11 of the Most Common Artificial Intelligence Roadblocks & How Marketers Can Avoid Them He failed in business in ’31. He was defeated for state legislator in ’32. He tried another business in ’33. It failed. His fiancee died in ’35. He had a nervous breakdown in ’36. In ’43 he ran for congress and was defeated. He tried again in ’48 and was defeated again. He tried running for the Senate in ’55. He lost. The next year he ran for Vice President and lost. In ’59 he ran for the Senate again and was defeated. In 1860, the man who signed his name A. Lincoln, was elected the 16th President of the United States. The difference between history’s boldest accomplishments and its most staggering failures is often, simply, the diligent will to persevere. One of my all-time favorite clients is one of those guys whose entire office is plastered with motivational quotes from world leaders, presidents, and football coaches. (Vince Lombardi was especially loquacious.) When I started working with John many moons ago, he had a couple of Successories posters his kids had given him about Excellence and Perseverance taped to his walls. These days, he’s got a football-field-sized office packed floor-to-ceiling with chi-chi-la-la original art housed in bougie frames, but the overall message is the same. You can do or be anything you want, no matter your circumstances if you just get to work. (Say what you will, but he started selling stuff out of his trunk with $0 in his bank account, and now he still has zeroes. There are just nine of them BEFORE the decimal point.) Long-term client relationships ebb and flow just like a marriage. Some of them are good. Others are not-so-good. The ones that last over time where you develop your own language? They’re golden. Several years ago, John and I hit a rough patch. We weren’t seeing eye-to-eye on, well, anything. Frankly, he should have fired me. Our most recent stuff was killing it for them, but he’s “too old and has way too much money” to deal with drama, and I was causing a lot of it. You see, John gives me free rein to create 5-10 new projects each year, and they execute them no matter what. There aren’t any limitations, even when it comes to budget. We just need to make sure that 1 out of 5 is a home run. (Yes, I know, this is a dream situation for a consultant. I’m very thankful.) Every project I had put up for the new year was AI-related, and every one of their managers was opposed. Diabolically opposed, in fact. John has a big team, and because his company has many methods of distribution, there are many managers. (More managers than Vince Lombardi quotes, if you can believe it.) Not one person thought they should tackle Marketing Artificial Intelligence. Not even the Customer Service Ambassador who reminds me of Mrs. Claus. Her vocabulary doesn’t include the word “no, “ yet she was a “No. Hard No. Not in this lifetime.” Not only did all the managers agree that they should not do it, they were very vociferous. Believe it or not, I understood it. This company has been around for a long time, and they have several legacy systems. Systems so big and cumbersome that I’m secretly convinced they’re powered by dinosaurs living in one of the restricted buildings. (When I say dinosaurs, I mean real T-Rexes, not employees with 40+ years of tenure, although they have many of those too.) The company was also late to the web game. (In the interest of full disclosure, this is why John gives me so much rope. He regretted not listening to me when I was squawking about the internet back in the days before Al Gore invented it.) They were so tardy and had so many prehistoric creatures, er, systems, that it was painful. Very, very, very, very, very, very, VERY painful. And even more expensive. Funny enough, they have more robotic technology than almost any company I’ve ever visited. Their new buildings are so smart that they make the geniuses they’re named after seem remedial. Their pick-pack operations are something out of a sci-fi movie. And yet, they had absolutely no interest in adding AI to their marketing efforts in a meaningful way. Sure, they had all the personalized recommendations this and all the dynamic upsells that, but the real needle-mover stuff that they were going to need to build in-house? They’d have rather had group lobotomies. They didn’t want to deal with AI, Marketing, or Marketing AI. Full stop. In this position, 10 out of 10 consultants would have quickly returned to the drawing board and come up with a bunch of new projects. Anything else was suicide. I doubled down. You see, this client had a monumental impact on who I am as both a consultant and a person. As villainy as I sound here, I believed I had a fiduciary responsibility to stand my ground. Even if it meant ending the relationship. I didn’t feel this way because I thought I knew everything. The older I get, the more I realize I know jack. This client has aggressive competitors, and because of how data is being used/sold in their industry, there was only room for one company to take the advantage. In other words, if they didn’t act, they would lose their position, and the numbers supported that they wouldn’t get it back. I