I’m going to say something that might get me kicked out of the Silicon Valley entrepreneur’s club.
Building an “AI culture” is a waste of time.
There, I said it. I’ve spent years in the trenches, building and selling companies like RemoteTeam, which was acquired by Gusto. I’ve invested in over 200 startups, including some of the biggest names in AI like Anthropic, OpenAI, and Scale AI. I’ve seen what works and what’s just corporate theater. And I’m telling you, the endless workshops, the rah-rah all-hands meetings, the “AI for Everyone” Slack channels… they’re mostly noise. They make executives feel good, but they don’t actually move the needle.
After selling my last company, I got obsessed with a question: why do some remote companies thrive with AI while others just spin their wheels? I’m a systems guy. I believe in finding the underlying patterns. So, I did what any good angel investor would do. I went digging. I analyzed six of the most successful remote-first companies in my portfolio and beyond. I’m talking about companies that are not just surviving, but absolutely crushing it. And I found something interesting. They all had one thing in common when it came to AI. And it had almost nothing to do with “culture.”
The Culture Trap
Let’s be honest. “Culture” is a fuzzy word. It’s hard to define, harder to build, and almost impossible to measure. When leaders say they want to build an “AI culture,” what they usually mean is they want their employees to be excited about AI. They want them to use it. But excitement is a feeling. It’s fleeting. You can’t build a business on feelings.
I saw this firsthand at RemoteTeam. We were a fully remote company from day one. We lived and breathed asynchronous communication and digital tools. When the new wave of generative AI hit, we were excited. We set up a channel, shared cool tools, and encouraged everyone to experiment.
And for a while, it was great. People were generating images for blog posts, summarizing meeting notes, and writing first drafts of emails. But then, the novelty wore off. The channel got quieter. People went back to their old workflows. The initial burst of activity didn’t translate into a sustained, competitive advantage. It was a sugar high, not a systemic change.
Why? Because we were focused on the what (using AI tools) without a system for the how and the why. We were building a culture of AI enthusiasm, not a system of AI leverage.
The One Thing That Actually Works: Systematized AI Leverage
So what was the common thread among the successful companies I studied? It wasn’t a better AI chatbot or a more enthusiastic CEO. It was this:
They built a system to continuously identify, prioritize, and automate high-leverage tasks with AI.
That’s it. It’s not sexy. It’s not going to be the theme of your next offsite. But it works. It’s a machine, not a mood. It’s a process, not a pep rally. It’s about making AI a core part of your company’s operating system, not just a fun app on your phone.
Think of it like this. A restaurant with a great “food culture” might have passionate chefs who love to cook. But a successful restaurant chain has a system. They have standardized recipes, a streamlined supply chain, and a repeatable process for training staff. The passion is still there, but it’s channeled through a system that ensures quality and consistency at scale.
Systematized AI Leverage is the same idea. It’s about moving from ad-hoc, individual experimentation to a structured, company-wide process for finding and exploiting opportunities for AI-driven efficiency and innovation.
How to Build Your System for AI Leverage
So how do you actually do this? It comes down to four key components. I’ve seen this pattern work for hybrid teams, fully remote companies, and everything in between.
1. The AI Audit: Find the Leverage Points
You can’t improve what you don’t measure. The first step is to systematically map out your company’s workflows and identify the tasks that are ripe for AI intervention. This isn’t a one-time exercise. It’s a continuous process.
- Start with a simple question: Ask every team and every individual: “What are the most repetitive, time-consuming, or low-value tasks you do every week?”
- Create a central repository: Build a simple database (an Airtable or Notion page works fine) to log these tasks. For each task, capture:
- Task Name (e.g., “Manually updating the sales CRM”)
- Team/Individual responsible
- Time spent per week
- Potential for AI automation (Low, Medium, High)
- Prioritize ruthlessly: You can’t boil the ocean. Focus on the tasks that are high-frequency, high-time-suck, and have a clear path to automation. A good rule of thumb is to look for tasks that take up more than 2 hours a week and are performed by multiple people.
