Tapping Artificial Intelligence for Your Next Best Actions


AI-Powered Chatbots in Customer Service and Engagement

Using AI for customer service in your company is a definite method to save time and money. If you’re like most business owners, you’re constantly searching for fresh, creative ways to improve your enterprise. We’re here to inform you that improving AI customer service is a simple and rapid win.

Read More »

Next Best Actions (NBAs) can be likened to “Insights On-Demand.” Instead of tapping Uber for a ride, you’re tapping your digital assistant for a tip to improve performance. NBAs get to the heart of how AI can help “everyone” to 10x their team’s efforts – by reducing the time cycle for data-driven, custom, real-time actionable insights to improve performance.

It’s impressive even if you have just one person doing this, but imagine what it can be like when your entire organization is tap-tap-tippity-tapping their way to Uber-like growth (and sidestepping some of their painful HR lessons).

How Can You Leverage AI to Make Better Decisions?

As we covered in AI as a Force Multiplier for Business Intelligence, AI provides several benefits to make it easier and faster to make better data-driven decisions. It automates massive amounts of labor in data collection, validation, and analysis. This makes the data available faster, increasing its value, and enabling decisions based on up-to-the-minute data. It synthesizes data into a Single Source of Truth to help simplify the decision-making process.

The benefits go on and on, suffice that it ultimately enables people to work on what matters most – finding and acting upon insights to assist all stakeholders to improve their organization’s operational performance and profitability. This takes the form of Next Best Actions.

What are “Next Best Actions”?

A Next Best Action is computer-generated guidance about how to produce the best outcome relative to specific combinations of conditions (data). NBA’s are in essence the translation of Big Data into Big Wisdom, or as we prefer to call it, Actionable Insights. Modern Analytics, Machine Learning, and AI work together to find and match complex patterns to generate Next Best Actions. As an aside, some may find it useful to refer to the DIKW Hierarchy (aka Pyramid or Chain) which helps explain the relationship between data and wisdom, and how data evolves. Here, we can simplify it as:

The DIKW Hierarchy

How Do NBAs Equate to Big Wisdom?

Our past few posts have talked about data cycle time from different angles – how fast data can be collected, analyzed, delivered, and ultimately acted upon.

The value of data is a central topic to AI systems and as one of our favorite chart’s shows, the value of data is roughly proportional to how fast you can act upon it. Knowing that gasoline was $2.20 a gallon a year ago is less valuable than knowing it’s $3.49 today.

Big Wisdom comes in reducing decision latency or the cycle time for insights to improve developer, team, and operational performance. It usually applies in numerous small, but still significant improvements that can act like compounding interest.

In software development, nearly all performance metrics are interconnected in some way and have at least some impact on overall Cycle Time. Developer skill, team size, composition, structure, work processes, degree of teamwork, and other elements all factor in, too.

It’s not uncommon for a typical team of developers (6-8) to cost $1 million a year. Wages are usually the single greatest expense in software development. A report by Stripe indicates software teams are on average about 32% inefficient.

However, there is (as yet) no upper limit to how “productive” developers can be.

Are Next Best Actions like SOPs?

Kinda-sorta. Standard Operating Procedures (SOPs) provide step-by-step instructions for routine types of tasks. A lot of companies place heavy emphasis on their SOPs.

Someone completely new to a job can use an SOP and workflow diagram to successfully complete sequences of tasks.

The value of an SOP resides in having a uniform way of doing things. This makes it easier to identify how and when defects are introduced into the workflow and to prompt remedial training. Many companies expect everyone to follow their SOPs.

But, SOPs are for routine tasks like how to pick, pack, stack, sort, do basic QA, and ship products in a warehouse environment. SOPs work well for decision points with two possible outcomes, like Yes or No. Complexity increases when adding more possible outcomes or dependencies, like Yes, If…

Many warehouse positions have requirements like, “Being able to pick up and move a 30-pound box” and “Move boxes and set them down in designated staging areas.”

Managing software teams is… more complex.

