Skip to main content

3 places to start when implementing automation

(Image credit: Image source: Shutterstock/Vasin Lee)

Automation is a business buzzword that's become unavoidable. Some may relish the idea of technology so advanced it can take on humans' work for them, while others fear their jobs will disappear and they will be replaced by robots. 

Research suggests that the former prospect is more accurate than the latter. A recent report from the McKinsey Global Institute found that modern technology can already automate 51 percent of existing job activities. 

Take a moment to analyze that argument. First, the report doesn't say more than half of today's jobs could be automated; it specifies activities. In other words, automation isn't taking away humans' jobs — it's shifting what tasks humans will focus on at work. 

Another aspect worth noting is that automation isn't currently taking on half of existing jobs; the technology exists, so it can. That said, there's a great amount of potential for companies to begin implementing automation into their operations. 

Automated tasks in action 

Even with this room for growth, automating tasks is not a new idea. Companies are already letting technology execute certain tasks on behalf of their employees. 

Amazon, for example, has 45,000 robots working alongside 230,000 people, according to a report by the Seattle Times. At the center of the company’s massive distribution network with famously fast and consistent delivery is Kiva Systems, a robotics company the e-retailer purchased for $775 million in 2012. 

However, even Amazon has been hit with cyberattacks recently as devices and the malware infecting them become more intelligent with time. Automation is becoming a necessary key in cyberdefense as organized, well-coordinated, automated attacks bombard networks. 

What makes cybersecurity companies like McAfee so successful is being able to collect and analyze vast archives of malware, viruses, exploits, hacks, and other attacks. Stepping into cybersecurity without this infrastructure is impossible, as global threat intelligence evolves daily. Automation is the only way to process enough information fast enough to be effective in a zero-day environment where attacks are always coming. 

Once threats are identified, protections need to be implemented across the network to remain effective. Whatever the malware is doing needs to be stopped, and it’s a lengthy process hindered by the fact that it’s continuing to move ahead. Automation speeds up this process to allow real-time containment of the attack while humans analyze the attack’s progression. Of course, this is only possible when network monitoring is automated to detect anomalies. 

Detecting, intercepting, and eliminating threats isn’t the only way automation is working with humans. With Google Translate’s automation technology, human-to-human communication is more accessible than ever before. Introduced in 2006, Google Translate has more than 500 million monthly users, making it one of the company’s most-used assets. 

With 7,099 living languages listed by the Ethnologue catalog of world languages, Google has a long way to go in translating them all into each other, and even the 100 or so it currently supports have many regional nuances and idioms that make direct translation difficult without the use of artificial intelligence and automation. 

Google’s developer teams have been hard at work rolling out AI for Translate that is as intricate as any voice assistant like Siri, Alexa, and Facebook’s M. Recognizing contextual cues and responding in a conversational manner are the goals of any AI developer. Advances made in this research become universally applicable when made under the Google umbrella, and there’s no telling in what directions the automation used by Translate will lead the company. 

That said, automation to reduce costs or risks or to increase productivity are not the only tasks within the realms of automation; tasks resulting in an increase in revenue or improvement in the customer experience are also becoming strong candidates for automation. 

Implementing with the big picture in mind 

In order for automation to affect the business positively, business leaders must approach implementation not in small, disconnected endeavors, but as an addition to the business that will benefit it holistically. 

For companies looking to automate tasks in a way that harmonizes with their business, consider three places to start: 

1. Non-critical processes: New endeavors are rarely a widespread success instantly, and automation is no different. Companies should start small and work their way up when they start automating. It's best to begin automating just one to three tasks — and it's important that these tasks are not critical ones that could negatively affect the business's outcome. If errors occur and a human cannot easily backtrack and correct those errors, look for a different task to automate. 

This way, companies can test automation's success in an environment that won't be detrimental to overall operations. After all, implementing automation will mean very little to a company if it's not implemented correctly. 

2. Repetitive processes: Scripts and processes that require low cognition and in which data sensitivity and privacy aren’t a concern are ideal for automation. Repetitive processes are ripe for the choosing. When companies automate these mundane, repetitive tasks, workers have more time to focus on tasks that require human intellect instead of busy work. 

Automating these tasks not only increases production, but it reduces errors by creating a uniform process across the company. The more repetition and less cognition needed to complete a task, the earlier it should be automated. 

3. Descriptive — not prescriptive — processes: Descriptive analysis summarizes raw data for human understanding to allow full analysis of the past. Sums, averages, and percent changes are examples of descriptive statistics, and they’re found through consistent reporting. Looking at descriptive data is like asking an employee to give you a report of a project's outcome after the project is complete. 

Prescriptive tasks, on the other hand, involve more forecasting than reporting, like asking a team member to recommend a course of action on a project and predict its outcome based on reasoning. That said, predictive analysis is inherently more challenging than descriptive — even for humans, it's easier to describe than it is to predict. For holistically sound automation, it's better to begin automating descriptive tasks first. 

Every job — from the most simplistic of entry-level jobs to executive seats — requires a mix of both mundane and complicated tasks. Thanks to automation and artificial intelligence, how we accomplish those tasks is soon going to change. It started with tech giants like Google and Amazon but is quickly finding its way into companies large and small across every sector. 

It’s only a matter of time before automation becomes more prolific, and business owners need a bouquet of services using a crawl-walk-run approach to successfully implement it. Rather than rush to get a seat at the table, it’s necessary to analyze the business benefits using a risk-backward approach to determine which automation projects to run. Examining work processes to determine which require human intervention and which can be automated is the first step toward weaving automation into your company holistically. 

K.R. Sanjiv, CTO, Wipro

Image Credit: Vasin Lee / Shutterstock

K.R. Sanjiv is the chief technology officer for Wipro, a global information technology, consulting, and outsourcing company. Sanjiv has more than 25 years of enterprise IT experience spanning multiple industries.