Organizations today continue to challenge themselves, seeking breakthrough performance by leveraging automation and cognitive transformation strategies. And while so many leading organizations have discovered and enjoyed the benefits of Artificial Intelligence, the discipline required to achieve the promise of digital euphoria requires a focus that has alluded many organizations almost since the dawn of the microprocessor.
In my experience, the two main obstacles we face today in achieving true valuable change that drives significant growth with first AI projects are still somewhat basic in nature.
The first challenge, is that many firms still lack the necessary skills or expertise needed to start and run a project effectively. This trait remains true for most firms that are trying AI for the first time, because AI is still on the bleeding edge of innovation for most industries. The AI talent base will increase and improve over time, but many organizations will still have to learn to manage the risks of experimenting to realize the positive effects of such cutting-edge technologies.
The second common ailment that hampers success of most AI projects, is related more to the management approach that companies typically use to govern them. Even today, after many failed attempts, there is an ever-present ‘technology first’ solution strategy that is still used to try and solve many business problems. I like to think of this as putting the “how we fix it” before having a true understanding of the “What is the goal?” related to the company’s needs and challenges. For a team tasked with innovation, getting to the heart of the matter of the business problem, requires both methods that will allow you to get to extreme clarity, and the executive agreement to move forward and spend BEFORE choosing a technology (i.e. platform, software, hardware).
I’ve outlined what I believe are the most critical elements to should consider as you venture into your first Artificial Intelligence or Augmented Intelligence project.
Do we build it? Or do we buy it? That is the question.
As your enterprise works through the throws of its first AI project, beware of the undying belief that you can “build it yourself”. It’s not that In-house projects are a bad idea or methodology….as many firms have done an amazing job with other technologies and their own personnel. But on the cutting edge within the likes of AI, it is seldom effective. If you do decide to take on a project like AI in-house, consider the true risks and challenges that remain with that approach.
Its attractive on the surface to start such an endeavor internally with a project like AI. Understandably, most people believe that using your own data scientists, for example, is an unquestionably a better option. And let’s face it - they are already employed by your organization and they likely are already very familiar with your data structures. And like most scientist types, they are naturally very smart and capable people. But they are also very likely to already tasked with accomplishing other important and latent projects like big data analytics, process automation, and machine learning projects. And since new AI projects can consume much more of their time, be sure they have the bandwidth, skills, and the tools to do it.
Hidden costs – extensive and expensive
Many Internal projects also often fail due to extensive hidden costs that crop up when you least expect it. Sometimes those expenses are so daunting that the project can be disbanded altogether. One of the biggest reasons why new projects fail was because there was inaccurate scope of the true requirements. Often when project scope gets expanded out of proportion, it’s because there wasn’t a firm set of project objectives set forth at the opening charter. Operating without solid success criteria, internal projects often continue to expand, and fail to demonstrate positive return on investment, or provide true business value. Unless you truly have the in-house experience and available skill-sets with some proven AI outcomes, it’s often easier and more valuable to hire the experts.
Expect progress and momentum, but not perfection
Everyone seems to strive for perfection, but it’s a rare occasion to attain it. Within business, as it is in life itself, true perfection is just an illusion. Because of this, trying to be perfect with a business outcome is not only pointless, it is detrimental - especially if it holds back progress or results that show necessary payback and ROI. If you aren’t getting the return in the timeline you expected, then shift your focus to incremental improvements to get you closer to achieving profitable financial outcomes even if at first they seem smaller than you would expect.
Remember that perfectionism often tends toward procrastination; so having an all or nothing strategy might work occasionally but its likely much more important to the team funding the project to see bottom line results early on. Try to always focus on improvements aligned to Key Performance Indicators (KPIs) that truly matter. Which brings me to the next point…
Keep your eyes on the “KPI’s”
Delivering results against the right metrics for the right people at the right time is critical to getting the next round of investment after your first AI project. Powerful measured objectives that lead to better outcomes, will empower your team to focus and act. To make better decisions, grow the company, and collaborate – requires the right data on what’s working (and what isn’t.)
I often tell my clients the KPI’s that matter the most, are likely the ones that are headline conversations between someone in finance (i.e. the CFO) and the key areas of current change – the executive suite of focus. For example, a good relationship between the CFO and the CMO, might give a realistic goal and timeline for marketing spend. The high level of collaboration will provide much better measurable KPI’s that without that cooperation.
Good KPI’s also have actual deadlines, in what I’ve dubbed “Quant-by-date” KPI’s. Something like “growing sales by 10% before Fiscal Year End” or “creating 15% reduction is costs to manufacture” is hard to argue with or misinterpret.
