Thirty Minutes a Day Is a Week of Your Life Back

Listen to the episode with Doug Dvorak

The short version

• Save thirty minutes a day and you have bought back a week of your year. That is the easiest AI win most people skip.
• We run agents that check other agents, and when something looks wrong I ask the AI to prove it. It pulls the source.
• An AI operator costs about what one SDR costs, but the second, tenth and hundredth role does not cost another sixty thousand.
• Follow up is where the money is, and it is the first thing a busy rep drops.
• I slept on AI, kept a team of four developers too long, and now write all the code myself.
• In two to three years we will hire humans who can manage a team of AI the same way we hire humans who can manage humans.

The surprise was not the big stuff

Doug asked me what AI result in my own business actually surprised me. It was not replacing a CRM or building new tech. It was how easy it is to get a little time back. If you claw back thirty minutes a day, every day, on something routine you already do, that is a week of free time in a year. There is a lot you can do with seven extra days.

So instead of chasing the big transformation, I broke it down to the small questions. Can it check my emails and flag the ones I actually need to respond to. Can it do a first round draft of the response with all the facts in it, so I just proofread and send. That was the epiphany for me, and it is still the one I give people first.

if we could get back thirty minutes a day, every day of the year, just something that we do routinely, that's one week of free time that we bought back in a year

What I do when I think the AI made something up

We build our own systems on top of things like ChatGPT and Claude, and the stack is a recursive learning system. One agent does the work, another agent checks it, and another agent checks that check. Two quality tests sit in the middle before anything ships. It is getting better and it still makes mistakes.

Here is the story I told Doug. Luna, my AI operator, replied for me and listed names I did not recognize. I assumed she had invented them. She had not. A client had buried their team members in a long email with a pile of attachments, I had skimmed past that part, and she had correctly mapped each person to their role in the task list. That is the habit now. When something looks wrong I ask her to prove it, and she pulls up the exact email.

For emails, spreadsheets and documents I would put the accuracy closer to 99 percent, and I still put my eyes on it. What it does that I cannot do is sweep every historical thread and organize all of it faster than I ever could. When we get something wrong in our company, we own it, we apologize, we say we are still training the models, and we move on. I would rather be early and make a few mistakes than wait for perfect. If you are in insurance or financial advice or another regulated field, be far more protective than I am.

AGI and agentic, without the jargon

At the highest level, AGI just means it is smarter than humans and can do things on its own. It does not need you to prompt and direct every step. Luna books me on podcasts. I know that method and I know how labor intensive the follow up is, so I taught her the lessons and now she teaches herself. She notices a response came back from Doug, marks it as a win, and goes back to analyze what produced it.

Agentic is the other half. ChatGPT, Grok and Claude wait for you to type. Agentic means it runs on a schedule, or when something happens. On the internet these are webhooks or actions. When an email arrives, act on it. When a support ticket comes in, try to resolve it. When money hits the account, text me. You do not say every morning at 7AM go check my email. You say check my email every fifteen minutes, follow up, and alert me if anything is urgent, and from that point forward it runs on its own.

The SDR math nobody runs properly

A human SDR with commissions, overhead and taxes runs you somewhere around 60 to 70,000 a year. Running Luna costs about the same. One to one, it is a wash. That is not where the argument is.

The argument is what happens next. With a human you get a computer, an office, training, and the capacity of exactly one person who has emotions and vacation time. Growth means a second SDR at another 60,000. With an agentic setup, the initial investment is roughly that first employee, but the second role does not cost another 60,000. It might tick up 5,000 because of more API usage. No retraining, no FICA, no turnover, and SDR turnover is brutally high.

We are testing this live right now with Russ Ruffino, who runs a large coaching company. Someone watches a video sales letter, clicks the button if they want a conversation, and has a full voice conversation with his AI operator that answers their questions and moves them into a more formal closing conversation. We are building the case studies as we go.

The difference is Luna can serve as 10 or 50 or a 100 employees and have infinite scale and infinite ability.

Follow up is where the money is

Doug asked where busy sales professionals waste the most time that AI can fix. The follow up. That is where the money is. Going through leads to figure out who is hot. Watching whether they visited the site, opened the email, clicked the link. Texting at the right interval. Noticing that five or ten of your potential clients are at the same conference you are going to, so you can take them all to dinner at once.

It does not just save time, although it is tiring work. It converts better, because it gets handled on the days you do not have the time or the energy. Same logic applies to the CRM. Every salesperson hates updating it unless they are marking a deal closed. Let the operator update it every time you email, text or call, because that data matters to the whole company.

I use the same approach on my own calls. I make promises in a sales conversation, then my AI operator turns the transcript into a deliverable or a statement of work I can send back as a proposal. I am not taking notes during the call and I am not listening to it again later to remember what I committed to.

In the follow-up. That's where the money's at.

What I got wrong, and what it cost me

I slept on it. I am an old school developer, I have written a lot of code, and I played with the early tools and walked away. Lovable got me to 90 percent and not to the finish line. I had a full tech team of four developers, and when I asked them whether we could start coding with AI they said it was not good enough yet. They were right at the time.

