Marketing & GTM — Crys Black (Brand, Content & Voice)

Republished with permission from the FrUn SMB AI Guide. Canonical chapter: Superhuman Docs → · All chapters

Crys Black · Fractional CMO & AI Strategist · crysblack.com · LinkedIn

1. Mindset — What an SMB Should Believe Before Spending a Dollar on AI

The first time a small business owner told me she’d been “meaning to figure out AI” for eight months, I watched her shoulders climb up around her ears. She wasn’t behind because she was lazy. She was behind because every article she read made AI sound like a second full-time hire she couldn’t afford to manage.

So let me give you the one belief to get right, because if you get it wrong it will cost you both time and money.

AI is not a marketer you hire. AI is a tool that makes you a faster marketer.

That distinction matters. When owners treat AI like an employee, they hand it the whole job, walk away, and then publish something generic, slightly wrong, and clearly not written by a human who knows their business. When owners treat AI like a power tool, they stay in the driver’s seat. They get a rough draft in thirty seconds and spend their time making it good. Same hour. Wildly different result.

Here’s where AI actually helps in marketing for a small business:

• Beating the blank page. First drafts of emails, social posts, web copy, and answers to the same customer questions you answer every week.
• Repurposing. Turning one thing you already made (a sales call, a blog post, a customer FAQ) into a week of content.
• Making sense of your numbers. Drop a spreadsheet of your sales, email results, or web traffic into an LLM and ask what’s working and what’s not. You get a plain-English read on your own data, no dashboard and no analyst required.
• Showing up consistently. The thing most SMBs fail at is not quality. It’s frequency. AI makes a weekly rhythm possible for a team of one.

And here’s where it just feels productive:

Tinkering. Generating forty logo options nobody asked for or rewriting the same headline eleven times. If you’re spending more time prompting than you’d have spent just doing the task, the tool is now a hobby, not a help.

What has to be true inside your business before AI pays off here? Three things:

• You know who you serve. Not “everyone.” The specific person who buys from you, in their words.
• You know how you sound. Friendly? Expert? A little irreverent? AI will average you into beige unless you tell it who you are.
• You have a few real examples handy. A past email that worked, a review a customer left, the way you actually describe your product out loud. AI is a mirror. Give it something good to reflect.

Get those three down on one page and a $20 tool will outperform a $2,000 one in the hands of someone who skipped them.

2. The SMB Tool Stack — Start Cheap, Stay Simple

You do not need an “AI marketing platform.” The best AI tool for you is usually the one you already pay for. Start there.

Here’s what most SMBs miss in 2026: the software you already use is growing AI features fast. Your CRM, HubSpot or Salesforce, has AI built in now. Canva writes copy and generates images. And the big AI assistants are integrating directly with these tools, which makes the stuff you already own easier to use than ever. Before you buy anything new, go turn on the AI features in the tools you’ve already got. That’s free, and it’s sitting right there.

Then, the short list I’d start with:

• An LLM of your choice, about $20/month. Claude, ChatGPT, or Gemini. This is your writing partner, researcher, brainstorm buddy, and first-draft machine. Pick one. Try the free version of each for a week and keep the one whose writing sounds most like you. This single tool replaces the blank page, the freelance blog post you kept putting off, and the hour you lost staring at an empty email.
• Your CRM, ideally free to start. If you already have HubSpot or Salesforce, use it and switch on its AI. If you have nothing yet, HubSpot Free gives you a real CRM plus email marketing (up to 2,000 sends/month), forms, and landing pages, free up to 1,000 contacts. It replaces the spreadsheet you’ve been using as a customer list.
• Canva for design, about $15/month or free to start. Templates for social, flyers, one-pagers, and simple ads, now with AI built in for images, backgrounds, and copy. Load your logo, colors, and fonts into Brand Kit once and everything you make after looks like you. It replaces the freelance designer for your everyday stuff.
• GenSpark for slides and images. When I need a presentation or polished visuals built fast, this is what I reach for. Add it once slide decks or strong visuals become a regular part of your marketing.

How I work: I use Claude Cowork because it connects to my other systems and the files on my desktop, so the AI works where my work already lives instead of in a separate tab. That’s where this is all heading: AI that meets your existing tools and files instead of replacing them. The more your tools talk to each other, the less you do by hand.

Tools to skip or postpone (for now):

• Big SEO suites (Semrush, Ahrefs, and similar), $100 to $200+/month. Overkill until you’re publishing real volume and have someone to act on the data. Postpone.
• “Autonomous AI marketer” or AI SDR tools that promise to run your outbound or content on autopilot. They need clean data, careful setup, and a human babysitting them. For a five-person shop, the autopilot crashes the plane. Skip until you have someone whose job is to run it.
• Anything that requires a dedicated admin, an annual enterprise contract, or a “book a demo” to even see pricing. If you can’t sign up and start in an afternoon, it’s not an SMB starter tool yet.

