The Mac mini arrived last night. I sat with it on the kitchen table this morning and didn't realise that I have no mouse. No keyboard either. Fifteen years a laptop guy and I have never owned either. The mini is going to be dedicated to an enterprise client, so fifteen minutes of setup and I am done with the mouse forever, so I refuse to buy a usb or bluetooth mouse for something that I will use for one time, and then it sits in a draw and gets lost in our sea of clutter.
So I texted my neighbour. She's a venture backed tech CEO herself, much bigger than me, running a company of at least 100 employees. She must have something like this lying around. Better her than me, and I'll use it for an hour and give back to her later today.
While I was waiting for her reply I sat down to write today's post and realised: today is the day I start automating marketing with ai for my own business. The post forces it. Sixty days of writing about ai ops and I have not run a single piece of cold outreach yet.
A small detail I would not have arrived at without journaling: the Mac mini was not really the problem. The borrowed mouse was today's first small win because I let myself slow down enough to think clearly about it. Who is close. Who would have a mouse. Victoire is close, maybe Victoire. Then Rachel's name came to mind and I knew. I sent the text.
While I am here, an honest note about the daily writing itself. It is starting to feel yucky. I am drifting into ai slop. I need to come back to a human voice. The post from two days ago was about exactly that. Today the post forces me to do the work I keep saying I will do.
The thing about the BIP writing with an ai co-author is that you get better and better at seeing what ai wants to do for editing. Now, when I scroll through my LI feed it is very easy to spot the ai slop, and it's like 80 to 90% of my feed. At apparently it works, because it's concise, tidy and a lot of lead gen bait with sometimes hundreds of comments. Comment x in the post, and I'll give you y. I've filled a few of these out, and the person never even reaches out to me. Man, I'm starting to despise LI, but as a marketer, sales person it feel necessary.
Today is prospecting. Today, seriously, I am starting.
Two prospecting bets I am running in parallel this week.
Bet 1: Apollo plus an openclaw fork
The first is openclaw forked and pointed at Apollo for outbound. I say cold, but the outreach will be signal-based. Companies that are already asking for help in areas where I can help them. AI ops, fractional CAIO work, the kind of thing I have been writing about for sixty days.
I only need to reach out to fifteen to twenty per day. The plan is to make each one as tailored as possible using ai tools to read the signal and draft the first touch in my voice. My voice bible is detailed, it's something I've been working on daily, so that part is easy. I do not want to scale ugly. I want to scale the version of cold outreach where the prospect reads the first sentence and thinks, "this person actually read about us, or he's very good at ai which is something we want"
The ai employee build is progressing. I hired the person doing the build a couple of weeks ago and I like her work. She could become an asset on the fulfilment side too, when Jan and I are too busy. Jan cannot come on full time yet because the monthly revenue does not justify it. The ai employee bridges that gap.
Bet 2: Meta ads direct to a WhatsApp ai agent
The second bet needs more setup. Possibly today.
Last summer Jan and I were building a WhatsApp ai SaaS. We shut it down. It was a lot of work for a market that was changing under us. But for a few months I was running Meta ads direct to an ai agent on WhatsApp, and the lead conversations and qualifying were the best I have ever seen. Since then the models have gotten a lot better. The agent I would build today reads warmer, qualifies tighter, and never misses a follow-up.
I started setting up my Meta account again this week and hit blockers on verification. I have heard the same from marketing friends. Meta is dealing with a wave of bot accounts and has tightened the rules. I am not a bot. I still cannot get fully verified. I can get verified enough to run lead-gen ads, which is the part I actually need.
The ads I have in mind will be headline-only. I have not worked out the copy yet. The math: twenty to twenty-five dollars a day in Meta spend is about seven hundred dollars over thirty days. For a $4,995/month package, one conversion covers that spend seven times over. That is the bet. One conversion in thirty days from a WhatsApp ai agent I built, control, and tune daily.
The buyer side of this
If you are running a small or mid-size business and you have looked at ai and felt the gap between "we should be doing more" and "we do not know where to start," the WhatsApp agent I am building will be where you can talk to me without a calendar invite. Headline-only ad, click straight into chat. Same conversation I would have on a call, no warm-up. Likely, I will be putting this on my website as well. If you want to test something that I can build in 1 day, probably a few hours, you have it.
The goal for the month is sales conversations. Not signed clients yet. Conversations. Apollo plus openclaw is the outbound side. Meta plus WhatsApp is the inbound side. Both are running this week.
On top of that I am still building the ai controller sandbox runs for an enterprise engagement, and the MCP connectors for a research assistant build I am starting today. I am not stopping any of the build work. I am adding the prospecting layer that the build work has been telling me to add for sixty days.
I hope Rachel texts me back. I will pick it up on the morning walk with the dogs and the Mac mini will be live before lunch, I hope.
Day 61 of 365.
Monthly Revenues $9,200 | Clients 2 | Prospects, starting today, I promise!