In hindsight I should have jumped back in and relearned coding with the tools that existed. We made the switch about a year ago. I let the full team go and I write all the code myself with AI. We have shipped far more, solved far more problems, and it is cheaper and a better solution for us. When Doug asked my favorite tool in the rapid fire round, I said Codex.

For sales leaders: call reviews and the empathy gap

If you manage three reps who are on the phone four or five hours a day, you do not have fifteen hours a day to review calls. So the easiest high impact move is automatic call reviews. Send a notetaker like Fathom, which is free, into every call, or pull recordings out of whatever system you use, and have a model score and coach every single call against your script or your framework. Your reps get feedback immediately and you learn what actually works.

That also hands you a marketing asset. The key to marketing is answering people's questions, and your sales team collects every question that exists. Pull them out of the transcripts and have marketing build content around them. Better content, better leads, more closed deals.

The leadership difference is context. Instead of asking AI what to say to Johnny because he had a bad day, which gives you something cheesy that people will laugh at, you walk up and say you did 14 calls today and that is not your normal, what is going on. Alex Hormozi talks about what life does to our business, and Gary Vee talks about empathy. If someone usually performs and suddenly does not, go to the life question before the performance question.

I think in the future, in the next two to three years, we're gonna look for humans that can manage a team of AI just the same as we look for humans that can manage a team of humans.

Start with a pen and paper

If you have thirty minutes this week and you have never used any of this, do not go shopping for tools. Get a pen and paper and write down what you do every day. Then open ChatGPT or Claude, both of which have free plans, paste in your list, and say here is what I do every day, what can you help me with. Read the answer. Ask a follow up question if you get confused.

It will come back with things like connect your Google Apps account and I will manage your inbox, or connect your Google Drive, give me a couple of sample proposals, and I will draft them for you. Then pick one. This is where people get overwhelmed, because they say go do everything. Do one task, get it right, take that weight off your shoulders, then go to the second and the third. Repeat.

On security, connecting to ChatGPT, Claude or Grok is not the risk people imagine. It is similar to connecting your calendar to Calendly or wiring something into HubSpot or Salesforce. Use normal password hygiene, turn on two factor, do not be reckless with permissions, and do not plug random things you found in a Facebook group into your accounts. For a nontechnical salesperson, Grok is the fun one right now because it makes it easy to spin up small bots with specific jobs.

The book that shaped how I think about selling is Delivering Happiness by Tony Hsieh, the founder of Zappos. They gave the customer service team authority to ship free, discount and refund without escalating, because happy customers build evangelists. I could learn a better closing tactic, but if the customers are happy with what they bought, the stream never ends. The other one I gave Doug was the Win Without Pitching Manifesto by Blair Enns, because it taught me to be a specialist and not a generalist.

the human to human relationship matters less because the human's not making the buying decision anymore. The AI operator is. So it's the AI to AI relationship that starts to matter.

The moments worth clipping

Short cuts from this episode, if you want the highlights before the full thing.

Questions people ask about this

How much does an AI operator cost compared to hiring an SDR?

A human SDR with commissions, overhead and taxes runs around 60 to 70,000 a year, and running Luna costs about the same. The difference shows up on the second hire. Another human SDR is another 60,000. Expanding the AI operator into another role might tick up 5,000 because of more API usage in the models, and it can serve the workload of 10 or 50 or a 100 employees.

How accurate is AI for emails, spreadsheets and documents?

For me it is probably closer to 99 percent, and I still read everything before it goes out. What it does better than I can is sweep every historical email and thread and organize the facts faster than a human. When something looks wrong, I ask it to prove the claim and it pulls up the exact source.

Is it safe to connect AI to my email and calendar?

Connecting to ChatGPT, Claude or Grok is comparable to connecting your calendar to Calendly or wiring something into HubSpot or Salesforce. Those companies are focused on security and we are not seeing widespread breaches. Use normal password security, turn on two factor authentication, do not be reckless with permissions, and avoid random tools you found in a Facebook group.

Which AI tool should a nontechnical salesperson start with?

They are all fairly interchangeable. Grok is the fun one right now for a nontechnical person because it makes it simple to create small bots, give each one a task like managing your inbox, and train them by talking to them. ChatGPT and Claude both have free plans and are the easiest place to start a first conversation.

What should I do first if I have never used AI?

Get a pen and paper and write down what you do every day. Open ChatGPT or Claude, paste the list in, and ask what it can help you with. Then pick one task, get it right, take that off your plate, and move to the second. People get overwhelmed because they try to hand over everything at once.

Will AI replace sales jobs?

It depends on your runway. If you have five or ten years left and deep relationships, those relationships will carry you. If you have ten plus years left and you do not adapt, I do not think you make it. As buyers rely on their own AI operators, the buying decision stops being purely human, and the company whose operator responds first is already in the conversation while you are still busy.

Watch or listen to the whole conversation

The full episode is on The Best Sellers Podcast, hosted by Doug Dvorak.