A Tribe Called A.I.

Sandy Carter tweeted an interesting article today about the different types of artificial intelligence and the communities that surround them.  It put them in groups or “tribes” which have similar abilities and are in some form of war with each other for supremacy, dominance, and an A.I. version of Darwinian evolution.

Whether you are a A.I. Evolutionist (developers of Genetic Algorithms that evolve themselves) or a Kernel Conservative (developers who match a particular A.I. template or approach to a problem and then develop the machine learning system to solve it) it does not come as a surprise that you want to find a more efficient solution to a problem.  Why else would you decide to use your chosen version of A.I. over another.

I think these “13 Tribes of Intelligence” will co-exist as we find their different strengths and weaknesses and apply them in the most efficient ways possible.  Knowing about all of these approaches, their strengths and weaknesses, and applying them based on quantifiable criteria instead of “which tribe you belong to” will save time, money, and maybe even lives.  (If the system is matching Kidney donors with receivers like the system in the U.K.’s NHS then lives are literally at stake.)

Self Organizing Teams at the Executive Level

Self organizing teams are one of the most productive forms of collaboration.  Individuals in these teams are more independent, communicate better, find answers to questions instead of letting them go, and continually enhance their own skills and abilities.  There are five core pieces to highly functional self organizing teams: Competency, Collaboration, Motivation, Trust and Respect, Continuity.  I will address how each piece of this pie fits the highly desirable C-Suite of a successful business and argue the case that at the Executive Level, the only successful teams are self organizing.

Competency: I have read that if you know more about the job you are hiring for than the prospect employee that you shouldn’t hire that person.  This feels correct following commonly held beliefs such as “If you are the smartest person in the room, you’re in the wrong room.” and “Anything someone can do 80% as well as you can do it should be delegated.”  What all this comes down to is the ability to get the job done.  Now, we want every member of our team, organization, company to be able to get the job done and typically those that can’t are not working with us for much longer, however, this is particularly important at the Executive level.  Specifically, you cannot know whether someone is a good salesperson or not when you have made your career in the software development field.  Is one customer good if they are big enough?  How good is 100 customers a month?  These questions are unanswerable unless you know how sales works.  In a self organizing team, each member is competent in their role and responsibilities so you don’t have to worry that your CFO is really doing a poor job but you just can’t tell.

Collaboration: Communication is essential to any organization and especially to the highest levels of business.  We spend days upon days and endless hours thinking of ways to better communicate with our customers, business partners, and colleagues.  In self organizing teams, teamwork is encouraged to help push the business forward in whatever way team members need to work together and because each member is competent, they can focus on what they do best and rely on their team to support them in all other areas of the business.  Focusing on each member’s strengths to push their contribution to the next level is highly valuable and sometimes essential for a business to succeed.  All of this comes back to the core of collaboration, which is to work with each other to achieve results that are greater than the sum of the individual effort.

Motivation: One of the biggest motivation killers is to have a manager breathing down your back before a critical deadline.  Sure, you’re there working hard at crunch time to solve the immediate problem but we all know what happens once that deadline has come and gone; you lose interest.  At the executive level, this can be seen as quarterly reports, reporting to investors or shareholders in general, and reporting internally to the CEO.  Without motivation, deadlines are missed or pushed back due to lack of wanting to work on a particular daunting project.  To encourage and motivate a team, coaching and guidance should be the order of the day and with members of the C-Suite this means sharing information, explaining projects and technologies, collaborating on exciting projects, and applying the coaching and guidance towards each other.

Trust and Respect:  Trust your team.  As a CEO you absolutely must trust your team.  Give them the reigns and let them handle more details of that aspect of your business, for that is what they are good at.  This gives them the ability to work the best way they can and frees yourself up to focus on what matters for your position and role.  Without Trust, and therefore Respect, the team will lose motivation, the desire to collaborate, and eventually competency because they just won’t care about the work anymore.  This makes Respect one of the most important conditions for a successful business.

