5 Questions to ask before starting your IoT/AI Project
Dominic Duffy, Head of Product Marketing, Mitie
As the UK’s largest facilities management company, Mitie knows a thing or two about buildings and building infrastructure. We know the often significant, sometimes profound, difference an improvement to maintenance regimes, energy consumption, space utilisation, and even cleaning schedules can make to a business. We know because we make them every day, on behalf of our customers. And it’s our pioneering Connected Workspace that provides the insight needed to make all the difference. But it wasn’t always the case.
Back in 2016, Mitie knew instinctively it was in the ideal position to leverage its domain expertise to help customers improve the performance of their buildings. We knew that by combining data from different business functions, the opportunities to improve processes and systems would be revealed. We also realised that collecting the data meant embracing the Internet of Things (IoT), and that meant getting to grips with an emerging technology, quickly.
Two years on, we’ve learned a huge amount. So much, in fact, we’ve turned our knowledge and experience to good use, with consultancy in IoT and data science now forming a significant and rapidly growing part of our business. We know a well-executed IoT project can yield significant benefits. Unfortunately, it’s a sad reality that most IoT projects are not well-executed
Having implemented Connected Workspace across many sectors and business functions, we’re used to asking the questions that provide the basis for successful projects. Below are five of the key questions we encourage you to ask before embarking upon any IoT project.
1. Does your project actually matter?
Let’s assume your project is the first of its kind. As such, it’s likely to form the pilot for subsequent projects. So, it makes sense for it to address a real business need and to have a clear business case. Only then will it have the visibility and importance to secure commitment across the organisation and gain crucial support from your InfoSecurity, Operations and IT colleagues.
2. Have you identified the right data?
Having aligned your project with a business goal, you’ll need to determine which data you’ll be collecting, and from which sources. This sounds a simple task, but it is absolutely pivotal to the success of the project. As such, it requires a considerable level of lateral thinking and domain expertise.
Maxwell Wessel, of Stanford Business School, cites the reduction of waste as an example of ‘Getting to the Right Data for the Job,’ by asking three specific questions: ‘What decisions drive waste in your business?’ Which decisions could you automate to reduce waste?’ and ‘What data would you need to do so?” By answering these three questions, he says, you’ll be in a good position to answer perhaps the most important question relating to the ‘right’ data: “If you could have any piece of information, however unbelievable, to make the perfect decision, what would it be?”
3. Do you have access to the right skills?
We’ve looked at the importance of identifying the right data, but what if you don’t have access to the skills needed to help you do it? Even if you have identified the right data, do you have the skills needed to interpret it in any meaningful way? The chances are, you don’t.
You’re not alone. According to Gartner, the single most significant challenge for businesses looking to adopt IoT/artificial intelligence (AI) is a lack of necessary skills – particularly in data science.
Without input from a data scientist, it’s quite possible you’ll either collect too much of the ‘wrong’ data, or be unable to derive any valuable insight, even from the ‘right’ data.
4. Where might your project lead?
Thus far, we’ve established the importance of aligning your project with a specific business goal, collecting the right data and ensuring you have access to the skills to gain relevant insight. So far, so good. Now, let’s assume your project was a great success; a lot of valuable insight was gained. What next? You’ll probably want to apply the same approach in a new project, aligned with a different business goal
Which all makes sense. But it would most likely be an opportunity missed if any subsequent project was executed in isolation of the first. After all, business functions are interconnected, so it follows that IoT projects should also be interconnected.
For example, at Red Bull Racing, an IoT/AI project was undertaken to identify the cause of complaints of tiredness and lethargy from the engineering team. Suspecting an environmental cause, a number of sensors were installed to collect data relating to temperature, volatile organic compounds (VOC’s), humidity, sound, movement and carbon dioxide. But rather than confine the monitoring purely to environmental parameters, the decision was made to monitor the power consumption of air conditioning units in the same area.
The result? Not only the isolation of the cause of the problem – very high levels of carbon dioxide – , but also identification of over £17,000 of potential savings in energy costs.
So, a project that began by focusing on environmental monitoring, expanded to encompass energy consumption, with very little additional work required.
5. Are you prepared to wait?
Although collecting the right data is key, you’ll still need sufficient volume from which to gain insight. This takes time. How much time is difficult, often impossible, to predict accurately. Don’t fall into the trap of assuming the more data collected, the better. Unchecked, data volumes can grow exponentially, which paradoxically, can actually slow the process of deriving insight.
It is therefore vital that all those involved are aware of the need to be patient. The good news is, having answered the questions above, your project is likely worth the wait!