One of the most common complaints about IT projects is that they take so long to deploy. Here’s how the Department of Immigration and Border Protection managed to fully automate its Border Risk Identification System (BRIS) in just 12 months.
Servers picture from Shutterstock
These observations come from Gavin McCairns, chief risk officer for the Department of Immigration and Border Protection, and came up during the GovInnovate 2014 conference in Canberra last month. (We quoted McCairns on the importance of fat data during the conference.)
The aim of BRIS was to automate the process of assessing passengers for immigration clearance. Prior to the system being used, only 10 per cent of passengers were sampled to identify potential risk. Post BRIS, that figure has risen to 100 per cent, using a constantly updated set of analytics to identify potential risks. Automation was deemed essential to deal with rising traveller numbers (15 million passengers to assess in 2014 is predicted to rise to 18 million in 2015.)
The key requirement, according to McCairns, is to ensure that you don’t waste too much time on the development phase. The prototype for the updated system was built in six weeks, and then piloted for six months.
By the end of 12 months, a production system was ready to roll out in airports and help identify risks. It continually updates risk profiles to identify passengers of interest.
“The computer can’t spot the sweat on someone’s brow — only a good officer can do that — but the computer can help to tell us who should be sweating,” McCairns said.
There’s also a recognition that automation may not always work. “We’ve built a self-licking ice cream, which I really like,” Mc Cairns said. “If the reality of what’s happening at airports or seaports doesn’t match the predictions, we start to shut it down.”
Three other useful rules? McCairns put it succinctly: Don’t break the law; don’t break the bank; don’t break the mainframe.
That said, it may be easier to take a “skunkworks” approach at time and develop something unofficially, at least in part. As McCairns puts it: “It is easier to seek forgiveness than to seek permission.”
Evolve is a weekly column at Lifehacker looking at trends and technologies IT workers need to know about to stay employed and improve their careers.
Comments
5 responses to “How Border Protection Rolled Out A Major IT Project In 12 Months”
Glad they are focusing on efficiencies in areas like this… although the “It is easier to seek forgiveness than to seek permission.” quote is always scary when it comes from a government employee, lol
Yeah, I hate that quip too. Its better to know that you can explore innovation without being blamed for failure, but the organisation will learn from failure to build better next time.
Yes, some project mgmt tips there but this is a major identity and border surveillance system. And this dude says ““It is easier to seek forgiveness than to seek permission.””. That should worry anyone.
This reminds me of that scene from Dirty Rotten Scoundrels:
“No Ruprect, what do we do afterwards? That’s right, we apologise!”
Still tells us little about the actual project itself, apart from keeping the development phase short.