Continuity:  The team sticks together.  There is a reason a CEO, CFO, or any other C-Suite Executive stays with a company for many years.  To get anything accomplished at a large corporation it will take years, especially if it is a change in culture, a pivot to adopt new technology, or even the launch of a successful and sustainable business.  If the CSO leaves 2 years into a business launch, picking up those relationships with key customers will not be easy and could send the company into an unrecoverable tailspin.  Form a team and keep the team for at least five years.

If you don’t believe now that a Executive Team is automatically a self-organizing team then take these principles and put them into practice (if you can).  You will find that things will change; A+ players will stand out and anyone who cannot acclimate or adjust to new standards of excellence will be ousted for what they really are.  Finding your Team’s Zen, where working with your team is second nature, will come with benefits:

  • They pull work for themselves and don’t wait for their leader to assign work. This ensures a greater sense of ownership and commitment.
  • They manage their work (allocation, reallocation, estimation, reestimation, delivery, and rework) as a group. They still require mentoring and coaching, but they don’t require “command and control.”
  • They communicate more with each other, and their commitments are more often to project teams than to the an individual.
  • They understand requirements and aren’t afraid to ask questions to get their doubts clarified.
  • They continuously enhance their own skills and recommend innovative ideas and improvements.

Now if all of this seems enticing, at the executive level and otherwise, there is a small formula that can help build these teams.  Nitin Mittal has an excellent write-up on Building Self Organizing Teams.

Visual Intelligence – Stephen Few

At Auximiti we are always moving forward and striving to provide the best information at all times.  Sometimes, this leads to releasing weird, awkward, or unhelpful features.  Our recent analytics iterations in our platform’s admin console were live for a couple days for each iteration.  Starting with getting data into a graph, we began to understand what we were working with.

First Version of Data Report – Auximiti

Initial Release: You can see the dates and the number of uses are highly visible but the colors don’t help.  You also don’t know which day of the week matches each bar.

These were obvious flaws that we could jump on and correct, which we did in our next iteration.

Second Version of Data Report – Auximiti

Iteration 1: Here we see a lot more information provided in an easy to understand presentation. Labels on the graph gave context.  We can see which date corresponds to which day of the week.

We wanted to dive deeper to get an analysis of the hour to hour data.  Little did I know that Stephen Few and a few of his 8 Core Principles of Data Visualization would guide our development and address our problem.

“The best software for data analysis is the software you forget you’re using. It’s such a natural extension of your thinking process that you can use it without thinking about the mechanics.”

– Stephen Few –

Iteration 2: We’re looking at his principles of Explore and View Diversely which address finding new ways to look at the same data.  Solving problems and proving new hypothesis posed by the user.  We wanted to view our data at a more granular level, hourly in our case, to see if new insights could be derived.  Unfortunately we had really run into the problem of creating a ridiculous chart.  Just look at all the colors!

Third Version of Data Report – Auximiti

From here we had an obvious question to answer:  What are we going to do about our color scheme?  Now at this phase, our fearless leader David had been channeling Stephen Few which he had read and was familiar with his ideas on data visualization.  He pointed me in the right direction with a simple statement.  “Don’t use color unless the color has purpose.”  What an amazingly simple statement to wrap up our problem.  In relation to the Core Principles, we we’re still on track with Stephen Few.

Iteration 3: We’re going after the principles of Compare and Attend which are defined as ways to show the user what is important and why, without getting in the way of their cognitive processes.  Our cognitive block? These distracting colors and no correlations. A busy hour switching from 1 PM Monday to 4 PM Wednesday would not change color.  This was a flaw which there was no choice but to address.  We wanted people to get information at a glance.  We settled on a purposeful use for color that showcased the highest volume of traffic.

Fourth Version of Data Report – Auximiti

Now, you can immediately understand the information.  When you have busy days/hours in red and easy to view comparisons to the rest of the data.

Looking forward we are planning to expand on the Compare principle by matching multiple locations and devices up with each other in a single chart, showing where business differs and where it is similar.  We will probably have to iterate again on colors and labels until we get the display to deliver actionable information at-a-glance.  This time we will be starting from a much better perspective.

The full list of Stephen Few’s Core Principles of Data Visualization

Update:  Our comparison charts look fantastic and let our users login for straight-forward information.  This is something we could port to wearable technology or augmented reality!