''' I once saw a multi-million dollar AI project go up in smoke. We had the best engineers, the fanciest algorithms, and a C-suite that was practically drooling over the expected ROI. A year later, the project was quietly shelved, the team was demoralized, and the only thing we had to show for it was a very expensive lesson in what not to do. The tech was perfect. The execution was a disaster. Why? Because we broke the unspoken rules of AI change management.
Everyone is scrambling to integrate AI, and they're all making the same mistakes. They think it's a technology problem. It's not. It's a people problem. I've been fortunate enough to be in the trenches for a few of these transitions, both at my own companies like RemoteTeam and as an investor in over 200 startups, including some of the biggest names in AI like Anthropic and OpenAI. I've seen the patterns. I've seen what works and what leads to chaos. And I can tell you that there are rules to this game that no one talks about publicly. I'm breaking the silence.
Rule 1: Don't Boil the Ocean
The biggest mistake I see is trying to do too much, too soon. The ambition is intoxicating. "Let's build a custom, end-to-end AI-powered operating system for the entire company!" It sounds great in a press release, but it's a recipe for failure. I saw a promising startup try this. They spent 18 months and burned through $5 million trying to build a monolithic AI system to manage everything from sales to HR. It collapsed under its own weight. They never even launched a single feature.
Start small. Seriously. Find one, specific, high-pain problem and solve it with AI. Pick a process that is manual, repetitive, and a bottleneck. At RemoteTeam, we didn't try to automate all of HR overnight. We started with a simple AI tool to help with onboarding paperwork. It was a small win, but it had a huge impact. It saved our HR team hours of tedious work, and it showed the rest of the company that AI could be a powerful ally, not a threat. Once you have that first win, you build momentum. You earn trust. Then you can start to tackle bigger challenges.
Another example I saw was a large e-commerce company that tried to implement a fully autonomous warehouse system in one go. They spent a fortune on robots and software, but the system was so complex that it never worked reliably. The result was chaos, with orders being lost and delayed. They eventually had to scrap the whole project and go back to a more manual system. They would have been much better off starting with a smaller, more manageable project, like using AI to optimize their picking routes.
Rule 2: Your People Aren't Luddites, They're Scared
If you walk into a room and announce, "We're implementing an AI that can do your job faster and better," you shouldn't be surprised when you're met with resistance. People aren't afraid of technology. They're afraid of being replaced. They're afraid of becoming obsolete. And frankly, their fears are often justified.
I remember a team of analysts who were incredibly hostile to a new AI-powered data analysis tool we were introducing. They dragged their feet, they found every possible flaw, and they complained endlessly. We were about to pull the plug when we finally sat down and had an honest conversation. They thought we were trying to get rid of them. We explained that the goal was to free them from the drudgery of data cleaning so they could focus on higher-level strategy and analysis. We showed them how the tool would make their jobs more interesting, not eliminate them. It was a turning point. They became the tool's biggest advocates.
Communication is everything. You have to be radically transparent. You have to focus on augmentation, not automation. Frame the AI as a tool that will make your team more effective, more creative, and more valuable. Show them a path forward. If you don't, they'll fight you every step of the way. We held a series of town hall meetings where we demonstrated the tool and answered every question, no matter how hostile. We also created a series of training workshops to help the analysts learn how to use the tool effectively. It was a lot of work, but it was worth it.
Rule 3: The "AI Guy" Can't Be a Lone Wolf
Another classic mistake is siloing your AI efforts. You hire a brilliant data scientist, give them a team, and tell them to "go do AI." They work in a vacuum, build something technically impressive, and then try to foist it on the rest of the company. It almost never works. The other departments don't understand it, they don't trust it, and they haven't had any input into its development. It's a solution in search of a problem.
AI is not an IT project. It's a business transformation. You need buy-in from every corner of the organization. When we started our AI journey at MovieLaLa, we created a cross-functional task force. It had representatives from engineering, marketing, sales, and customer support. They were responsible for identifying opportunities, defining success metrics, and championing the projects within their own teams. It wasn't always easy. There were heated debates and turf wars. But it was essential. By the time we launched our first AI-powered feature, the entire company was behind it.