pled my case to the managers as a group—a complete fail. I tried doing it individually—also a complete failure. I begged. I bribed. I contemplated resorting to blackmail. Nothing worked. So, I went to John. “One project. Scrap the rest of the list. It’ll be a homerun or….” “You’re willing to bet the house on this?” He interrupted. (If we’ve ever worked together, you know “bet the house” is a whole thing with me.) “Yours,” I said. No smile. Not even a little. “Okay. Just one. You should get X involved because he’ll take a bullet for you. Keep away from things involving A, Q, R, and S because they’ll be purchasing all the ammo west of the Mississippi. And kid, don’t come back to me with anything that’s not a win.” Long story long, this old-school, legacy company is now doing more Marketing AI than most big guys. The Pteranodons and Allosauruses are smashing it. I’m not saying it was easy. It wasn’t. I’m not saying I didn’t have second thoughts. I had second, third, fourth, and ninetieth thoughts. I’m saying that it’s possible. Not in a rah-rah way. Take one look at me, and you know I’m not cut out to be a cheerleader. I’m telling you in an “if I can do it, you can too” manner. So, if you’re in a similar position about Marketing AI – your company hates it, but you 100% believe it will be a game-changer for you – start small and grow big. I’m sure there are many other ways to make this all work, and I can only speak to what I know. I’ve written before about how to choose a marketing AI project, selecting a vendor; data and why it’s essential; and how to figure out whether you should build or buy. I’ve also written about why most AI projects fail. Today, I want to discuss the 11 most common roadblocks you may encounter when adding AI to your marketing mix and how you can avoid them. So, let’s begin…. YOU CAN’T EASILY GET THE CUSTOMER/PROSPECT DATA YOU NEED FROM YOUR IT DEPARTMENT, THE LEGACY SYSTEMS, SANTA CLAUS, OR WHOMEVER/WHATEVER Yep. Been there. Felt that. Hated every second. Totally understand your concern. The good news is that you can do many projects without customer data. And before you ask, yes, they’d likely be better with file data but if it’s your biggest roadblock, then just concentrate on the things you can do without customer/prospect info. Projects involving content, conversion, email, social, SEO, paid ads, and the like can all be done without having customer/prospect data. Plus, when you finally have that data, you can overlay it and/or use it for disruption. Is this optimal? Not really, but AI learns as it grows, and the sooner you plant the seeds, the better. This may sound cliché, and it’s important to remember that all this AI stuff is evolving F-A-S-T. Even though the process is often likened to “building another eCommerce site” by vendors, it’s far more sophisticated than that. Getting your feet wet while the water is warm and still shallow is going to be easier – and far more lucrative – than doing it when it’s 30,000 feet deep and frigid. YOU DON’T FEEL LIKE YOU HAVE THE RIGHT EDUCATION/TRAINING TO PULL OFF MARKETING AI PROJECTS Almost all industry surveys show that this is the #1 reason why marketers don’t start AI projects. Folks say they lack the training, education, and/or overall competence. I get it. You should start anyway. Even though AI has been around since the 1950s and used in marketing for decades, the advances in machine learning, computer vision, and natural language processing have taken it to a new level. It’s just the beginning of the Gold Rush, and we’re all beginners. (Yes, even the big guys.) You may say, “well, I’ll just wait a couple of years and then get started.” You do you, and chances are, you’ll regret this. There’s definitely a first-mover advantage with some of this stuff. It’s also important to note that there are things that you’re going to want to claim now. Plus, the longer your models are around, the more sophisticated they’ll likely be. A LOT OF THE VENDORS SEEM SMARMY AF, AND IT’S DIFFICULT TO KNOW WHOM TO TRUST Right now, the Marketing AI world is very similar to the WORLD-WIDE WEB or INTERNET SCHMINTERNET world of the 90s. Snake oil salesmen run amok. The good news is end-users are much better at identifying which companies/vendors are good and which ones to avoid. With that said, Marketing AI is new(ish), and some of these rapscallions, er, vendors, are very clever, so I wrote a whole article on choosing an AI vendor just for you! Access it here. YOU DON’T HAVE A BIAS EXPERT/LEGAL EXPERT/PRIVACY POLICE/SECURITY GURU/FAIRY GODMOTHER WHO SPECIALIZES IN AI AND DON’T KNOW WHERE TO FIND ONE QUICKLY Frankly, this is a concern for even the biggest of companies. The legal/privacy/ethics aspects of AI are new and ever-changing, and no matter how much cash you have, it’s challenging to keep up with the latest and greatest. I’m not a lawyer, nor do I play one on TikTok, and most starter projects will indubitably use your existing resources. (Yes, even if you’re in Financial or Healthcare.) If you are uber-concerned with not having the proper counsel/resources, it may be a good reason to buy your AI instead of building it. (There’s a whole article on