In recent months the media have been awash with articles about how artificial intelligence (AI) and machine learning will change the way we work, live and even love. From archaeology to agriculture, medical imaging to cars, it seems that no sector will remain unaffected. So, what does this mean for Field Service and how can it benefit from these new technologies? This blog presents some of the merging trends and their potential impact on Field Service operations.

Scheduling. Whenever I mention AI to service managers, scheduling seems to be the first thing that springs to their minds. The reason? Automated scheduling systems have been available for years now and their benefits are well established. I would argue though that these aren’t true AI systems. They are clever, for sure, but generally algorithmic. They repeatedly follow a set mathematical algorithm to home in on the optimal schedule. AI is more than that. It has the ability to read and interpret data, at times with a fuzzy structure, and make sense of. Beyond that, it can also make decisions and then take action based on these. Add to this machine learning – the ability to learn from experience and automatically find new patterns in data, giving you so much more than the ability to crunch a large set of numbers.

Chat bots and natural language processing will increasingly become part of field service contact centres. We are all now familiar with chat bots and voice driven telephony menus for call centres. Today, Microsoft Chat-Bots can be deployed to automatically retrieve inquiry specific data, say the status of a work order, or confirm whether a spare parts order has shipped. Soon we will see these technologies being combined and enhanced with machine learning so that they can handle issues of increased complexity. For Field Service this will mean that work orders can be logged, and queries answered with limited or no human involvement.

Real-time language translation is another aspect. Contact centres covering multiple geographies often struggle to recruit multi-lingual staff to cover their shifts, often leading to smaller contact centres being retained in-country. As real-time language translation matures, we may well see this leading to consolidation into regional contact centres with all the accompanying cost and scale benefits.

Checks, checks and more checks tend to consume a lot of service admin effort. Service organisations check, review and authorise a huge volume of transactions every day. Purchase invoices are captured, matched and authorised. Parts Requisitions can only be ordered from the supplier once it has been confirmed that the part is indeed appropriate to the fault and the cost within authorisation limits. Service contracts may contain a spend limit per call whereby, depending on the engineer’s cost estimate, the customer needs to authorise the cost before the repair can proceed.

The list can be endless and regulatory requirements such as Sarbanes-Oxley can further add to the effort involved.

Current document management systems can scan documents such as Purchase invoices and log their contents via OCR (Optical Character Recognition), but accuracy is often poor as is the ability to cope with non-standard document layouts. Artificial Intelligence is likely to offer much improved performance. Microsoft’s AI can already be used to extract and log the relevant data items from a scanned document without the need for a fixed layout. This enables its use for any supplier without adjustment.

I have worked with companies where service managers read every service report and re-code the work hours, parts and comments into invoices. In some cases this is done to conform with the schedule of rates for the contract, in others its simply because they know what is (and isn’t) acceptable to their customers and therefore will prevent an invoice query. AI will soon be able to ‘learn’ how to perform this task, releasing service managers from a tedious chore.

AI will also start playing a part in Parts Requisitions and orders. Is the part appropriate for the issue diagnosed by the engineer? Does it fit with the parts used on similar faults? What is the history of the engineer regarding parts requisitioned but not fitted? AI can draw on the data from other work orders to establish a pattern and highlight suspect requisitions. Depending on the industry, the benefits can be considerable; reducing the number of wrong parts ordered reduces parts and visit costs and, of course, leads to a quicker fix.

Object Recognition is another area directly applicable to some Field Service sectors. Here at HSO we have used Microsoft’s object recognition AI to speed up the registration of Customer Assets. We tend to think of customers having a small number of large machines. In industries like Fire Safety the opposite is the case. A site can have hundreds of smoke detectors, emergency lights and sprinklers and their registration as assets is usually time consuming. With AI, the operative points the camera at the asset and Object Recognition gets invoked to identify the asset type, make and model from the picture. Serial or Asset numbers and any expiry date on the asset itself can be captured via another image and the asset is registered almost effortlessly.

As with most new technologies, once AI matures and becomes more widespread, many more applications for it will emerge. Service Managers and their teams ought to start thinking how they can use this technology and realise its potential benefits. The evolution is coming.

Danny Weider, Field Service Consultant, HSO

Danny Wieder

Field service management software