One of the biggest challenges we faced was getting the different departments to speak the same language. The engineers would talk about algorithms and data structures, while the marketing team would talk about customer engagement and brand identity. We had to create a shared vocabulary and a common set of goals. We also had to be careful to manage expectations. The marketing team was expecting a magic bullet that would solve all their problems, while the engineering team was more realistic about the limitations of the technology. It took a lot of patience and communication to get everyone on the same page.
Rule 4: Garbage In, Garbage Out. Still.
This is one of the oldest adages in computer science, and it's more relevant than ever in the age of AI. Your fancy new algorithm is only as good as the data you feed it. I've seen companies spend millions on AI models, only to have them produce nonsensical results because they were trained on messy, incomplete, or biased data.
Before you even think about building an AI model, you need to get your data house in order. This is the unglamorous, behind-the-scenes work that nobody wants to do, but it's absolutely critical. You need to invest in data infrastructure, data governance, and data quality. You need to have a clear understanding of where your data comes from, how it's collected, and what it represents.
I once advised a company that was trying to build an AI-powered recommendation engine. The recommendations were terrible. It turned out they were training the model on a hodgepodge of old, unstructured data from a dozen different systems. We had to pause the project for six months and undertake a massive data cleaning and consolidation effort. It was painful, but it was necessary. We had to manually go through thousands of records, correct errors, and fill in missing values. We also had to build a new data pipeline to ensure that all new data was clean and consistent. Once we had a clean, reliable dataset, the model started to sing.
Rule 5: Don't Outsource Your Brain
There's a huge temptation to just throw money at the problem and hire a big consulting firm to come in and "do AI" for you. Don't do it. While external partners can be valuable, you can't completely outsource your AI strategy. You need to build in-house expertise. You need to have people on your team who understand the technology, the data, and the business context.
I saw a company make this mistake. They hired a well-known AI vendor to build a custom solution for them. The vendor came in, built a black box system, and then left. The company had no idea how the system worked, and they were completely dependent on the vendor for any changes or updates. When the vendor inevitably raised their prices, the company was stuck. They had outsourced their brain, and now they were paying the price.
Building an in-house AI team is a long-term investment, but it's one that will pay off in the long run. You'll have more control over your destiny, you'll be able to move faster, and you'll be able to build a sustainable competitive advantage.
Rule 6: Measure What Matters
It's easy to get caught up in the technical details of an AI project and lose sight of the business goals. I've seen teams celebrate the fact that they've built a model with 99% accuracy, only to find that it has no impact on the bottom line. Why? Because they were measuring the wrong thing.
Before you start any AI project, you need to define what success looks like. What are the key business metrics you're trying to move? How will you measure the impact of the project? And how will you know if you're on the right track?
I once worked with a company that was using AI to predict customer churn. They had built a very accurate model, but it wasn't having any impact on their churn rate. It turned out that they were only predicting who was going to churn, but they weren't doing anything to prevent it. We had to go back to the drawing board and redefine the project goals. We created a new set of metrics that were focused on customer retention, not just prediction. We also built a series of automated interventions that were triggered when a customer was identified as being at risk of churning. The result was a significant reduction in their churn rate.
The Real Secret to AI Success
I could give you a dozen more rules, but they all boil down to this: AI is not a magic wand. It's a tool. And like any tool, its effectiveness depends on the skill and wisdom of the person wielding it. The companies that will win with AI are not the ones with the most sophisticated technology. They're the ones that understand that this is fundamentally a human challenge. They're the ones that lead with empathy, communicate with clarity, and build a culture of trust and collaboration. The technology will change. The unspoken rules of human nature will not. What are your unspoken rules for AI change management? Share them in the comments below. '''
Frequently Asked Questions
What if I disagree with some of the advice?
Good. That means you're thinking critically, which is exactly what a good founder should do. Take what resonates, test it, and discard what doesn't work for your specific situation. No advice is universal.
How often is this guide updated?
I revisit and update my guides regularly as I learn new things and as the market evolves. The core principles tend to stay stable, but specific tactics and tools get refreshed based on what's working right now.
How should I work through this guide?
Don't try to absorb everything in one sitting. Read through once to get the big picture, then go back and work through each section as it becomes relevant to your current challenges. Bookmark it and return to it regularly.
Who is this guide designed for?
This guide is written for founders and operators who want practical, actionable advice rather than theoretical frameworks. Whether you're just starting out or scaling an existing business, the principles here apply across stages.