Big risks, bad data, venn calendar, predictive analytics and unstructured data
Pearls of wisdom from Episodes 66-70 of the AI & Intelligent Automation Network podcast
Taking big risks is necessary, eliminating bad data is imperative, enabling venn calendars ensures a zen team, predictive analytics are only as good as the thinking behind the input and iterative change could be death by 1,000 cuts
Taking big risks is necessary
Episode 66: Cathy O’Dowd, Mountaineer
You don't take the big risks that lead to the big successes if you are paralyzed by the fear of failure.
Cathy O’Dowd is the only woman to summit Everest twice. She doesn’t speak in metaphors. When she speaks of life and death circumstances, she’s discussing one of her friends and colleagues either living or dying based on how they approached the job at hand- summiting a mountain. So when she speaks of big risks and big successes- she means, that if she didn’t put her life on the line because of the fear of it ending, she wouldn’t have succeeded.
And so, when you are making decisions on the risks to take at your global corporate enterprise regarding the use of artificial intelligence, realize the stakes are high, but they could be higher.
Eliminating bad data is imperative
Episode 67: Ricardo Badillo, Western Union
Determine the cleanliness of your data. If you automate trash, it’ll be trash in trash out. If you automate bad data, you're going to get bad data.
If taking big bold risks is now on your list of things to do thanks to Cathy O’Dowd, consider that data cleansing and data mining are big bold risks in waiting. Meaning, if you don’t truly understand the data and how you’re informing your artificial intelligence, your output will be flawed or just wrong. Don’t let the attempt of outpacing of change be the reason you lose the race. Put yourself in the story of the turtle and the hare and make the right choice of which protagonist part you’d like to play.
Enabling venn calendars ensures a zen team
Episode 68: Chris Gilmore, Aetna
We recognized that we needed to create a collaborative environment. When you think about all the different players; we have developers, testers, project managers, business experts, and IT experts. If you don't know how to work together- either all in the room together or virtually all at once- you're often having to do follow ups. So, the first thing that we did was we elongated that common periods of work. We instituted a minimum of two hours and sometimes as many as four each day. That meant in some cases the offshore team was working earlier and in some cases, our business team came in earlier. But the result was a common period with people talking to each other, which made a huge difference.
The technology is just a tool. Prior to this quote, Chris was discussing where he was finding his talent. He discussed new hires, involving integration partners and other solutions. And so he wound up with myriad people in myriad placed. All Chris is saying here is that people have to work together to be able to work together. With current technology, we don’t have to be in the same room- we can have virtual interaction. But the interaction must occur to ensure the best possible results.
Predictive analytics are only as good as the thinking behind the input
Episode 69: Martin Ruane, Engie
It's using data and analyzing that data on a large scale to be able to better predict for outcomes.
The promise of predictive analytics is great. All we’ve got to do is apply some machine learning to our historical data and voila- more, better, profits. Yes- you have to have found and selected the right technology and partners for your enterprise to succeed. But when you’ve done that you then dive into your historical data. And it’s the result of that plunge that makes the true difference. In asking the question, do you have what you need- consider the question of, do you know what you need? Where you’ve been is not where you’re going. What future are you attempting to predict? What will your enterprise deliver to your future customers? Preparing for an unknown future is difficult. But plotting out a future today based purely on the past won’t help you tomorrow.
Iterative change could be death by 1,000 cuts
Episode 70: Michael Baldauf, Raiffeisen
We know that we want to chip away at that the tremendous amount of unstructured data. We have to increase it enormously otherwise we can't use AI for much more than the next step. It's impossible to do that. Getting a grip on data is basically our number one priority.
Michael is sharing that wholesale change through artificial intelligence cannot happen unless you understand your unstructured data. And he plainly states that getting a grip on data is the number one priority. The good news is that global corporate enterprise is still collectively in a place where spending time on your data before you implement scaled artificial intelligence solutions is where you should be. But if you don’t do what you should be doing now- or worse, if you spend all of your time on simply iterative change, you’ll be left behind.