WorkDone 101: A Technical Explainer

October 2nd, 2018
We know that WorkDone’s AI-based automation platform is the future of work—but like any technology that’s just now entered the realm of the possible, it can take a minute to fully wrap your head around what an “AI-based automation platform” actually means.
The British sci-fi legend Arthur C. Clarke (pictured above) famously wrote that “any sufficiently advanced technology is indistinguishable from magic”—and using WorkDone really can feel like you have an invisible assistant helping you through the day. But WorkDone is based on a lot of very real, cutting-edge technology.
Arthur C Clarke on the set of 2001: A Space OdysseyArthur C Clarke on the set of 2001: A Space Odyssey

So let’s break it down a little bit. First of all, what can the WorkDone technology actually do?

On the most basic level, it automates repetitive tasks that involve multiple popular SaaS platforms. You install a browser plugin—the WorkDone Monitor—and it tracks what you do over the course of the workday. Once the Monitor detects a pattern that it can repeat on its own, like copying and pasting a certain field from your email client into SalesForce (for instance), it can generate a WorkDone Agent to do that task automatically. We call that whole process, from watching you work to generating an Agent, “Expertise Capture.”

In concrete terms, that means that a little notification pops up in your browser asking if you’d like to let the WorkDone Agent take care of that task for you (just like how a password manager might ask if you’d like it to record a password), and then you can choose to let it take over the copy-pasting work from now on whenever that specific type of email hits your inbox.

So from just watching you work, WorkDone can create a custom digital assistant that takes care of a mundane task (or any number of distinct mundane tasks). In the past, automating tasks like that would have taken specialized programming knowledge—now you can just let WorkDone generate that custom program for you.

Once WorkDone creates a WorkDone Agent, you can share that Agent with everyone else in your company—so that they now have identical digital assistants taking care of that particular task—or share it on the online WorkDone marketplace, where people around the world can pay you to use the Agent you created.
That’s the core of what WorkDone does—it watches you work, then creates applications that can take some of the repetitive tasks off your hands.

Now let’s walk through the basics again, this time with some higher resolution on the nuts and bolts of what’s going on.

For starters, how does the WorkDone Monitor work? You install it as a browser plug-in, and then it uses content analytics technology to track where your mouse moves, what you click on, what’s on your screen, what data is being transferred in the background—in short, it watches everything you do.

All of that data is hashed with 256-bit encryption within the browser, and then the encrypted data is sent to the cloud, where WorkDone performs behavioral pattern analysis on all of the data that the Monitor has collected. Only you (the customer) retains the private key for decryption purposes (unless you’ve specifically agreed to give WorkDone a copy), so we can’t see any of that user or process data.

Once our behavioral pattern analysis (which uses machine learning technology to process all of your data) spots a work pattern that can be automated, it asks you if you’d like to start running a program in your browser to automatically complete the task that’s been identified—that program is the WorkDone Agent.

Since our technology is browser- and cloud-based, our target market is cloud-forward customers who use popular SaaS platforms. For now, WorkDone can’t identify tasks involving on-premise software, since that falls outside the scope of our WorkDone Monitor’s ability to track what’s going on as you work.

If you want some more specific examples of how WorkDone gets work done, you can also read our use cases on purchase orders and invoice processing.

And if you’re curious, we’ve also compiled a quick glossary of some of the technical terms that keep coming up:
  • Machine learning
  • Cloud
  • Content Analytics
  • Behavioral Process Analysis
  • WorkDone Agent
  • WorkDone Monitor
  • Expertise Capture