Short answerAutomation is the better first choice for rule-based, repetitive work (reminders, invoicing, data transfer, scheduling confirmations), while a virtual assistant is better for judgment-based work at low volume (inbox triage, research, customer conversations, exception handling). The costs differ by an order of magnitude — $50–$300 a month for an automation stack versus $800–$2,500 a month for a part-time VA — so the highest-return sequence is: automate the rules first, then hire a VA for whatever judgment work remains.
About the numbers: Unless a named source is linked, ranges and dollar examples are planning illustrations—not industry benchmarks, client results, or promises. Measure your own volume, time, pricing, and adoption before making a decision.

What to do next

The comparison owners actually face isn't abstract — it's a pile of overflow work and a decision about where to throw it. So sort the pile with one question per task: does this task ever require a decision I couldn't write down in advance? Chasing invoices: no decision — the rules are fully writable (remind at these intervals, escalate at 30 days). That's automation, and paying a human $20 an hour to execute writable rules is paying a premium for worse reliability, because humans doing robot work get bored, and bored humans make errors that software doesn't. Handling a confused customer email: real decisions — tone, judgment, when to escalate. That's human work, and forcing it into automation produces the canned replies customers rightly hate.

There's a hidden cost asymmetry worth naming. Automation's costs are front-loaded: setup thinking, configuration, a few weeks of tuning — then it runs at near-zero marginal cost forever, at any volume. A VA's costs are permanent and linear: every hour of work is an hour billed, forever, and doubles when volume doubles. There's also management overhead owners forget to price — instructing, checking, re-explaining — which also consumes some of the hours a VA saves. None of this makes VAs a bad buy; it makes them a specific buy: judgment work, moderate volume, tasks that vary too much to script.

The expensive mistake, made in both directions: hiring a VA to do robot work (paying $1,500 a month for what $50 of software does more reliably), or forcing judgment work into brittle automations that mishandle your customers. The businesses that get this right run both, layered — software executes the rules, the VA handles the exceptions the software flags, and the owner only sees what genuinely needs an owner. Getting the layer boundaries right requires knowing which of your tasks are actually rule-based versus judgment-based — a distinction owners consistently misjudge about their own work until someone walks through it with them. That sorting is a core output of the FrictionList audit: every friction point tagged automate, delegate, or keep, with the costs of each path side by side.

Quick answers

When is a virtual assistant better than automation?

For judgment work at low-to-moderate volume: inbox triage, research, scheduling with negotiation, customer conversations, and handling exceptions automations can't.

How much cheaper is automation than a VA?

Typically 5–10x for equivalent rule-based workloads: $50–$300/month in software versus $800–$2,500/month for a part-time VA.

Can I use both together?

That's the optimal setup: automation runs the rules, the VA handles flagged exceptions, and the owner handles only what genuinely requires an owner.