The fully automated hotel and the limitations of capture technology in an AP automation programme

Posted by Pete Loughlin in AP Automation 28 Oct 2015

There’s a hotel in Japan called the Henn-na which is fully automated. Robots at check in, robot cleaners – the lot. Check out this piece from the BBC

In reality it’s not what it seems and for the more unusual guest requests, there’s a human behind the mechanised façade. This hotel may be a showcase of automation but it’s not that ground breaking. There are hotels that I’ve stayed in that have eliminated the check in desk even if they haven’t replaced it with a humanoid robot. In these hotels there are staff about the place to help if you need it but it is a largely automated process to check in and check out. But can the fully automated hotel ever work? Or perhaps a better question to ask is how good do the robots need to be for us to be satisfied that the automated hotel does work? And this question is not unique to hotels. It can be applied to any kind of automation aspiration where we want machines to take the place of humans – AP automation is a case in point.

Site LogoCapture technology that extracts invoice data from paper documents is like the robot at reception. It’s the first point of contact when an invoice arrives at the customer. A single line invoice that matches a purchase order is as complex as a pre-booked single night stay for a regular guest in the hotel. The robot at reception has no problem and is even fine tuned to deal with this frequent visitor.

But what about the next invoice? An unfamiliar multi-line invoice from a supplier that isn’t recognised? Think of this as the family that is booking into a hotel for an indefinite period of time because their house has been flooded. The insurance company is paying, not the guest. There’s five of them. They want adjoining rooms. There’s a baby. It’s 4:00 am.

But it’s not just the once-in-a-blue-moon oddities that will throw the robot, even everyday exceptions will blow a fuse. There is the option to configure the robot more. Fine tune and adapt it. Teach it to learn. But for how long? How much effort is justified? And this is the issue that faces anyone trying to implement an AP automation programme. How much effort do you put in to configuring the capture software to accommodate the numerous idiosyncrasies of suppliers? How sophisticated does the robot at reception need to be to make the customer experience acceptable?

It all depends on how sophisticated the P2P function is. How sophisticated the hotel itself is.

Think of a traditional hotel. It has lots of staff looking after guests’ varied and individual needs. It would be a lovely place to stay but it wouldn’t be cheap. A robot at reception in this hotel just wouldn’t work. It’s like a traditional paper based finance operation. No computers. Lots of human interaction. Probably quite accurate in terms of how things get reconciled but certainly not efficient and probably quite expensive.

But what about a modern, efficient hotel. A coffee machine in the room, mini bar, movies, self checkout via the TV, minimalist rooms that are easy to clean. You can see how a robot at reception might work. Just as scan and capture solutions work very well when there’s a sophisticated P2P function.

Here’s the point. Scan and capture solutions work well – but only if they are deployed in a mature purchase to pay environment. If electronic purchase orders are raised most of the time and if there is a high level of compliance to procurement policy, capture technologies work well. Even better, if invoices are received electronically, the capture solution has less to tackle.

It might seem paradoxical to some but the more the purchase to pay processes can be automated, the less the capture technology has to do, the more effective it is.

So coming back to the fully automated hotel, what is the best approach? Do you make the robot on reception sophisticated enough to deal with all possible customer scenarios? Or do you make the hotel efficient enough so that the robot is faced with a relatively small number of routine situations? For me it’s the latter. Trying to get the robot at the front desk or the capture software to handle old world complexity makes no sense. Bringing the hotel or the P2P function into the the 21st century first and then automating the front desk – that makes sense. And this is where P2P technologies and capture technologies create genuine synergy.

Pete Loughlin can be found on twitter @peteloughlin

Post a comment