“IF x, THEN y” works great for an SOP. In software development performance and behavioral analytics, a simple example of the logic used could look something like: “IF a, b, c, and d are true, and if e is false, THEN action.” That implies at least 32 possibilities.

Every additional variable exponentially increases the complexity of deciding the Next Best Action. A great many variables are involved when evaluating NBA’s in software development. Change just one variable and the recommendation will most likely change, too.

In practice, “IF a developer’s productive lines of code, efficiency, test coverage are all good (high), and code churn is also good (generally low), but their defect rate is also high, THEN examine the effectiveness of their tests.”

Programming language adds another dimension. The defects might be associated with having an Objective-C developer using Swift for the first time. Then we get into the team view, where your NBA might be to pair them up with your best Swift developer. And if you don’t have one, then the NBA will be to hire a Swift specialist or focus on Swift training – perhaps depending on the number and severity of defects tagged in JIRA.

What Does a Next Best Action Look Like?

We recently launched our Early Access Program to allow both new and existing account holders to follow our efforts to upgrade our analytics dashboard. We are moving toward implementing an AI-powered Digital Assistant focusing on behavioral analytics for software development. It could change some, but there are at least four things that the AI needs to show when users make a “Google-like” query:
  1. The data – with visualization aids like charts and graphs
  2. An explanation of each data point
  3. A summary of what the data means
  4. Primary and Secondary Next Best Actions
Something like this:

Next Best Actions as Advice

AI serves as an advisor, NBA’s are advice. For now, at least, humans must retain the final authority for making decisions. But Why?

You’re likely to have a lot of real-world awareness that no AI system could even dream about that can influence how you respond to its NBAs. You might know that your startup is still finishing its Series A funding and will have to wait on hiring that Swift developer.

There are things that AI systems don’t (and should probably never) know…

Imagine an NBA that advises, “You must train John that his wife is always correct to avoid arguments leading to long sleepless nights on the couch watching re-runs of the Jerry Spring Show responsible for an 82% drop in his productivity over this past week.”

That’d be creepy. That’s stuff that even as a manager you would probably never know. But, if you’re engaging with your team and have earned their trust, they’re likely to volunteer to you that they’re having relationship issues and that it’s impacting their work.

Alleviate the Skill Crunch

A large swath of the software development paradigm (methods, strategies, tools, processes, etc.) encourages the cultivation of cross-functional teams. With so many companies contending with skill shortages, this is imminently logical for promoting continuous skill development while also automating everything you can.

Startups – Take the case of an early-stage tech startup recently winning its Series A funding round. Congrats! In all likelihood, you’ll need to hire more people. The more skill and experience required, the harder those people are to find and the more expensive they’re prone to be.

NBAs make it easier for your existing team members to step into promotions as team leads, engineering managers, project managers, and more. Even without formal training, they can be introduced to aspects of DevOps, DataOps, and Business Intelligence.

Non-Technical SMEs – More companies are finding that software development can enhance or supplement their core business. They can expand in-house, outsource through freelancers or a development agency, or use IT staffing agencies to augment their small teams. The principles are fundamentally the same, suffice that with NBAs it’s possible for a non-technical project manager to still make sure the project is delivered on time and on budget – and even work as intended.

EnterprisesOrganizational inefficiency tends to increase the larger a company gets. So, the potential value of NBAs increases in proportion to the scale of operations. First, you have more people asking questions about how they can improve organizational, project, team, and individual contributor efficiency and productivity. Secondly, especially if your organization is active in conducting One-on-One meetings and striving to meet Objectives and Key Results, you can align everyone’s efforts on your organization’s most important challenges.

Did you like our content?

Spread the word

Subscribe to Our Newsletter

Don't miss our latest updates.
All About Software Engineering Best Practices, Productivity Measurement, Performance Analytics, Software Team Management and more.

Did you like our content?

Spread the word

Subscribe to Our Newsletter

Don't miss our latest updates. All About Software Engineering Best Practices, Productivity Measurement, Performance Analytics, Software Team Management and more.