutions that leverage the power of Elemez and our professional services and AI / data science teams, please contact us.