If you have zero budget:

Use the free tier of ChatGPT, Claude, or Gemini, plus Canva’s free plan, plus a Google Doc holding your one-page brand brief. That’s a complete marketing setup for $0. Get results there first, and upgrade only when you hit a wall you can name out loud.

3. The First 30 Days — Where to Start

The single highest-ROI thing you can set up first is not a campaign. It’s your brand brief: a one-page document that teaches AI who you are. It does two jobs at once. Every draft starts sounding like you instead of like a robot that skimmed your website once, and it turns a generic chatbot into a real thought partner that can actually strategize with you.

Build that first. Then use it to do everything else faster. Here’s the four-week order of operations. Treat it as a checklist.

Week 1: Build your foundation

• Open one document. Write down who you serve (the specific buyer), what you sell, the five questions customers ask you most, and three words for how you want to sound.
• Paste in two or three real examples: an email that worked, a great review, how you describe your business out loud.
• Give it a permanent home. Load all of that into a reusable container: a Custom GPT, a Gem, a Claude Project, or a Skill. Now your AI remembers your brand every time instead of starting cold.
• Turn it into a thought partner, not just a writing tool. Treat it like a co-CMO, even a co-CEO. Hand it a real challenge out loud (the voice feature is great for this) and have an actual back-and-forth: “Here’s my situation and my goal. What campaign would you run?” You’ll surface ideas for campaigns and programs you simply hadn’t had before.
• Pair that with deep research on where your ICP and personas actually spend time and consume content. A partner that knows your brand, plus real research on your buyer, is where the genuinely new ideas come from.

Week 2: Pick one repeatable output

• Choose a single recurring thing: a weekly email, one social post every other day, or a monthly customer update. One. Not all of them.
• Take something you already have (a past sales call, an FAQ, a blog post) and ask your AI, using your brand brief, to turn it into two weeks of that output.
• Set up Canva: load your logo, colors, and fonts into Brand Kit, and pick one template you’ll reuse every time.

Week 3: Put it on a schedule and map your calendar

• Use your assistant’s schedule feature. Most of the big AI tools now let you schedule a recurring task. Set yours to draft that weekly or monthly output automatically: “every Monday, draft this week’s newsletter from my latest updates.”
• Watch the first couple of runs and refine. If the output isn’t quite right, adjust the prompt or the GPT behind it until it is. This is your tuning week. A scheduled task is only as good as the instructions feeding it.
• Build a simple content calendar. Map what goes out regularly, social, blog, email, and when. The calendar is the plan; the schedule is the engine that fills it.

Week 4: Organize and review

• Inventory your AI tools. List the Skills, Custom GPTs, and Gems you’ve built, what each one does, and the date you last updated it. Keep that list somewhere you’ll find it. Now you know what you have, and when each piece is due for a refresh.
• Let AI read your numbers. Export your results (email opens, clicks, web traffic, sales) into a spreadsheet, drop it into your AI, and ask: what’s working, what’s not, and what should I do more of? You get a plain-English read without building a single dashboard. Strip out customer names first.
• Write your repeatable weekly routine in five lines or fewer. That routine is the actual product of this month.

What “good” looks like at the end of 30 days:

You have a reusable brand brief and a thought partner that knows your business. One regular output is going out on a schedule instead of relying on you to remember. You’ve got a simple content calendar, and your AI tools are documented so you can actually find and maintain them. And you can point to real hours you got back. You’ll feel it most in one specific way: marketing stopped being the thing you keep meaning to do.

4. A Real SMB Use Case

The business: a small B2B marketing agency. Lean team, serving high-trust, high-consideration service businesses, the kind of clients who hire an agency because every piece of content has to sound exactly right. Their growth depended on two things: winning and onboarding new clients well and delivering consistent quality without burning the team out.

What was broken: their AI use was ad hoc. Everyone prompted their own way, so content quality and brand voice drifted from person to person and client to client. New-client discovery research depended on who ran it and how much time they had that week. Reporting was scattered and hard to turn into a clear story for clients. And the handoffs, discovery to brief to content, ran mostly on gut feel. Good people, inconsistent system.

What we did: we started with an assessment, interviewing the team about the friction, then built a connected set of custom GPTs inside the ChatGPT account they already had. Each one standardized a step. A Deep Research GPT paired with ChatGPT’s Deep Research feature so client discovery produced the same structured, source-backed insight every time, no matter who ran it. Per-client content and SEO GPTs trained on each client’s brand voice. A reporting GPT that takes a monthly analytics export and turns it into a plain-English read on what’s working and what to fix. Plus a shared prompt library, brief and SOW templates, and an Asana task playbook so work handed off cleanly into their project tracker.

