14 Apr 2021
Head of data analytics at BAI Communications in Canada, Jeremy Foran, believes that although automation and machine learning have been the cornerstone of business optimisation, many organisations still underestimate its importance. Read on to understand the business imperative of driving better business operations through automation.
It was 2012 in Australia and the stakes were high: the Australian government was looking to clear 126 MHz of wireless spectrum in the 700 MHz range and a band of mobile network operators (MNOs) were each willing to pay millions for their share. However, before the government could provide access, the analog broadcast channels occupying the spectrum would have to be turned off and the digital television channels moved to a lower band. Clearing the 700 MHz band meant over 1,400 existing digital transmitters would have to be re-tuned.
BAI Communications (BAI) was appointed to coordinate the program across more than 20 organizations, scoping and commissioning the work to be carried out at 426 sites spanning the continent. Some estimates put the scale of the project at 15 years’ worth of work. BAI completed it in three. How was BAI able to complete the project in that timeframe and under budget?
The team knew from the outset that – given the scale and timeframes to which it had committed – the work had to be done differently. Automation would be critical. Getting the processes, data, and relationships right was essential, but that took valuable time from the program and meant that some work had to be done manually. So, the importance of the automated system – developed in-house – was even more essential.
The team’s reliance on automation paid dividends. By programmatically generating statements of work, testing procedures, and approval flows, the team moved from delivering a handful of sites a month to 60 sites a month at the peak of the program.
Without this innovative automation throughout the program, each of the design, review, and commissioning teams would have needed to be 3-5 times larger or the project would have taken up to five times longer, which would have been unacceptable for the government or MNOs Instead, BAI was actually rewarded by the government for its timely execution.
Automation contributes to differentiation
A 2020 global survey of business leaders from a wide cross-section of industries conducted by McKinsey & Co. found that 66% were piloting solutions to automate at least one business process, up from 57% two years earlier.
Some of today’s most innovative companies distinguish themselves through their ability to automate:
- Uber disrupted the taxi service industry by automating the dispatching service.
- Tesla has become the most valuable car manufacturer in the world by leading the world in highly automated manufacturing methods.
- Amazon is the 800-pound gorilla with highly automated fulfillment centers that enable billions of packages deliveries a year. The fulfillment centers are a marvel, and if you can take a tour, I highly recommend it.
Automation is so powerful because it shifts company resources from tasks to decision-making. Similarly, accountants don’t spend their time focusing on arithmetic. Accountants use Excel so they can focus on interpreting the numbers and making budgetary decisions.
To start an automation department, you need at least a single person willing to learn how to write computer code. Writing code is how you automate everything outside of Excel (as well as inside Excel). There is a plethora of free resources online for anyone looking to get started. A language like Python can be picked up over the course of a few weekends. Python has a diverse ecosystem that provides powerful libraries so beginners can hit the ground running. It won’t be long before your different systems, such as CRM, DMP, NMS, POS, etc., are all talking to each other.
The evolution of automation: machine learning
Machine learning seems to be present in every facet of business discussion today. This is because it represents the next evolution of automation as it provides a means of automating decisions. In 2015 I was part of the BAI team that deployed the public Wi-Fi service in the Toronto subway system. The network is designed to support 1.6 million riders per day. A network of that scale requires a high level of complexity, which necessitates a robust monitoring system.
Network monitoring systems collect and present KPIs from all the underlining equipment to the network operations teams. The team’s job is to monitor and interpret the data and decide if the network is behaving normally or abnormally. Interpretation is required because a network switch running at 1% utilization is normal at 3:00 am when the subway service is not running, but very much indicative of a problem during a Monday rush-hour.
Using anomaly detection, a form of machine learning, BAI ‘trained’ a machine learning model on our ‘normal‘ Wi-Fi usage patterns. The model would then alert the team when it observed patterns that were abnormal. This reduced the number of team members we needed watching screens all day, trying to detect problems. It even notified us of issues we may have missed. It operates 24×7, zero sick days, and costs almost nothing to run.
To bring automation into the business you don’t need to be in the automation business
Most companies are not accounting firms, yet they have accounting departments. In a similar way, the intent of an automation department is not to be in the automation business, but rather to bring automation into the business. An automation department should play a supporting role in serving the business objectives and providing a framework to grow, adapt to market changes, scale-up production, and reduce overhead: the smaller the company, the more gain for growth.
In my experience at BAI , I have both witnessed and participated in a history of adopting automation practices. Today, we leverage all opportunities to automate processes to be sure that we are delivering optimized solutions for our customers.
While Warren Buffet has said that “accounting is the language of business”, I believe now need to consider that automation is the cadence of business. The forward-thinking organisations that have already incorporated automation in their daily operations will continue paving a path for improved innovation, efficiency, and growth.