By Sunetra Chakravarti, Monday 05 March 2018
Machine learning, artificial intelligence and robotics have dominated every technology conference, tradeshow and summit in the past six months.
Sophia, the humanoid robot, has been a star of many of these events and appeared on ITV’s Good Morning Britain last year repeating pre-programmed sentences with very unhuman-like jerks and twitches of its wire-festooned ‘head’. This is the media-friendly side of a technological revolution that is also taking over many of the jobs previously done by humans.
So, is automation linked to machine learning and artificial intelligence set to revolutionise the printing industry?
It’s a difficult question to answer, but one thing is clear: there are areas in which some of this technology is still very unsophisticated. In spite of its huge success, Amazon is regulary the butt of jokes that focus on its inability to understand how to serve up product recommendations. A customer may have just ordered, say, a roller-blind, which invariably prompts the website to flag up every other roller-blind available as a potential follow-up purchase, as if the customer must be some sort of obsessive roller-blind collector.
However, shoppers in the not-too-distant future may have a different experience with a smart AI personal shopper, which is able to better analyse what that customer is looking for and is able to create very targeted digital catalogues delivered to the customers email or indeed printed catalogues sent to their home address.
Similarly, on a very basic level, social media platforms may soon be able to generate personalised photobooks for users through an ability to analyse the relationships and frequency of interaction of the people in photos posted to the site.
On the shop floor, however, manufacturers are more interested in better predictive maintenance and enhanced security as well as in greater understanding of device utilisation and performance. All that, AI can deliver.
In a survey by Boston Consulting Group and MIT Sloan Management Review of more than 3,000 business executives, managers, and analysts in 112 countries and 21 industries, 72% of respondents in the technology, media and telecommunications industries expected AI to have a significant impact on product offerings in the next five years.
Dragana Pavlovic, senior vice-president, Global Development Group at Xerox, thinks AI will completely change the way business is done.
“Artificial intelligence can improve every part of the printing process. For example, in Xerox FreeFlow software today, there are algorithms that can figure out different document layouts, such as optimising imposition in order to minimise printed waste. In direct mail and catalogues, as more information about the recipients becomes available, it is possible to use that data to create even more relevant mailings by automatically customising the job content for the recipient.
“In job submission, Xerox FreeFlow can monitor which presses are busy and route new jobs to available presses. Presses, like the iGen 5, constantly monitor themselves with sensors and use this information to make real-time in-process adjustments for things like paper alignment and image quality to provide best possible print without human intervention.
“Today’s machines also connect back to Xerox where machine data can be analysed, compared to expected performance and allow for adjustments or software updates remotely. In the future, data could be used to predict the need for service before the machine fails.”
These technologies can improve efficiency and effectiveness across the whole printing process, from print job creation all the way to continuous production and machine service optimisations.
For example, sometimes the new technology helps to automate tasks which might have been bottlenecks in the process before, which then enables the whole business to work more efficiently end to end and to be able to take on more jobs. Other times, the technology can offload some of the more routine work in order to enable the workforce to focus on higher value tasks. AI and machine learning have that kind of potential for positive overall impact in the printing business.
However, they work best when handling a well-defined process rather than broad systems, as Richard Charnley, technology consultant at NXP Europe, points out: “If you take chaos and try to automate it, you end up with more chaos.”
He adds: “People on a budget don’t realise what you can do with automation. It is the way the industry is moving. More and more people are clued up to the fact that manual processes are not cost- or time-effective any more.
“The problem within the industry is that although people are interested in it, they can’t see what can be done on a budget. Any business process with a logical step can be automated. And to do that, all it takes is rolling your sleeves up.
“Enfocus Switch, for example, can automate almost every business process by doing the heavy lifting – in fact 90% of all print tasks can be automated this way. You can work a lot more efficiently and quickly – can get Switch to program the printing presses and it will run long-run jobs freeing up operators.
“I would advocate Enfocus because you get to the proofing stage in three to four minutes rather than have a studio person on it running it manually. The software does the work and gives the user full control. They approve jobs quicker and you can pass it on to route workflows.”
Time to bed in
However, no new technology arrives completely free of drawbacks and these aren’t necessarily apparent in the early days. As with all revolutions, it takes time to accurately assess the wider impact.
Nick Murray, managing director of Wellington Press belives there are limits to what automated systems will be able to handle. He says: “In the past few years, bigger companies have started moving to automation. Jobs, like sending a one-off calendar to a customer, will ultimately be completely automated.
“However, the quick turnaround work required to print 200 financial brochures overnight to reach the London Stock Exchange tomorrow morning at 7am will never be automated. Jobs like that need people – the files will be sent through at night, then there will be emails exchanged and client liaison and such like.
“For me as a business, investment into full automation is not completely justified at the moment. However, there is a real market out there.”
As James Belton, operational planning and solutions manager at Harrier told a group of printers recently: “You try to automate as much as possible, but you still need a human to watch it. Most of us have a diverse product range so we can’t automate everything; there are too many variables. It is better instead to spend money on efficiency throughout the whole business.”
This is a UK-based customer retention technology business focusing on personalising the ‘customer journey’. The company was challeneged by one of its customers, Amara Living, to come up with a way of delivering personalised dynamic content at the point of the first order, in order to encourage customers to make repeat orders.
Retain.me used XMPie PersonalEffect TransMedia Pro to create and dispatch a customised delivery note with tailored marketing content. This included highly targeted marketing messages based on customer status, algorithmically generated category recommendations, natively translated content based on the language of the website visited (English, French or German), and QR codes for encouraging the customer to give a product review. Amara Living saw a second-purchase improvement rate of 13.1%, and a 23.3% improvement in the existing customer repeat rate.
This was a win for the printing industry too because it required the creation of new full-colour digitally printed pages and therefore new revenue that did not previously exist.
The company’s main focus was to simplify and expedite the online purchasing experience by empowering customers to “self-serve” answers to commonly asked questions.
By developing an artificial intelligence plug-in, AnswerDash, Moo made it easy for customers to self-serve answers to repetitive questions that had previously stretched its live chat channels.
When used alongside LiveChat, AnswerDash serves up common Q&As before escalating a chat with Moo’s human support agents, freeing them up for more personalised, high-value conversations.
During the six-week A/B test of AnswerDash, sales conversions on Moo’s US site increased by 1.4% and the average order value increased by 1.8%. Beyond increased sales conversions, Moo’s customer support team experienced an 18% reduction in live chat volume during the same time frame.