One of the companies I advise, a fast-growing fintech startup, did this and found that their support team was spending a combined 80 hours a week manually categorizing and routing support tickets. That became their number one priority for AI automation.
2. The AI Toolkit: Standardize Your Stack
The world of AI tools is a chaotic mess. There’s a new “game-changing” app every day. If you let your team use whatever they want, you’ll end up with a fragmented, insecure, and expensive mess. You need to standardize your toolkit.
- Choose a core set of tools: Based on your AI Audit, select a small number of powerful, flexible tools that can address the majority of your high-leverage tasks. For most companies, this will include:
- A powerful LLM for text generation and analysis (like the ones from OpenAI or Anthropic)
- A tool for workflow automation (like Zapier or Make)
- A specialized tool for your core business function (e.g., a code generation assistant for engineers, a content generation platform for marketers).
- Go deep, not wide: Instead of chasing every new shiny object, become experts at using your core toolkit. Provide training, create internal documentation, and build a library of best practices and prompts.
- Integrate with your existing systems: The real power of AI comes from integrating it into your existing workflows. Use APIs to connect your AI tools to your CRM, your project management software, and your communication platforms. For async communication, this is critical. Imagine your project management tool automatically summarizing progress and flagging bottlenecks using AI. That’s leverage.
3. The AI Onboarding: Train for Impact
Onboarding is your first and best chance to instill your system of AI leverage. Don’t just give new hires a list of tools and wish them luck. Train them on the system.
- Make it part of Day One: Your AI system should be as fundamental to your onboarding as setting up email or Slack.
- Focus on workflows, not just tools: Instead of a generic “Intro to ChatGPT” session, show new hires how to use AI to perform specific tasks within your company’s workflows. For example, a new salesperson should learn how to use AI to research leads, draft outreach emails, and update the CRM—all within your standardized toolkit and processes.
- Build an AI “buddy system”: Pair new hires with an experienced employee who can mentor them on using the AI system effectively. This is especially important for digital nomads and remote employees who can’t just lean over and ask a coworker for help.
4. The AI Flywheel: Create a Feedback Loop
This is the most important part. Your system needs to learn and improve over time. You need a flywheel that constantly feeds new ideas and improvements back into the system.
- Measure everything: Track the impact of your AI initiatives. How much time are you saving? Are you closing more deals? Is customer satisfaction improving? Use this data to prove the ROI of your system and justify further investment.
- Create a formal process for suggestions: Set up a simple way for employees to suggest new tasks to automate or new ways to use your AI toolkit. This could be a form, a Slack channel, or a monthly meeting.
- Reward system-builders, not just users: Recognize and reward the employees who are not just using AI, but are actively improving the system. These are the people who are creating new workflows, building new automations, and finding new ways to create leverage. They are your force multipliers.
One of the most successful companies I’ve invested in, a leader in the data infrastructure space, has a quarterly “Automation Award” for the employee who builds the most impactful AI-powered workflow. The prize isn’t huge, but the recognition has created a powerful incentive for everyone to think like a system-builder.
Stop Chasing Culture, Start Building a Machine
Look, I get it. “Culture” is an appealing concept. It feels human. It feels collaborative. But in the fast-moving world of AI, it’s a trap. It’s too slow, too vague, and too disconnected from the bottom line.
Don’t waste your time on AI pizza parties and prompt-a-thons. Stop trying to manufacture a feeling. Instead, get your hands dirty. Build a machine. A system. A process. A flywheel.
Start with a simple audit. Standardize your tools. Train your people on the system, not just the tech. And create a feedback loop that makes the machine smarter every single day.
This is how you win with AI. Not by talking about it, but by systematically embedding it into the core of your business. It’s not as glamorous as a “culture of innovation,” but it’s a hell of a lot more effective. And in the end, results are the only thing that matters.
Frequently Asked Questions
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.