What AI did vs. what the team did: AI made the knowledge consistent and the documents repeatable. The team kept the strategy, the relationships, the judgment, and the final quality check on everything that went out. The system did the grunt work so the humans could do the human work.

The tools and the cost: mostly tools they already paid for. ChatGPT for the custom GPTs and Deep Research, working alongside the Asana, Harvest, and Slack already in their stack, with Lovable for quick prototypes. The investment was an upfront assessment plus a lightweight monthly retainer, roughly ten hours a month of my time, not a big platform purchase.

The result: discovery research got consistent and faster, so the team walked into early client conversations clearly better prepared. Content held each client’s voice without anyone reinventing the prompt. Reporting became repeatable and easier to act on. And cleaner briefs and SOWs meant less rework on the handoff into delivery. Same team, mostly the same tools, a much more consistent machine.

And you don’t need a whole ecosystem to start. Pick one: a single research GPT you reuse before every discovery call, or one reporting GPT you feed your monthly numbers. Build that, get it solid, then add the next. Consistency is the whole trick.

5. Guardrails — What an SMB Should NOT Do

Here are some mistakes that quietly cost you trust, and trust is the only marketing asset a small business can’t rebuy.

The mistakes I see most:

• Verify before you publish. Always. This is the big one, so I’m putting it first. AI writes confidently even when it’s flat wrong. It will invent a statistic, the wrong hours, a guarantee you never made, or a feature you don’t offer. Read every word, and fact-check anything with a number, a name, a price, or a promise in it before it goes out. AI hands you a draft, never a verified fact.
• Don’t let AI invent proof. This is verification’s evil twin. Never let it write a customer testimonial, a review, or a result you didn’t actually get. Fake testimonials aren’t just a bad look, the FTC treats AI-generated or fabricated endorsements as illegal. Only publish words a real customer actually said.
• Chasing tools instead of finishing one workflow. Ten half-set-up tools help you less than one workflow you run every week. Pick the workflow. Ignore the shiny thing.
• Letting your voice drift to generic. Left alone, AI sands every business down to the same smooth, forgettable corporate tone. Your voice is your edge. Keep editing until it sounds like a person, specifically you.

Keep humans on these, no matter how good the tools get:

• Replies to real customers, especially complaints and refunds. A bot answering an upset customer turns one bad day into a screenshot.
• Reviews and reputation. Respond as yourself.
• Any claim about pricing, safety, health, money, or legal terms. If it’s regulated or it’s a promise, a human owns it.
• The final yes before anything publishes. Always.

Privacy and data issues SMBs underestimate:

• Customer data in consumer AI tools. Don’t paste customer lists, emails, payment details, or anything personal into a free chatbot without checking its data settings. On consumer tiers, your inputs can be used to train the model. Turn that setting off, or use a business tier, before you share anything private.
• “Free” images with strings attached. AI-generated images can resemble real brands or people, and stock-looking images sometimes carry usage rights you don’t have. For anything customer-facing, use Photoshop’s or Canva’s licensed library or images you made or own.
• Disclosure where it counts. Don’t let AI generate customer-facing claims about your product, price, or results without a human who can stand behind them reviewing first. And if AI writes something that reads as a real person’s words, be honest that it did.

6. Lessons Learned + One Thing You’d Tell Every SMB Owner

Five things I’ve learned using AI in my own marketing work, written so you can steal them:

• The foundation beats the tool. A clear one-page brief makes a $20 tool brilliant. No brief makes a $2,000 tool useless. Spend your first hour on the brief, not on shopping.
• The fast bad draft is the gift. AI gets you to a rough draft in seconds so you can spend your time making it good. The draft was never the deliverable. Your judgment is.
• Consistency beats brilliance. A pretty good email every week will out-market a brilliant one every quarter. AI’s real superpower for a small team is helping you show up on schedule.
• Name your voice or lose it. I got tired of editing the generic out of every draft, so I built an actual brand voice Skill that captures how I really write: my rhythm, my hard rules (no em-dashes, ever), the kinds of stories and references I reach for. Now the tools I use sound like me on the first try instead of beige corporate AI. You don’t need to build a Skill to start. Write your voice down in a paragraph and feed it to your AI every time. But the moment it’s worth it, make it reusable. Your voice is your edge, and AI will quietly sand it off every chance it gets.
• Start with one workflow, not ten tools. Get one thing working end to end and running weekly before you add anything. Momentum compounds. Tool sprawl doesn’t.

The one thing I’d tell every SMB owner reading this:

You are still the marketer. AI is Yoda, not the hero. The point isn’t to sound like everyone else using these tools. It’s to show up consistently as the one thing no competitor can copy: you. Build the foundation, pick one workflow, keep your hand on the wheel, and let the robot do the typing.

Now go send the email you’ve been meaning to send for eight months. I dare you.

Part of the FrUn SMB AI Guide. Read on canonical guide →