“A self-learning, truth-eating, digital parasite.” Mission Impossible 7: Dead Reckoning – AI Is The Enemy
The AI Hollywood hype machine continues to steamroll, where many lacking critical thinking skills fall for its fallacies. The US government has also gotten into the AI regulation and promotion game, imposing an AI Bill of Rights. Those with a short memory are wise to remember what happened during COVID-19 when the government imposed restrictions, leading to the economic problems that we have today. Take a look at your LinkedIn feed; many companies and individuals have taken to ‘AI-Washing’ overhyping the capabilities of AI, similar to its distant cousin, “Greenwashing,” where the chair of the SEC, Gary Gensler, issued stern warnings against this practice, according to a WSJ report last December. Yet, the AI armchair quarterbacking continues, where unsuspecting victims sometimes fall prey to these AI charlatans who peddle false hopes and outrageous promises.
Is It Really AI?
According to a report by Bloomberg, AI mentions during earnings calls have dropped precipitously. This trend is a good sign as the C-suite is now starting to ask questions on how AI can help their business rather than blindly throwing money at it.
There’s no unified definition of what AI is. Most will say it’s getting a computer to think and act like a human, an umbrella term encompassing many things – natural language processing, computer vision, or robotics. Truthfully, it’s machine learning algorithms – instructions you give the computer doing the work, not AI.
Machine learning (ML) analyzes past patterns from data to predict what will happen next. Predictive analytics is the crux of what machine learning does, mistakenly interchanged with the sexier moniker of AI. Now that the horse has left the barn, many use the term AI instead of correctly using ML by those in the know and those who aren’t.
Beyond Vanity To Generating ROI
Getting C-suite executives to talk about AI is easier now than ever. However, once you get your foot in the door, it’s time to distance yourself from the AI hoopla and discuss tangible ways to help your target company improve operational efficiencies that lead to top and bottom-line growth, gaining a solid ROI.
ChatGPT and the whole generative AI have been a blessing. It’s opening up a path for those who’ve been in the analytics business to educate company leaders who see AI as something transformative that helps their business grow as the US continues to suffer from inflation and labor shortages during a time when the employment rate is at an ultra-low 3.7%.
However, deploying AI, or rather an ML consulting engagement, isn’t as easy as typing in a prompt in ChatGPT. It takes hard work, discipline, and a keen eye for astute business principles that have been around forever.
Here are the crucial strategies clients and companies selling AI consulting and software must agree on for a successful engagement.
- Identify what business problems you are solving: For many companies, whether in a retail or warehousing environment, it’s driving repeat business, a major sore spot. Other issues can include logistics optimization, demand forecasting, or predictive maintenance. Companies constantly face internal and external threats. Most businesses that have existed for over five years have treasure troves of data they can draw from, laying the foundation for any ML project.
- Have a plan for sourcing and cleaning data: Data provides the fuel to make things happen. Garbage in, garbage out. Cleanliness of data matters for accurate predictions – a small error in input data can lead to the butterfly effect, providing inaccurate predictions that can lead to disastrous consequences for your business. Data cleaning is the less glamorous part of a consulting engagement but accounts for 80% of the process.
- Come up with a math model that enables you to solve your pain points: Now that the data is clean, it’s time for math to take over and insert the clean data into the predictive analytics engine. C-suite executives need to ask their AI consultant or software company the how or the why behind their predictions. Time series analysis may be appropriate for those looking to improve inventory forecasting. Pattern-matching algorithms may be the right solution to help a company determine what a customer will purchase, or clustering algorithms may help determine what groups of customers will do based on geographical, demographic, psychographic, or behavioral characteristics.
- Take action with your findings: All theory and no action is a waste of time for everyone. We’ve all come across consultants who provide a good game plan in fancy PowerPoint presentations or have the data sit in pretty dashboards created by Tableau and then leave once the check has cleared, leaving your company to their devices to figure out the next steps. Worse is the software company, which gives a login and password for an ultra-low monthly subscription, leaving you (as the consultant did) to figure things out on your own. Both are mistakes. Once the AI-ML has made the predictions, it’s time to prioritize and execute the game plan to solve the originally stated business problems.
The Loudest Isn’t Always The Smartest
There’s a lot of noise on the interwebs, particularly on LinkedIn. Just like in politics, the loudest screamers get the most attention. Resist the urge to engage with the newly-minted, self-proclaimed AI experts and research this technology.
In the end, C-suite leaders want results. Investing in a true AI-ML consulting or software isn’t for the faint of heart. It requires a visionary willing to step out of their comfort zone to realize the immense potential predictions can have for your company’s profitability.