Griffin on Tech - We need to learn from Ports of Auckland's automation misadventure
The mantra of the last few years in the tech world has been that we need to embrace automation to boost our productivity and shift humans to higher-value tasks.
But the reality is that according to consultancy firm EY, 30 - 50% of robotic process automation projects initially fail. Tech vendors and their customers talk up the success stories, but we hear less about the projects that went off the rails.
Ports of Auckland this week said it was abandoning its terminal automation project and would write off $65 million worth of automation software and guidance systems as a result. It's a painful admission for a company that runs a critical piece of national infrastructure in our largest city.
What's the real cost?
The automation project had been years behind schedule and with the labour shortages and supply chain disruption caused by covid, it was having a knock-on effect on the whole economy that probably puts the actual cost of its failure in the hundreds of millions of dollars.
As Andrea Fox writes in the Herald: "The cost of business lost as ships waited up to two weeks to unload, or skipped Auckland entirely; the costs to the economy of resulting supply congestion; the extra costs to Auckland importers and exporters of congestion charges imposed by frustrated shipping lines; the cost of lost export business and opportunities because empty containers have been stranded in Auckland."
Ports of Auckland's cranes are integral to the country's supply chain
Some of the crane equipment will now need to be converted for manual use, defeating the purpose of the project, which was to speed up container processing by taking the pressure off port workers and getting them out of harm's way. Ports are dangerous places, as a recent spate of worker deaths reminds us.
So what went wrong at the port? Auckland Council, the port's owner, has launched an investigation into the failure. It could prove to be an instructional case study of the things to avoid when embarking on a major automation process.
As I wrote on Tech Blog yesterday, RPA is well-embedded in businesses now and a new wave of automation tools, such as IBM's Watson Orchestrate service, is adding artificial intelligence to make automation more intuitive and flexible for a wide range of tasks.
Industry 4.0 too expensive, too hard
But the really expensive automation projects marry intelligent software with hardware in factories, transport infrastructure and on farms. New Zealand firms have been relatively slow to embrace the "Industry 4.0" movement, scared off by the high costs of implementation and a lack of expertise to design, install and run automation systems.
The Ports of Auckland debacle won't do anything to inspire confidence in company leaders who are being urged to automate. That's why the investigation should serve as an opportunity not to pin blame on various people, but to fully understand the processes, technology and leadership failures that led to such a costly mess.
The business case for automation remains just as strong. Ports are automating parts of their processes around the world. But the high failure rate of projects is a reminder that the complexity of automation can't be overlooked. Ports of Auckland should and can serve as a lesson for anyone heading down the automation path.
EY's top ten reasons why RPA projects fail:
1. Not considering RPA as business-led, as opposed to IT-led.
2. Not having an RPA business case and postponing planning until after proof-of-concepts (POCs) or pilots.
3. Underestimating what happens after processes have been automated.
4. Treating Robotics as a series of automations vs. an end-to-end change program.
5. Targeting RPA at the wrong processes.
6. Applying traditional delivery methodologies
7. Automating too much of a process or not optimizing for RPA.
8. Forgetting about IT infrastructure.
9. Assuming RPA is all that's needed to achieve a great ROI.
10. Assuming skills needed to create a PoC are good enough for production automations.
You must be logged in in order to post comments. Log In
The main reason these things fail is because almost no one involved in the process actually really knows what they're doing. There is no fitness for purpose on either the supply or demand side, and the only ones who win are ethically suspect suppliers who're simply better at specifying contracts than their hapless, marketing-led customers who they're happy to exploit. Of course, all the secrecy surrounding these failures also ensures that the guilty parties are unpunished in the marketplace by the reputational damage they deserve. All the parties to these sorts of debacles should be named and given the opportunity to explain themselves publicly. I suspect that would immediately change the way things are done (or not done, more to the point).
Is there an NDA that prevents the parties involved being named? And if so, should there ever be one when public money is involved?