whether to DIY or outsource here.) That way, your vendor(s) can help you with some heavy lifting and/or make solid referrals to the people they use. As for bias… It’s usually the sticky wicket. All sorts of people sell themselves as “bias” experts. Many (most?) of them are good with theory but not necessarily with what really happens in practice/the real world. I wish there were a one-size-fits-all answer on handling bias, but alas, there’s not. My recommendation? Identify the bias type(s) that you need help with and then ask your existing resources for referrals. Incidentally, many bias concerns are data-related issues, so be sure to loop in your internal and external data teams. One of the best things I’ve found is to document the problems and what you tried to do to prevent them. (Your lawyer may advise against this. Please check with them first.) MANY/SEVERAL/ALL THE DEPARTMENTS THINK THAT AI SHOULD BE THEIRS BECAUSE IT IMPACTS THEM AND/OR THEY’RE ALREADY DOING IT Yep, this happens a lot, especially in Customer Service and Operations projects. (Chatbots, for example.) For many companies, AI doesn’t have a home yet, so everyone and their pet pig doing something AI-related says they “own” it. If you find yourself in this predicament, John’s wall and Harry Truman would tell you, “It’s amazing what you can accomplish if you do not care who gets the credit.” On the other hand, your third-grade teacher might say something like “keep your eyes on your own paper” and focus on just the marketing projects first. There’s a lot you can do till things really intersect. You could also choose the project and team you’d work best with and forge an alliance. (I see that work too.) Some marketers fight for AI to just be in Marketing. They spend a lot of time developing a strategy, building a plan, going to the mat for their position at the top, etc. I haven’t seen great results with that, but some folks try anyway. (To me, it’s a waste of time and energy. Not to mention it dampens a lot of the other departments’ enthusiasm which seems to come back and haunt you in the future.) You know what’s best for you. Just remember, the more you do AI, the more you’ll want to do ALL the AI-related projects. At some point, you’ll likely need everyone on board, so don’t burn the bridges. YOU ALREADY USE AI IN MARKETING, AND THE RESULTS HAVE BEEN OK, BUT IT HASN’T BEEN THE REVOLUTIONARY, NEVER-BEFORE-SEEN MONEYMAKER THAT YOU WERE PROMISED People have so many excuses for not doing Marketing AI that just thinking about it exhausts me, but I hear you on this one. LOUD AND CLEAR. Vendors/consultants have been selling Marketing “AI” for over a decade now, and a lot of it has been 0% AI and 100% great sales pitch. This may sound like hyperbole, but a chimpanzee can build if/then statements in Excel or Google Docs. There’s no true AI in that logic — or those types of projects — and that’s what many vendors have sold as “artificial intelligence.” Real AI evolves. It processes oodles and oodles of past and current data and changes in real-time and at scale. It learns. It grows. This means that even though a project may start as “meh,” it has the potential to get dramatically better over time. (This does require human supervision.) True AI isn’t just a yes/no formula or decision tree with a few scattered branches; it’s a complex prediction machine that works toward the goal(s) you set to achieve as efficiently and expediently as possible. Some Marketing AI projects will save you time. Others will save you money. Many do both. The only way that you can see how they work for you is to try them. (Yes, you can also build a model to predict whether they’ll work or not, but those are often inaccurate unless you really know what you’re doing.) If you’re gun-shy about trying something new, look at where you can add/enhance AI to your existing projects. Pick the juiciest one, implement it, test it, and then roll it out. Don’t set it and forget it; ensure you’re watching it and adjusting it as necessary. Chances are good that you’ll see what you’ve been missing. Will it make you revolutionary, never-before-seen money? Depends on the project. It certainly has the most potential. MARKETING AI SOUNDS UBER EXPENSIVE, AND YOU DON’T KNOW IF YOU CAN AFFORD IT Marketing AI can be uber expensive, and it can also cost peanuts. It depends on what you’re doing. Most starter projects are cheap and pay for themselves rather quickly. Focus on what you can get done with the budget you have. Remember, Marketing AI isn’t just one thing. It encompasses many different aspects/projects in Voice, email, social, conversion and optimization, advertising, organic traffic, retention, acquisition, analytics, and so on. No matter how much (or little) money you have, you’ll be able to find something to improve with AI. YOU DON’T HAVE THE RIGHT AI PEOPLE ON STAFF – NO DATA ENGINEERS, NO MACHINE LEARNING EXPERTS, ETC. AND/OR YOUR TECH TEAM IS FROM BC (AND I DON’T MEAN BRITISH COLOMBIA) First, it’s essential to know that you’re not alone. This is far more common than you might think. When starting, you may need to lean on outside consultants/vendors to strategize, design, and implement your