🎧 Prefer to listen?

I was paying a virtual assistant $500 a month to do things I could now automate in an afternoon. That’s not a hypothetical — it’s what was actually leaving my bank account every month for tasks like drafting client emails, pulling research, organizing data, and following up with leads. When I added it up, I felt sick.

I’m not anti-VA. Good ones are worth every penny. But I was paying someone to do repetitive, pattern-based work that AI tools handle faster and more consistently. The kind of work where you’re essentially paying for someone’s time rather than their judgment. If you’re spending money on a VA mostly for tasks that follow a template, this post might save you a few hundred dollars a month.

I tested a bunch of tools over three months. Most of them were either too complicated, too limited, or too expensive for what I needed. Three stuck. Here’s exactly what they are, what they replaced, and what the math looks like. If you’re curious about building your first automation before diving in, start there — this post assumes you’re comfortable with the basics.

The tasks I was actually paying for

Before I could replace my VA, I needed to understand what I was paying for. So I tracked every task for two weeks. Here’s what came back:

Email drafting and responses. About 8 hours a month. Client check-ins, proposal follow-ups, answering the same five questions every new lead asks. My VA had templates, but she still had to personalize each one.

Research and competitive analysis. About 6 hours a month. “Find me three competitors in this space,” “What’s the average price for this service,” “Summarize this 40-page report.” Time-consuming but completely mechanical.

Data entry and organization. About 5 hours a month. Moving info from forms to spreadsheets, updating CRM records, organizing invoices. Soul-crushing work that nobody should be paying $25/hour for.

Social media scheduling. About 4 hours a month. Formatting posts, finding hashtags, scheduling across platforms. Not creative work — logistics.

Follow-ups and reminders. About 3 hours a month. “Did the client reply? If not, send the second follow-up.” Pure pattern matching.

That’s roughly 26 hours a month at $20-25/hour. The math was clear: I was paying $500+ for work that didn’t require human judgment, creativity, or relationship skills. The tasks that DID require those things — strategy calls, creative direction, actual relationship building — I was already doing myself.

Tool 1: ChatGPT Plus ($20/month) — the email and research machine

ChatGPT Plus replaced about 14 hours of VA work per month. Here’s how.

Email drafting. I created custom GPTs for my most common email types. Client follow-up? I paste the last message thread and say “draft a friendly follow-up referencing our last call about X.” It writes something better than my VA’s templates because it has the full context. Proposal follow-up? Same thing. New lead response? I built a GPT that knows my services and pricing, so it drafts responses that actually sound like me. I review and send — takes 30 seconds instead of waiting 4 hours.

Research. This is where ChatGPT really shines. “Find me the top 5 project management tools for freelancers under $30/month — include pricing and one pro/con each.” That used to be a half-day task for my VA. ChatGPT does it in 60 seconds. Is it perfect? No. I still verify key facts. But the 80% of research that’s “gather and organize information” — done.

Document summarization. My VA used to spend hours reading reports and pulling key points. Now I upload the PDF to ChatGPT and ask for a summary with action items. Three minutes instead of three hours.

The trick is building custom GPTs for your specific workflows. Generic ChatGPT is useful. ChatGPT trained on your business context is a replacement for someone who’s been working with you for six months.

Tool 2: Make.com ($16/month) — the automation backbone

Make.com replaced about 8 hours of VA work per month, and it works while I sleep. If you’ve read my post on automating client follow-ups, you already know the basics. But I’ve built way more since then.

Lead follow-up sequences. When someone fills out my contact form, Make.com automatically sends a personalized response, adds them to my CRM, and schedules three follow-up emails over two weeks. If they reply, the sequence stops. My VA used to manage this manually — tracking who got which email, remembering to follow up, updating the spreadsheet. Now it just happens.

Invoice and payment tracking. Make.com watches my Stripe account. When a payment comes in, it updates my spreadsheet, sends a confirmation email, and adds the client to my onboarding sequence. When a payment fails, it sends a gentle “hey, your card didn’t go through” email automatically.

Social media formatting. I still write my own posts (AI-written social content sounds like AI-written social content). But Make.com handles the formatting, resizing, and scheduling across platforms. I drop content in one place, it goes everywhere.

The free tier of Make.com handles 1,000 operations per month, which is enough to test. The $16/month Core plan gives you 10,000 operations — more than enough for a solo business. My VA spent 8 hours a month on the tasks Make.com now handles in the background. And Make.com doesn’t forget, doesn’t take holidays, and doesn’t need me to explain the same process twice.

Tool 3: Claude Pro ($20/month) — the thinking partner

This one surprised me. I got Claude Pro initially for writing, but it ended up replacing the more analytical parts of my VA’s work — about 4 hours per month.

Report analysis. My VA used to pull data from multiple sources and compile reports. Now I upload spreadsheets to Claude and ask it to find patterns, flag anomalies, or summarize trends. It’s faster and catches things she missed.

Proposal writing. For complex proposals, I’d give my VA bullet points and she’d flesh them out. Claude does this better because it can reference previous proposals I’ve uploaded and maintain consistency across documents.

Competitive deep dives. ChatGPT handles quick research. Claude handles the “read this 50-page industry report and tell me what matters for my business” kind of analysis. It’s better at nuanced, long-form reasoning.

The two tools complement each other. ChatGPT is faster for quick tasks; Claude is better when I need to think through something complex. Having both costs $40/month total. My VA cost $500/month for the same analytical work, minus the speed.

The actual cost breakdown

Let me be honest about the numbers:

Before (VA): $500/month for ~26 hours of work.

After (AI tools):

  • ChatGPT Plus: $20/month
  • Make.com Core: $16/month
  • Claude Pro: $20/month
  • Total: $56/month

That’s a $444/month savings, or $5,328 per year. But there are real caveats.

Setup time. It took me about 15 hours over two weeks to build all the automations, create custom GPTs, and learn the tools. That’s a one-time investment that pays back in the first month.

What I still do myself (or hire out differently). Tasks requiring genuine human relationships — sales calls, partnership conversations, creative direction — I do myself. Tasks requiring visual design — I hire a designer per project instead of monthly. The AI tools handle the repeatable 80%, and I focus on the 20% that actually needs a human.

What I lost. My VA caught things AI doesn’t — like “this client seems off, maybe don’t push for the upsell.” That kind of emotional intelligence isn’t in the tools yet. I’ve had to build that awareness myself. For solopreneurs who need customer message handling, there’s a middle ground worth exploring.

What I’d do differently

If I were starting over, I wouldn’t replace everything at once. I’d pick the task that costs the most time per dollar — for me, that was email — and automate that first. Get comfortable with one tool before adding the second. The overwhelm of switching everything simultaneously almost made me quit and go back to the VA.

Also: keep your VA for two weeks after you think you’ve automated everything. You’ll discover tasks you forgot about. I found three things I hadn’t accounted for — two of which I automated in Make.com within an hour, and one I decided to just stop doing entirely because it wasn’t actually producing results.

The bottom line

Replacing a VA with AI tools isn’t about being cheap. It’s about being honest about what you’re paying for. If your VA is doing work that follows patterns, templates, and rules — that work can be automated for a fraction of the cost. Save your human budget for things that actually require a human.

Start with one tool. Build one automation. Save yourself 4 hours this month. Then decide if you want to keep going. If you’re not sure which tool fits your needs, check out the AI Tool Advisor — it’ll point you in the right direction based on what you’re actually trying to do.