The Real Bottleneck Was Never the Software

I'm a software engineer. A good one. When a client hands me a broken codebase, a half-built product, or a blank slate and a deadline, I know exactly what to do. That part has never been the problem.

The problem was everything else.

Not because I didn't know how to do it. I know how to send an invoice. I know how to follow up on a contract. I know that a newsletter takes thirty minutes to set up and that a CRM with no contacts is useless. I knew all of this. I just didn't do it. Or did it badly. Or did it late.

Every time I had to stop building software and switch to something administrative — chasing a signed contract, logging a call, updating a deal stage, drafting a follow-up email — the cognitive switching cost was enormous. Not because those tasks are hard. Because they're not the thing I'm wired to do. My brain wanted to be solving an engineering problem, and sending an invoice felt like putting down a scalpel to sweep the floor.

So the floor went unswept. For months at a time.

HubSpot sat mostly empty. A newsletter I'd been "about to start" for the better part of a year never launched. Invoices went out late. Contracts were tracked in my head. Sales conversations died after the first meeting because I hadn't built any system to keep them alive.

This wasn't laziness. It was a structural problem. I was running a business that needed at least three people — an engineer, a business developer, and an operations person — and I was one person only really operating as one of those three.

What If You Could Just Describe What You Need?

A few weeks ago I started using Claude — Anthropic's AI — in a way I hadn't before. Not as a chatbot or a writing assistant, but as an operational layer connected directly to the tools I actually run my business with: my CRM, project management, email, document storage, newsletter platform, contract signing, and prospecting.

The shift wasn't that AI did the work for me. The shift was that describing what I needed — the same way I'd brief a project manager or business partner — was enough to make it happen. The activation energy of every non-engineering task dropped to nearly zero.

The best way I can describe what it feels like now: I have a team of six without the overhead of employing anyone.

There's someone who knows my CRM and keeps it clean. Someone who drafts emails in my voice. Someone who knows my clients well enough to catch when a message is slightly off. Someone who manages my project tracking. Someone who runs my newsletter infrastructure. Someone who challenges my assumptions before I make decisions I'd regret.

None of them are people. All of them are conversations with an AI that has access to my actual tools and enough context about my business to operate inside them intelligently.

Here's what that looks like in practice.

Nothing Falls Through the Cracks Anymore

The highest-value change isn't any single tool — it's that the operational stuff that used to evaporate now happens consistently.

Leads stay alive. I used to come home from networking events tired, with a stack of LinkedIn connections and a few numbers in my phone. Within 48 hours most of those conversations were gone — the context, the energy, the specific thing that made the connection worth following up on. Now, the same night, every meaningful contact goes into my CRM with notes about how we met, what they do, and what the next step is. When a follow-up email bounces, a task gets created automatically to follow up by text. Nothing sits in my head.

Twenty-something new contacts in two weeks doesn't sound like a lot in isolation. But multiply that by every event I go to for the next year and you have a compounding pipeline built from conversations that would have otherwise disappeared. The number isn't the point. The consistency is.

Contracts become projects in one step. I had a client engagement with a scope document sitting in Google Drive. Converting that into a structured project used to mean reading through it, deciding what the tickets were, writing them up, and entering them one by one. An hour of work I'd been putting off. Instead: pull the document into the conversation, describe what I need, and a project gets created with properly structured tickets, acceptance criteria, and sprint assignments — directly from the spec. The document and the project management tool aren't siloed anymore. They inform each other through a conversation layer that understands both.

The newsletter actually launched. I'd been planning it for almost a year. I had the topic, the audience, the content categories, even a few posts written. What I didn't have was the will to sit down and configure an email platform. In one session I chose the platform, worked through the tradeoffs, set up subscriber groups and automations, and wired the signup form on my website. When I assumed a platform had a native integration that didn't exist, I got corrected before I built on a false assumption. When I tried to structure something in a way that would have sent unsolicited emails, the logic of why that was wrong got explained clearly. That's different from just executing tasks. That's having someone in the room who pushes back when you're about to make a mistake.

Emails and contracts go out on time. Follow-up emails that I'd been putting off get drafted in context — referencing the actual conversation I had with a lead, calibrated to where they are in their decision process. Contracts get sent with properly formatted, professionally worded envelope messages pulled from the real engagement details. None of this is revolutionary in isolation. What's different is that it happens now, every time, instead of sitting on a to-do list I avoid.

The Connective Tissue Is the Real Story

The value isn't just in having AI help with individual tools. It's in what happens when those tools are connected through a shared context layer.

CRM contacts flow into newsletter groups, which trigger automations, which drive signups, which become qualified leads back in the CRM. Scope documents become project tickets. Deal context informs email drafts. Contract messages reference the right language because the scope document is already in the conversation.

Each tool is more useful because it knows what the others know. The connective tissue between systems that used to require manual orchestration — or more honestly, that I just never got around to — now happens naturally.

What Actually Changed

I'm still doing the engineering. I'm still making the decisions that require taste, accountability, and judgment. The work that only I can do hasn't gone anywhere.

But the operational stuff that used to fall through the cracks doesn't fall through anymore. Because it doesn't require me to context-switch into a different version of myself to do it. It just requires me to describe what I need — the same way I would if I had the capital to hire the team I actually need.

I don't have that capital. But it turns out, for a lot of it, I don't have to.

If you're a solo founder running a business that needs more than one person, reach out. I'd be glad to talk about what this looks like for your setup.