If you are looking for areas to focus on getting started in your first AI project, and feel you need ideas for better KPI’s than you already have, then check out this earlier blog we published that covers five strong entry points for a first project. Each of these key areas has its own measure for success. And easy example might be “5% increase in product adoption by end of this version release”. Pretty easy to track and likely translates a value that is easy to understand. For any cutting edge or new investment AI, seek the Highest Return, Lowest Complexity project you can get your hands on.
So, this begs the question “What isn’t a strong KPI?” I would suggest that if a measurement of an activity or a metric does not directly influence your successful achievement of a business goal, then it is not a KPI, it is simply a metric. Not all measurements are equally important, therefore, not all metrics are KPIs. Know the difference.
Think Big, Start Small, Move Fast
One important concept that I would always encourage when it comes to creating your first AI project is to “Think big, start small, and move fast”. Like it says right on the cover of the book title written by the Mayo Clinic Center for Innovation, “thinking big,” really means that successful innovators have to consider the full range of possible futures. They make sure they understand the emerging technology context, rather than just assume that their current assumptions are right. If you want to truly accomplish something great; be willing to start with a clean sheet of paper and pursue the most rewarding opportunities. Shoot to enable solutions that might just rewrite the rules of a category or even entire industries.
But since every journey has to start with a first step, “starting small is about taking that big idea and putting actions against it that are manageable. One of my favorite quotes of all time is the one by Henry Ford, which said: "Nothing is particularly hard if you divide it up into small jobs". Spend the time working through key initiatives and then sort them based on which one might gain the most fanfare WHILE creating enough financial return that you have a satisfactory outcome.
Once you have a big idea as a starting point, then seek ways to “move fast” and create incredible velocity. One way to get people rallied, is by giving your project its own brand name and be bold with it. Spend the effort; perhaps even write an Ethos for the project.
Another way to create keep the project moving fast is to outline the risks beforehand. A little bit of risk assessment and fear planning can go a long way later on when you hit speed bumps. Create a plan that will account for everything from the highest leadership decisions to the most practical concerns, including (but not limited to) the location of key documents, re-routing procedures for critical activities, and even identifying short-term replacements for key positions during the project. Murphy’s law has taught us all that if you don’t plan for someone to leave, or get sick, or (heaven forbid) get hit by a bus, is when the event is most likely to happen. Business people are very familiar with putting together lists that demonstrate what will happen during a project – but few outline what to do if something undesirable event decides to rear its ugly head.
Get Some Confidence
Hiring a confident team is probably the single biggest contributor of success for any first project, especially in AI. Smart people are important yes. Teamwork is also ultra-important too. But the “get it done” attitude is probably the single biggest contributor of success for any project.
Self-confidence is the fundamental basis from which leadership grows. Trying to teach leadership without first building confidence is like building a house on a foundation of sand. Likewise, the team for an AI project has to work with such conviction that they will make change happen without fear and are willing to stare naysayers in the face without any doubt.
I think back to the film Zero Dark Thirty where the red-haired character Maya (who was played by Jessica Chastain) was not only persistent by writing on her boss’s window every day up to the 121st day that they decided to take action. But in the movie, the Navy Seals who were putting their lives on the line needed to see the conviction to get rallied around the dangerous mission. To realize her dedication. There is a scene in the final stage of the film when before departing, the character played by Chris Pratt asked “do you really believe this story? I mean….Osama Bin Laden …”. Then the character returns with the question “So what convinced you??”. And the squad leader played by Joel Edgerton turns and points his hands to Maya and answers, “Her confidence”. And there Maya continues her look of undying belief and faith in the objective, when seconds later the president gives her the ok to launch the mission. Look to capture a team with the same confidence and belief that Maya has in getting the job done.
At Cognitive Scale, we are experts at operating Augmented Intelligence projects that lead to truly better outcomes. We’ve implemented numerous enterprise projects that deliver effective insights and change within 10 weeks, and true return on investment in under 10 months.
This project guideline you’ve read here is based largely on a culmination of our prior successes, and I hope you find some value in it. If you would like to hear more about the Cognitive Scales approach to success, please request a time to learn about our 10-10-10 methodology to learn about how we approach first AI projects. We’ve created a Getting Started Guide that helps to outline the steps necessary to be successful, and some ideas around Case Studies that work for various industries.
If you would like to discuss your project, please request a complimentary consultation by reaching out to us and meeting with our experts.
Chuck McMurray is a client executive with Cognitive Scale. Chuck brings 20 years of field experience selling and consulting in Big data, advanced analytics software, and consulting services.