How Does Alex Hormozi's Team Use AI in Support and Sales?

How does Alex Hormozi's company use AI?

Alex Hormozi, founder of Acquisition.com, says his team uses AI in four main places. Five support agents handled roughly 120,000 tickets during a book launch, and he says 90% were resolved with no human involved. Editing is split into small AI workflows. Meeting transcripts feed content ideas. A single AI text agent sells around 200 units a month. Humans still handle phone sales and big decisions. These are his own statements from talks, podcasts and posts, not audited results.

How did five AI agents handle 120,000 support tickets?

For the launch of a book, Hormozi's team set up five agents to answer incoming support requests. He described the volume as "somewhere in the neighborhood of" 120,000 tickets and said 90% were resolved without any human intervention.

The reason this works is in how he described the tickets. Most were people asking where to get the free stuff that came with the book. That is a narrow, repetitive question with a known answer. An agent does not need judgment to point someone to a download link. It needs the right information and a clear set of instructions.

The same pattern shows up in a smaller example. A woman named Tina ran cohort programs out of Telegram, and her team kept answering the same questions. Hormozi told her to build an agent that searches past answers and suggests one, so a team member can check it and hit send. A person stays in the loop. Months later she reported she had left Telegram for a custom dashboard and built nine other tools for her business.

If your own inbox is full of the same ten questions, that is your version of this.

How does Hormozi's team use AI for video editing and content?

His framing is the useful part. The old thinking, he said, is "I need to hire an editor." The new thinking is to ask what the five things an editor actually does to make a video are, and turn each one into a workflow.

For a clip from his Hormozi Hotline, he described a workflow that:

1. Reads the transcript.
2. Finds the highest-tension moments.
3. Starts the clips at those points.
4. Removes the ums and ahs and cuts the clip together.

He said the team's media output "tripled in the last month" with no added headcount, because prompts are taking over 30, 40 or 50% of the workload. His stated direction is that one person does the work of three, then five, then ten.

Content ideas work the same way. He records and transcribes his meetings, then has AI scan them along with his calendar to pull out the most interesting moments. His videos then cover things that happened to him that week, not invented stories.

A prompt template for mining your own meetings

If your calls are already recorded and transcribed, you can try a version of this. This is a starting template, not Hormozi's actual prompt.

Here are transcripts from my meetings this week and my calendar.
Find the 10 most interesting moments: disagreements, surprises,
mistakes, decisions, or lessons. For each, give me:
1. A one-sentence summary
2. The exact quote or passage
3. Why someone outside my company might care
Do not invent details that are not in the transcripts.

Does Hormozi let AI write for him?

No. He is clear on this. He said tweets used to take about an hour and a half a day and now take about 90 minutes a week, which he calls seven times more efficient. But he said, "It's not AI making my tweets." AI helps him generate ideas to iterate on.

He is also blunt about AI-written content in general. He dislikes the patterns that give it away, especially forced negation ("it's not this, it's that") and short staccato sentences. His fix for people who do write with AI is to give it about 12 rules it can never break and 16 samples of your own writing, which he says makes the output about five times better.

His broader argument is that reality is the advantage. If two people post the same six tips and only one has built ten real sales teams, he expects the second to win.

Can an AI SDR replace human sales reps?

Hormozi's own timeline is worth reading because he has made the predictions in public:

So text-based selling is working for his team, while voice is not there yet. He said humans are still winning on the phone even after two years of work on voice AI. If you are evaluating an AI sales tool, that split is a reasonable benchmark: test it on text channels first and keep people on live calls.

When does AI waste money?

Hormozi's strongest warnings are about spending. He told a story about a business paying about $11,000 a month for 11 virtual assistants doing data cleaning. The work was fine. The company then spent $350,000 building an AI system to replace them, and he said the task was not even the constraint on the business.

His own company ran into a version of this. In a September 30 memo, he said his team spends 16 times more on AI per employee than the average tech-forward company, and that "company revenue is flat compared to AI usage increases." His team gets no more AI budget, only training on what is worth working on and how to use fewer tokens (the units AI providers bill by) for the same output.

His test is simple. If you use AI all day and are not making more money, you probably have the same problem you had before, "except this time it's costing you tokens." Ask three questions:

• Am I getting more views since I started using AI?
• Am I getting more leads?
• Am I making more money?

If the AI bill rises and none of those do, find the real constraint (usually leads, sales or delivery) and aim AI there. He also said it is more efficient to decide a task is not a priority than to automate it.

Is AI taking jobs at Hormozi's company?

The source says he has never claimed to have fired anyone because of AI, and that his team is still hiring. He compared it to AI labs, which employ thousands of people.

His advice to other owners is harsher. He said to raise the bar for the whole company, and that people who cannot meet it will not stay. He told one owner of a $20 million website company to spend six months rebuilding workflows and cut headcount in half, which he said would roughly double the margin.

For employees, his answer is to get there first. He calls it "bring your own agent": train agents to do what you do, so you can offer a business the output of a whole department. The creator of the video that summarized his comments said that in a survey of over a thousand people in his community, 70% of those who used AI on a real work problem more than once saw a real career change, though that is a self-selected group.

What should you take from this?

Hormozi's approach comes down to a few habits you can copy at any job.

• Break your job into tasks. List the five or six things you actually do, like an editor's cut, review and cleanup steps. Each one is a candidate for its own workflow.
• Start with repetitive, answerable work. Support questions, status updates and first drafts fit. Keep a human check on anything with consequences.
• Block time for it. His suggestion is four hours a week, and in each block you automate or amplify one specific task.
• Keep decisions yourself. He says he asks several AI models the same hard question and gets different answers, and that delegating decisions makes you dumber. His rule is to outsource the typing, not the thinking.
• Measure against a business number. Views, leads, revenue, hours saved. Not tokens used.

His view on what stays human is risk. AI can give recommendations, but someone has to own the decision and the money behind it.

What could you build yourself from these ideas?

The support example translates well. A team that answers the same questions over and over could use Remy to describe a small internal tool in plain English: one that searches past answers and replies, drafts a suggested response from them, and lets a person approve it before sending. Remy builds and deploys full-stack apps with a database, sign-in and document search with cited results, so a shared team dashboard is within reach. It beats juggling a spreadsheet and a chat app, and you own the result. A tracker that logs AI spend against leads or revenue is another option.