On Wednesday morning I set a hard date: prospecting kicked off before the weekend. It is Friday. No leads are coming in. So I am going out in public and calling out the prospecting plan I wrote about earlier this week. Two channels. One is an openclaw fork pointed at Apollo for outbound. The other is Meta ads into a WhatsApp ai agent built to qualify leads before I ever get on a call.
The openclaw fork runs as an ai employee: daily Apollo prospecting against companies showing signals they need help, with a daily report back to me. There is a possible junior developer who could come on to run it if she works out. Jan checked her background and approves. He would manage her, since this sits on the tech side of the business. Until he is full time, it is really me managing her in between everything else to get things moving. No idea how it turns out, but learning how openclaw works is half the reason I am running this channel at all.
Channel two: Meta into WhatsApp.
The year Jan and I already lost on this
Jan and I spent about a year building a WhatsApp and Instagram ai agent. It did not work out. We were getting plenty of leads, but they were not serious, and chasing leads at $99 a month is not worth the time. I would get four or five sales calls a day and then spend the day chasing them down. It is not worth a senior salesperson getting on calls for $99 leads.
What did impress me was how accurately the WhatsApp agent qualified. With about five or six standard questions, the lead would self-qualify and tell us how ready they were for a call and how serious they were.
The questions that do the qualifying
Five or six questions, and the lead basically tells you how serious they are. The set I keep coming back to:
1. What is the one job in your business you would hand to a machine tomorrow if you could? 2. Have you tried any ai tools yet, or would this be your first real attempt? 3. If we got that one job off your plate, what is that worth to you a month? 4. Are you the one who decides on something like this, or does someone else sign off? 5. Is this a start-this-month thing or a sometime-this-year thing? 6. One to ten, how ready are you to actually hand part of this over?
The lead answers six texts and qualifies themselves. No call until they have already told me they are real.
Why I built my own instead of buying
Based on my experience qualifying leads with a WhatsApp agent, and being able to customise it far more than the clunky legacy software like ManyChat and the others (terrible UI, over-complicated, confusing), I decided to build a simple Cloudflare Worker WhatsApp agent tied directly to my Claude vault. The ai knows everything about me, our offer, my writing style. That is far more powerful than a WhatsApp SaaS that is ten years old and trying to catch up.
I have run LinkedIn ads before and they are terrible. $30 a click terrible, not doable for a solo founder paying ad spend out of pocket. Meta is the bet, even though it is a mess right now and has been for months. Support is terrible, account setup is a problem, my legitimate account was banned for no reason I can see. I am going with them because I have enough confidence in the Meta to WhatsApp qualifying process that I can get $1 to $3 leads. For a $4,995 a month package, that is worth it.
The headlines, and what sixty days taught me
The deepest frustration I have heard, over and over: they know they need ai, but they have no idea what to do or how to start. Worse, no idea who to trust. Picture a domain expert who is not technical at all. They cannot even ask their own IT department, because this stuff is too new and changes daily.
The headlines I am testing come straight off that:
1. "You know you need ai. You have no clue where to start." Says their exact frustration back to them. The most common thing I heard. 2. "It is too new to ask your IT team. So ask someone who does this all day." Names the isolation and puts me where they can actually reach me. 3. "Fractional ai for experts who are not techies." Plain offer. Pulls in the right person, repels the wrong one. 4. "Stop saving ai tools you are never going to set up." The do-it-yourself paralysis they are living in. 5. "Most ai consultants will confuse you. We just fix one thing." Answers the who-do-I-trust problem with simplicity. 6. "An ai employee that does the work, and reports back every day." Concrete outcome, tied straight to the openclaw channel.
The bet
This is all theory until the ads run. The Meta campaigns are built and sitting there, ready to switch on the moment the Twilio WABA approval comes back.
I have spoken with a few founders, not many, because I have been head down on fulfilment. The need is there. One enterprise founder asked what I was doing these days. He had not caught up with me in a while, and I am always working on something new.
My one liner: I am building a fractional ai agency. I help people who know they need ai but do not know where to start.
Hmm. That sounds like it could be basically everyone, he said.
Yes, it kind of does, I answered.
We will see.
Day 64 of 365.
Monthly Revenues $9,200 | Clients 2 | Prospects (Meta to WhatsApp campaign built, waiting on Twilio WABA approval) | Team: Me + Jan (CTO)