Fix My Business

I Keep Building AI Agents and Nothing's Working

John writes: "I'm six months in, building AI agents for follow-ups, email, lead management. I've spent weekends on it. Nothing's working. I feel like I'm getting further behind." The belief shift: AI agents work best when the process is already proven. If you're six months in and still figuring out what success looks like in sales, lead management, and pitching—you can't hand that to AI. That's like hiring an employee and saying "figure it out." They'll struggle too. The hidden tax: AI agents require onboarding, maintenance, troubleshooting. In the early days, that falls to you. Scott spent Easter weekend keeping his AI agents running after Anthropic changed their terms. The real shortcut is a human. You can add a VA within 48 hours. Train them. Get traction. AI and automation come later—once you know the processes work. The five levels: You need oxygen first (cash, deals). Then profit. Then order. AI falls within order. If you're automating before you have sales, you're out of sequence. The prescription: Stop building AI. Go to a human. Fix one task. Get them good for 90 days. Then integrate AI back into the workflow. The math doesn't work when you're shouldering AI maintenance and trying to launch at the same time. Got a business question? Ask Scott here: scotttodd.net/ask [https://www.scotttodd.net/ask]

I Thought I'd Be Further Along by Now

Andrea started her business 10 months ago as a side hustle. She thought she'd be further along. More revenue. More profit. Less time. She's considering quitting. The wall is real. Around 9-10 months, the initial excitement burns off. The wins don't come fast enough. The inner voice says: call it a day. Scott almost tapped out at nine months. If he had, he'd be in a corporate job today. It's physics, not failure. Starting a business takes rocket launch energy. A side hustle splits that energy three ways: day job, family, business. If you're launching on a third of the fuel, the timeline is longer. Give yourself grace. There are no bad goals. There are unrealistic time horizons. If you thought you'd be out of your job in a year, make it three. The goal isn't wrong. The timeline was. The 30X training rule: If a task takes an hour a week, be prepared to spend 30 hours training someone to do it. The control trap kills side hustlers. You only have a few hours. If you spend them running everything, you'll never get leverage. When you're feeling defeated, stop. Look at your tasks. Is it $10/hour work or $1,000/hour work? If it's $10/hour, it's not meant for you. The J-curve: Profitability dips when you hire. Then it rises. That's the trade. The ones who break through aren't working harder. They're letting go of work they were never intended to do. Congratulations. You are right on schedule. Got a business question? Ask Scott here: scotttodd.net/ask [https://www.scotttodd.net/ask]

I Keep Writing Ads But Nothing's Working

Josh writes: "I keep writing ads but I'm not getting leads. I've hired people. Tried different headlines, different platforms. My team says the ads are good. But no leads." The diagnosis: You're writing to everybody. Which means you're writing to nobody. The market is always talking to you. View counts show people are seeing your ads. If they're not converting, the audience is telling you something. The question is whether you're listening. The Visibility Trap: When you don't understand your customer, you manage with your gut. And your gut is often wrong. You lack data on who your customer actually is and what problem they're trying to solve. 8 billion narratives: Everyone walks around with a story in their head. Your job is to match your message to one of those stories. The fishing metaphor: Pick one problem. Write to that problem. If nobody responds, that's not the bait they want. Try another problem. Keep fishing until something bites. Scott's ATV story: Didn't know why people wanted land in one area. Discovered buyers wanted to ride ATVs. Started writing to that. More ATV people came. Applied the same approach in Nevada—asked a truck driver why he was looking for land. Used that intelligence. It compounds. The fix—four questions before you write: 1. Who is this for? 2. What do they believe? 3. What are they afraid of? 4. What words do they use? When you put those four elements into your copy, more people like that begin to respond. Got a business question? Ask Scott here: scotttodd.net/ask [https://www.scotttodd.net/ask]

I Spent Months Building an AI Agent. Here's the Part Nobody Talks About

Business owners are wasting massive amounts of time building AI agents when they should be figuring out what their business actually needs. Scott isn't anti-agent—he has them in his business. But there's an untold tax: onboarding, maintenance, troubleshooting. In small businesses, guess who does it? You. The Easter Weekend story: Anthropic changed their terms. 3pm Saturday, Scott's OpenClaw stopped working. He spent Easter weekend switching to ChatGPT. His wife was not happy. The hidden burden: Every new OpenClaw release broke what he'd already fixed. Nights and weekends troubleshooting. Time spent with AI agents is time not spent with the team. Digital dust: The AI is producing outputs Scott doesn't have time to review. It's not being used. It's digital dust. The new default: When work needs to come off Scott's plate, the desire is to give it to AI. But his default now is to give it to a human. A VA can be onboarded in 48 hours. An AI agent might take 100+ hours of teaching, testing, and overseeing. Humans first: Humans allow you to pressure test processes. They know if something's working. They can find a better way. If a process doesn't work with humans, it won't work with AI agents. You're just automating chaos—and chaos amplifies. The broken speaker: Turn up the volume on a broken speaker and the distortion gets louder. Apply AI to broken processes and the chaos happens faster. The close: If you've been putting off the AI agent thing because something feels off—trust that instinct. Got a business question? Ask Scott here: scotttodd.net/ask [https://www.scotttodd.net/ask]

How to Grow as a Leader

Two scenarios: Your team won't stop asking questions. Or you're solo and think this doesn't apply. It does. The nine-year employee story: A longtime team member messages: "I gave a refund, they got rebuilt. What happened?" Scott's response: "I don't know. Check with the back office team." The reply: "But you've always been the one that knew." True. But if Scott jumps in, he deprives the team of the ability to learn. It took every amount of restraint not to swoop in. But that restraint is what allows growth. The capability ceiling: Your business will only rise to the capability of your weakest team member. Every time you answer a question you shouldn't, you pull yourself down and limit the team's growth. The ego trap: "It's gonna feed my ego—'man, they still need me.' But my team didn't grow. They're still reliant on me. And that means the business is reliant on me forever." The tactic: Act like you don't know, even if you know. Redirect. "I don't know. What do you think?" Virtual assistants: We hire them for one task and never explore their full capability. One of Scott's VAs was massively underutilized—until he got to know them. The football analogy: Your quarterback should be the kicker. Your wide receiver should be the quarterback. You have talented people in the wrong positions. Strong team, no wins. The solo-preneur trap: "I tried hiring. The people weren't any good. Nobody wants to work." Maybe true. But Disney has 60,000 employees. They didn't magically find 60,000 perfect people. They invested in a process that brings people to standard. Choose Your Own Adventure: Every decision turns a page. Each choice moves you closer to your goal, keeps you sideways, or pulls you back. The question: Which adventure are you writing? Got a business question? Ask Scott here: scotttodd.net/ask [https://www.scotttodd.net/ask]