Which AI Tools Actually Run a Solo Marketing Business?

What does a real AI-powered solo marketing workflow actually look like?

Most "AI tools for marketers" lists read like vendor catalogs. This one doesn't. Lilach Bullock runs a solo marketing consultancy and has published a candid breakdown of which tools she actually uses, what each one replaced, and where each one falls short. The takeaway for any SMB operator: the gains are real, but they come from replacing specific tasks, not from transforming your whole business overnight.

Which tools replaced which parts of the workflow?

Bullock's stack isn't exotic. The tools she lands on are ones most operators have already heard of. What's useful is the specificity of what each one replaced.

| Tool | What it replaced | Honest tradeoff | |---|---|---| | ChatGPT / Claude | First-draft writing, brainstorm sessions | Output still needs a real editorial pass | | Perplexity AI | Manual research and source-hunting | Can miss nuance; verify anything critical | | Notion AI | Scattered notes, meeting summaries | Only as useful as your Notion hygiene | | Canva AI | Basic graphic requests to a designer | Not a substitute for brand-level design work | | Zapier / Make | Manual handoffs between tools | Requires setup time upfront |

The pattern here is consistent with what we see across client engagements: AI earns its place when it takes over a defined, repeatable task, not when it's asked to "help with marketing" in some vague sense.

What did AI actually save in time and money?

Bullock doesn't throw out a dramatic headline number, which is exactly why her account is credible. The savings she describes are in hours per week on drafting and research, not in headcount elimination or overnight revenue jumps.

That matches the broader data. McKinsey's 2023 State of AI report found that marketing and sales functions see the most measurable productivity lift from generative AI, with respondents reporting meaningful time savings on content-related tasks. But the same report noted that fewer than one in three companies had moved past early experimentation into consistent deployment.

The gap between "I tried it" and "it's actually in my workflow" is where most solo operators and small teams are stuck right now.

Where did AI tools fall short?

This is the part most tool-review posts skip. Bullock is direct about the limits:

• Content quality requires human editing. AI drafts get her to 60–70% of a finished piece. The rest is judgment, voice, and knowing what a specific client actually needs to hear.
• Research still needs verification. Perplexity is faster than a Google rabbit hole, but she checks anything that matters before it goes to a client.
• Automation setup has a real time cost. The Zapier and Make workflows that now save her hours took hours to build and test.
• Brand consistency is still a manual job. Canva AI produces usable assets, but keeping everything on-brand across a client portfolio is still a human task.

The tools that stuck are the ones that made a specific task faster, not the ones that promised to replace a whole function.

This is the clearest signal for any SMB evaluating AI tools: if a vendor is promising to replace a function, be skeptical. If they can show you how they speed up a specific task you do repeatedly, that's worth testing.

How should a solo operator or small team decide what to automate first?

Bullock's implicit framework, and the one we use with agency clients, comes down to three questions:

1. What tasks do you do repeatedly that follow a pattern? Drafting similar emails, summarizing calls, pulling weekly metrics, formatting reports. These are the candidates.
2. What's the cost of a mistake in that task? Low-stakes first drafts are safe. Client-facing deliverables with legal or financial content are not.
3. What would you do with the recovered time? If the answer is "more billable work" or "better client relationships," the ROI is obvious. If the answer is unclear, that's a signal to slow down.

A solo consultant recovering five hours a week on content drafting and research isn't a small thing. At a $150 per hour effective rate, that's $750 per week in capacity. Compounded across a year, the math on a $20–$30 per month AI tool subscription is not complicated.

Is there a single AI tool that does everything a marketer needs?

No, and anyone selling that product is overselling it. The operators Bullock describes, and the ones we work with, are running stacks of three to five targeted tools, each doing one thing well. The integration layer (usually Zapier or Make) is what ties them together into something that feels seamless.

The all-in-one promise is appealing but consistently disappoints in practice. A purpose-built research tool beats a general-purpose chatbot for research. A purpose-built image tool beats a general-purpose platform for visuals. The stack wins over the suite.

What we'd actually do

• Audit before you buy. List every task you do more than twice a week. Mark which ones follow a repeatable pattern. That list is your automation backlog, in priority order.
• Test one tool for 30 days against a specific task. Not "use AI more." Replace one specific task. Measure the time difference. If it's not meaningfully faster by day 30, move on.
• If you want the shortcut, come to the community. We've mapped tools to SMB workflows across dozens of industries at skool.com/aiforbusiness. You don't have to run the experiments solo.