projects. Sure, you can try to hire all the people you need but recruiting takes time, and chances are, you may not be ready to commit to permanent employees. It’s okay to start small with some solid outside help. A lot of initial AI marketing projects take a good marketer and a great vendor. As you get more projects under your belt, you’ll likely want to build an AI marketing team, but it doesn’t take a big crew to get started. One of the reasons why people think Marketing AI takes so many people is because there are no real standards in job titles yet. It sounds like you need six people to do something when you only need one. Many AI teams are a handful of people. Small but mighty. If you’re in an industry where you have no choice but to build things in-house, and you don’t have the existing people and/or won’t get them soon, try to figure out how where you can upgrade your current projects using AI and/or look at things that concern email, ads, content or one of the other easier things to pull off. YOUR LEGACY SYSTEMS AND/OR TECH INFRASTRUCTURE IS FROM THE LAST CENTURY… MAYBE EVEN BEFORE THAT… AND YOU DON’T KNOW HOW TO WORK AROUND IT – or — THE STUFF (“STUFF” USUALLY MEANS ORDERS, QUOTES, OR INQUIRIES) CAN’T BE PORTED INTO THE BACKEND SYSTEMS TILL XX/XX/XXXX AND XX/XX/XXXX IS A LIFETIME AWAY Build into your budget to do it manually. I know. I know. You’d think there would be a better answer in this day and age, but alas, for many businesses, this is the quickest, least-amount-of-drama answer. Most companies already have a roundabout way to get information into their legacy systems. Once you find out how that works, all you need to do is get your information into the same format. Or, you know, go old-school and just key it in if you must. Once you’ve proven that the project works, the task to get the data to port into your systems automagically is sure to get fast-tracked. Remember, one of the biggest advantages of doing Marketing AI projects is that you’ll have a first-mover advantage. The more you learn, the more time you’ll have to refine your systems. Refinement = $$$. YOU DON’T HAVE THE SUPPORT OF YOUR BOSS AND/OR UPPER MANAGEMENT, AND YOU’RE WORRIED THAT YOU’RE BEING SET UP TO FAIL Every time I give a speech, I get this comment. It’s a prevalent problem in many companies, especially old-school direct marketers, catalogers, and legacy retailers. Having worked around this for years, my best advice is to show them the money. The fastest way to get other folks (especially the muckety-mucks) to adopt/support your Marketing AI projects is to increase sales or save the big bucks. You can also do projects that save time. There’s no doubt those are often some of the worthiest, and I’ve found that translating time savings into dollar amounts works best. When some think of time savings + AI, they go right into the whole “robots are replacing me” spiral, which is often hard to get out of. However, your mileage will vary, so do what’s right for you. Many consultants/vendors will tell you to identify the biggest money-maker you can and then bet it all on that project. I’m a huge risk taker; in most instances, I avoid that recommendation like the plague when convincing non-believers. I’ve found that it works better to take an existing project that’s already good and then show how lucrative it will be with AI. (In the interest of transparency, it can also be the easiest and/or more cost-efficient.) If you prefer to start something new, do something that involves a tight-knit group; doesn’t take months or require lots of resources, and it is easy to show the results and communicate your wins. (Communication is a huge part of AI success.) I’d be remiss if I didn’t mention that one of the things that I’ve also seen work on is changing your project positioning. In other words, don’t sell AI projects as AI projects but sell them as plain ole’ Marketing projects or “advanced” Marketing projects. Then, when things work, you say, “AI was your secret weapon.” DEEP DOWN, YOU THINK AI IS THE ROOT OF ALL EVIL AND WANT TO AVOID IT LIKE THE PLAGUE. People whisper this to me all the time. They’re genuinely afraid of AI and what it will do to the world. Depending on what side of the fence you’re on, you may think they’re entirely crazy or you may be right there alongside them. Either way, though, the AI ship has already sailed. It’s ubiquitous in our daily lives, personally and professionally. You may avoid(ish) it in your daily lives, but from a business perspective, AI is built into search engines, social engines, ISPs, ESPs, and everyone else in between. Even if you flat-out refuse to use it in your business, you will need to learn how to accommodate for it to survive/thrive. The sooner you start, the more influence you’ll have over your models, and the better (and more just) you can mold them. What other reasons are holding you back from Marketing AI projects? Are you already doing them and have a tip you’d like to share? A question you’d like to ask? Tweet @amyafrica or write firstname.lastname@example.org. Parting Wisdom from John’s wall and Theodore Roosevelt: It’s hard to fail, but it’s worse never to have tried to succeed.