Today’s workplace is rarely a physical one; video calls, chat messages, shared documents, and digital workflows now define the way we collaborate. Almost all of our work takes place within collaboration tools that generate data, and more data invites the use of machine learning: algorithms that infer patterns in datasets. Machine learning can make your business smarter, but not when its only value is as a marketing gimmick. Unfortunately, there’s a lot of ML for ML’s sake marketed to businesses these days.
But there are uses for ML that aren’t a gimmick. Think about how you use technology to do your job. You send messages, set up meetings, respond to notifications, share projects and documents, locate materials within a vast trove of company data. These are all repetitive tasks that require little thinking, but that nonetheless take time, composing the glue of everyone’s days.
What if you could skip all those repetitive tasks? What if you never had to go looking for the people or information you need to get your job done?
Capabilities that fix those issues might not sound interesting or flashy, but they’re incredibly useful. At Atlassian, we call this “smarts” – the subtle, often unnoticeable features that use ML to make work easier. They might just be a game-changing application of machine learning that changes the way you interact with the core tools you rely upon every day.
Smarts in action
Consider a feature you probably take for granted: Autocomplete. You start typing an email address into the “to” field, and as soon as you strike the first key, you see a suggestion pop up for the exact person you were trying to reach. You press enter to add the email, then three more email addresses show up as suggested options.
While that capability may sound simple, it saves you time on every message you send. Imagine if you had to remember how to spell the email address of every person you contacted, or look them up in a directory every time you reached out to them.
Think about how much time that little feature saves you over the course of a day. Multiply that by the number of days you work in a year. Now consider how that convenience scales across your organization.
What if all the tools you used were just as convenient? That’s smarts.
An assistant for everyone
Software that integrates smarts can fuel productivity. To understand the big picture, it helps to first understand how small tweaks can reshape the experience of an individual employee.
The status quo for collaboration tools is a lot of repetitive, manual work typing exact terms to surface the right materials. Smart collaboration software offers search that responds to natural language, and that can predict what you might look for based on who you work with and what you work on. These next-gen tools automatically suggest the right colleagues to message, or to tag on a document. They can also give you visibility into what your closest colleagues are working on and connect you to experts throughout your organization, so you can better collaborate.
Improving the individual experience in this way is like giving everyone an executive assistant, allowing contributors to focus on the parts of their job that matter most, while leaving routine tasks that don’t require their creativity to the machine.
With so many teams now working remotely, that convenience is crucial. Doing our jobs in isolation wears us down. A survey by employment platform Monster found that 69 percent of respondents reported burnout symptoms while working from home during the pandemic.
How often do you have to ask a colleague to send you a link because you can’t find it in your own email? With teams working remotely, if someone forgets to document where their work lives, whether it’s data or tickets or code, it can make the rest of their team’s job that much more difficult. Smarts can make the task of finding the content you need much less painful, so you can spend less time searching. Intelligent search doesn’t just help you find what you’re looking for – smarts can help surface related content you didn’t even know existed.
Most people’s work involves collaborating with the same small group of colleagues within a larger enterprise – their team. And most teams work on common, repeatable types of projects, like shipping bug fixes or features, or moving customers from marketing funnels to sales pipelines. Smarts can identify patterns in that work and eliminate repetitive tasks, so you don’t have to spend extra time searching for your team members or others who do work adjacent to yours. This ease of connection is even more important when teams are working remotely. In the age of indefinite work from home, you can’t take a stroll through the office looking for the right person. But that doesn’t mean you want to spend your time combing org charts and emailing different people looking for an answer.
You tell software what you care about every time you click on (or ignore) a notification. So why do you keep seeing the notifications you habitually dismiss? Historically, ensuring the notifications in your collaboration tools are relevant to you means enduring the tedium of manually tuning, and repeatedly revising, precise sets of preferences – but it doesn’t have to be this way. Software should extrapolate your preferences based on your workflow and behaviors to determine what needs your attention.
Everything you work on lives in your company’s IT systems in one form or other. You set deadlines for projects, arrange meetings to talk about them, and share materials with colleagues in a common set of software tools. Amidst all that data, software should be able to help you set priorities so you can do the most important thing that others’ work depends upon first. Identify looming dependencies and blockers like this, and employees can meet goals faster and have more time for the important work.
One of the biggest time sucks for enterprises is triaging technical and support tickets. Organizing a single request doesn’t take long, but triaging a deluge of requests with labels, components, and assignees is a mammoth task. And that can be extra frustrating when you realize how much time it takes away from actually serving the customers making those requests, or fixing the bugs leading to those tickets.
When your tools are smart, you can leave triage to the software and focus your time on helping your colleagues or customers. This improves productivity on the technical support side, and time to resolution for customers.
And that’s just one example of applying automation. Thanks to advancements in generative language models like GPT-3, we’re starting to see more opportunities for ML to take on repetitive tasks we thought only humans could do efficiently, like summarizing a status update from a large document, or extracting key actions from meeting notes.
At the company level
Individual convenience adds up to organizational competitiveness, and that’s what makes smarts so powerful for an enterprise. For the solitary employee, smarts feels like having an executive assistant; for an organization, smarts is what it would be like if every employee had a dedicated executive assistant. It’s your company, super-powered.
What if every person on staff saved 10 seconds on 100 tasks every day? Applying small optimization across your whole organization’s work for the entire year means getting a lot more done in a lot less time, accelerating company performance while improving morale by eliminating the least interesting parts of work.
When large swaths of your workforce suddenly have to work from home, smarts becomes even more important. With employees conducting all their interactions virtually, a lot more data becomes available to train software. And when you train software to be smarter, you make employees’ jobs easier, eliminate the mundane, and reduce the risk of burnout.
A massive shift
With smarts, businesses can create efficiency in subtle ways that may not have occurred to them until now. It doesn’t come from implementing one killer feature, but by proliferating many intelligent experiences throughout the tools companies use. That’s exactly what Atlassian is doing with smarts.
ML enables us to build systems that learn user patterns, and use that information to predict what the user will do next. Given the ever-increasing amount of data created by collaboration, and the sophistication of machine learning techniques at our disposal, we can build experiences that recommend and automate simple but essential employee activities.
This has major implications for collaboration within modern enterprises. Technology that helps teams work better together will enable companies to grow at a faster pace, and eliminating small, repetitive tasks means more time to focus on the big picture.
The real power of AI isn’t in widgets that draw attention to themselves. It’s in subtle experiences you barely notice, but that make collaboration smoother in ways you’ve always craved.
Get more stories about teams and tech in your inbox