Bots: How to Enhance User Productivity Through Intelligence

According to Rebecca Freeman of the OECD Statistics Directorate, labor productivity boils down to total output (produce) divided by total labor input; or, in other words, each employee hour worked should be directly assignable to the total product or services rendered.

If that’s true, then increasing an employee’s throughput should directly increase the overall organization’s productivity. The problem? You can only get so much from a single person, no matter how many different methods you use to optimize employee output.

From that point forward, the only way to achieve higher efficiencies is through automation. But how can “people” be “automated”? Answer: Bots!

To be clear, this is not about replacing or reducing people. Rather, this is about reducing the mundane or tedious parts of a real person’s daily workload, so they can spend more time focused on providing the most value to their organization.

What is a Bot Anyway?

Wikipedia says that a bot:

“An Internet Bot, also known as web robot, WWW robot or simply bot, is a software application that runs automated tasks (scripts) over the Internet. Typically, bots perform tasks that are both simple and structurally repetitive, at a much higher rate than would be possible for a human alone.”

On the other hand, Microsoft’s definition is a bit more specialized:

“A bot is an app that users interact with in a conversational way using text, graphics (cards), or speech. It may be a simple question and answer dialog, or a sophisticated bot that allows people to interact with services in an intelligent manner using pattern matching, state tracking and artificial intelligence techniques well-integrated with existing business services.”

Microsoft’s introduction of the Bot Framework in 2016 was neither exciting nor interesting. Skype was the only feasible access channel, so it didn’t make sense to increase complexity without adding significant value. Since then, however, the Framework has been improved considerably and the list of available bot channels continues to grow (now there are 14). Also, users are generally more tech-savvy and have begun demanding other—and arguably, better—ways of accessing those same backend systems.

The Bot Opportunity:  What Tasks Can Be Automated?

Today, there is an incredible opportunity to harness the power of bots in general. They can be accessed via practical, realistic and popular ways, and can be implemented faster than ever before. What’s more, we stand on the precipice of a productivity revolution fueled by technology—one that will allow practical organizations to enhance, extend and increase customer engagement that, in most cases, won’t require end users to lift a finger.

The evolutionary leap itself is easy, and—using tooling already in place—the road to providing real productivity enhancements is surprisingly short. The hard part is figuring out which repetitive, manual tasks are ideal for automation. With this in mind, Bots are best suited for singular tasks or orchestrated, discrete sequential tasks that make up a complete unit of work. A few examples of available bots are those that:

• Send informal, dictated messages to small groups of people
• Tell you what your daily schedule consists of
• Tell you when you should leave for your next appointment
• Order lunch on your behalf
• Read your important emails aloud while you’re driving

Still don’t think bots would help your employees? What if employees could use bots to:

• Accurately track their time (billable or not) by speaking in short, natural sentences instead of filling out a timesheet?
• Distill a list of qualified LinkedIn candidates by asking one single question instead of hours of scouring profiles?
• Schedule a meeting by saying, “Setup a meeting this week with {person name}” instead of playing tag to sort it out?
• Schedule an emergency service visit by simply saying, “The refrigerator is leaking” instead of researching top repair services?

Once you target the task(s) you want to automate, implementing a bot for it is relatively quick and straightforward. And, if you use the Microsoft Bot Framework, it only takes a small amount of configuration to allow any—or all—of the channels to leverage that bot once implemented.

Intrigued? Excellent! Here’s a simple roadmap for implementing your first bot, from start to finish:

1. Identify Repetitive Tasks

Even though this may sound daunting, it really isn’t. Most of your employees are likely trained to perform regular tasks around the same time each day. So simply asking “What do you do when you first settle in each workday? What do you do before you leave the office?” will probably do the trick. If not, then “What do you find yourself doing at least twice a week?” or, “What do you wish you could do without having to do it yourself?” should work just as well.

2. Determine Which Tasks Will Benefit From Automation

Bots can’t work miracles, but they can accomplish wonderful things when they work with existing APIs; and with APIs, little to nothing is required from the backend to work with the bot. For organizations with systems that lack APIs, a good candidate task is one that will be API-enabled soon. But, even if neither is applicable, a bot can still automatically analyze or generate formatted data that users would have otherwise had to prepare manually.

3. Rank That Filtered List By Impact

The purpose of this step is to highlight which tasks are most beneficial. Decide which metrics determine the inherent value of a given task, such as input time saved, response time shortened, and potential maximum throughput. These metrics are not just limited to productivity either—they also have an inferred expense savings or revenue generated aspect. Regardless of what metrics are adopted, use them to figure out how much “bang for your buck” you can expect with each automated task.

4. Rank The Filtered List (From Step No. 2) By Estimated Implementation Time Required

The benefit of this step is that it illustrates which tasks can be implemented quickly and it introduces a new way of getting things done. For example: Imagine that a team identifies the “most impactful” bot, but realizes it would take seven people an entire month to implement; on the other hand, let’s say the team finds an easy, win-win task that would require a bot that could be implemented by one person in one day. Which one should they choose? The latter, because it helps employees acclimate to bots faster. It’s a healthy dose of practice that makes the transition to the “more impactful” bot much easier for those slower on the uptake.

5. Figure Out Where Steps No. 3 And No. 4 Meet—Then Implement “That” Bot

Once you’ve figured out the task, your ideal metrics and the amount of implementation time required, here’s how to implement the bot:

• You can use the Microsoft Bot Framework, which requires that custom bots hosted in Azure are implemented using C# or JavaScript.
• You can also use self-hosted bots, which can be implemented using any technology as long as they are accessible via HTTP/HTTPS and can understand JSON.
• Azure-hosted bots can be “serverless,” or they can be full-on applications. At present, the serverless model makes the most sense. In most scenarios, it should cost the least and should be the easiest to maintain. Exceptions to this include:
• New bot developers should expect to spend anywhere from one day to about a week getting intimately familiar with the Framework. After that, the effort is just like any other .NET Framework-based effort.

Increased productivity is only one of many benefits that bots can offer. Want to learn more? Contact us to learn what bots can do for your organization.