Small Team, Big Output: Rethinking Marketing Capacity in the AI Era

Slingshot is not a marketing agency. We build software, augment teams, and consult on AI strategy. So why are we writing about marketing capacity?

Because our own marketing department happens to be one of the more interesting AI case studies in the building. A two-person team running blog, social, email, events, sales, and contract generation, all while sitting inside a company that lives and breathes AI development.

The tools we build for clients, the experiments our developers run, the workflows our consultants design, all of it ends up guiding how the marketing team actually operates day to day. That cross-pollination is rare, and its lessons are worth sharing.

So this is a peek behind the curtain. Less ‘here is how to do marketing’ and more ‘here is what happens when a small marketing team has full access to the AI brain trust of a software company.’ The takeaways apply well beyond marketing, but the story starts there.

From Asking AI to Building With It

Last November, AI looked like an assistant. Ask a question, get an answer, move on. That stage is over for teams that take this seriously.

“What’s changed is that we used to use AI to do tasks faster. ‘Do this for me,’ ‘do that for me.’ But now we’re building systems that have our brand, our voice, and our institutional knowledge baked in,” said Whitney Powell, Sales and Marketing Coordinator at Slingshot. “AI starts every job already knowing what we do.”

The shift sounds subtle; it’s not. Treating AI as an assistant means every output starts from zero context and ends with heavy editing. Treating AI as an operations layer means the brand voice, formatting rules, audience nuances, and company knowledge live within the tool from the first keystroke. The output lands closer to finished, and the team spends its time on judgment, not janitorial cleanup.

That’s the difference between marketing teams that look efficient on paper and marketing teams that’re actually compounding.

The Headcount Math Has Changed

Capacity questions used to feel straightforward: more output meant more people. Larger marketing teams break their work into siloed roles: an email manager, an event manager, a content manager, a social manager.

So a two-person team running email, events, social, sales, contract, and blogs shouldn’t, by traditional math, be possible.

“If we were trying to do the same output we do now (one blog a week, multiple social posts a week, an email newsletter, all the sales efforts, and event planning for the Louisville AI Exchange), I would expect it would take at minimum four people, if not more, without AI,” said Savannah Cherry, Director of Marketing and New Business at Slingshot.

Four roles compressed into two. That ratio gets thrown around a lot in AI conversations, often without much under it. What makes it authentic here is that the compression is not magic. It runs on a few specific things: domain expertise that predates the AI tools, custom systems built on top of those tools, and an inclination to keep refining both.

Drop AI into a team without that foundation and the math doesn’t work. Which is exactly the trap most leaders walk into.

Expertise Is the Multiplier, Not the Subscription

There’s a tempting version of the AI capacity story that goes like this: buy the seats, hand them out, watch productivity jump. Spoiler alert: It doesn’t work that way.

“You can’t just throw one person an Anthropic membership and expect them to know what they’re doing,” Whitney said.

The reason is simple: AI accelerates whatever judgment already exists. It doesn’t create that judgment. A marketer who has spent years writing brand voice, building email funnels, and running events can use AI to do four jobs at once because they already know what good looks like in each lane. They can prompt with specificity, review output with a trained eye, and catch what is off-brand before it ships.

“It’s important that as you’re building out a team that’s AI-enabled, they have an understanding of what the proper output should be, and then giving them the tools,” Savannah explained. “Just giving one marketing person AI doesn’t mean that they can do every single area of marketing effectively. It means they can do more within the areas they already have expertise in.”

Leaders need to understand that AI amplifies existing foundations. Strong fundamentals lead to stronger results, while shaky ones become apparent more quickly. The key takeaway is that investing in your team’s fundamentals is crucial, as AI will magnify what’s already present.

The Real Output Is the System, Not the Asset

The clearest sign a team has moved past the assistant phase is what they build, not what they produce. A campaign is an output. A campaign generator is a system. The first ships once. The second ships every week, gets better over time, and works whether or not the person who built it is in the room.

There are several examples of what that looks like in practice at Slingshot. Whitney has built two of the team’s most-used systems. The first is a social caption generator. “I’ve built an artifact that’s an interactive tool where you can pick the post type and the platform, and it produces a caption in our exact structure and voice,” she said. “It’s reusable over time. You can edit it, update it.”

The second is a design system loaded into Claude. “I dropped our entire design system into a Claude skill: fonts, colors, components, the whole brand kit,” Whitney said. “What used to be a full day of building out a 50-slide deck is now about an hour. The system already knows what Slingshot’s brand looks like; I’m directing it, not designing from scratch every time.”

Other marketing tech examples the team has built include a sales call question tracker that pulls themes and follow-up items from transcripts in real time, and a contract generator that handles the first pass on standard agreements. None of them are flashy; all of them claw back hours that used to disappear into repetitive work.

Savannah touched on weaving AI into the rhythm of the workday. “AI routines are where it really clicked for me,” she said. “I start my day with a briefing that shows me what’s on my plate, what came in overnight, what I said I’d follow through on. I’m not spending the first hour of the day figuring out where to start. That’s an hour back, every single day, just from one tool.”

The common thread across Whitney’s and Savannah’s systems is that they’re reusable frameworks, not prompts. While generic prompts lead to routine responses and require constant re-explanation, systems keep their rules, making them useful assets. AI tools can speed up the process of changing individual expertise into shared knowledge, provided teams recognize their purpose.

What AI Still Cannot Do

For every system worth building, there’s a part of the work that should stay slow on purpose. Judgment and strategy still belong to humans, and arguably more than they did a year ago.

“Relationships are still the most important part of sales, probably more important now than ever, because of the overwhelming amount of content AI generates,” Savannah said. “Anything that has to do with trust, making judgment calls, and reading a room; those are all things AI physically can’t do.”

The flood of computer-generated content has raised the value of anything that feels unmistakably human. Outreach that reads like a template lands in the trash. Outreach that shows a real understanding of the recipient stands out more than it would have just a bit ago.

The same applies to strategy. “AI is great at the execution part once we know what we’re doing, but the ‘what-are-we-even-trying-to-do’ part needs to stay with humans,” Whitney said. 

The team decides who to call, what is worth saying, and where the company is going. AI fleshes it out, drafts the email, and builds the deck. The direction is still a human call.

Leaders who blur that line, automating the judgment work along with the execution work, tend to find out the hard way which is which.

What Leaders Should Actually Be Asking

The most common question executives ask about small marketing teams sounds reasonable and is almost entirely wrong. It usually goes like this: ‘Should we hire another marketer, or invest in AI tools instead?’

“That’s the wrong question,” Whitney said. “The right question is, do I have someone who can build the systems, not just the tools? The real question is whether your people are encoding their knowledge into tools, or just typing faster prompts. The leverage isn’t the AI; it’s the person turning their expertise into something the whole team can run on.”

Reframed that way, the choice stops being hire-versus-tool and becomes something more interesting. Where is the team spending repeatable, pattern-based manual hours? What expertise lives in one person’s head that should live in a system the whole company can use? And if we want to move into an area no one on the team actually knows, are we hiring in that expertise, or just piling more work on the people we already have?

“It’s less of a question of hire vs AI. You should be asking how you can invest in the people you already have via AI,” Savannah said. “Give your team the tools to be able to do more, so that then you as a company can grow, instead of cutting people to stay stagnant or save on your bottom line.”

That’s the work. Not picking a tool, not approving a budget line, but identifying the patterns inside the team and figuring out which ones deserve a system built around them.

The Punchline

Small teams with AI aren’t impressive because they replace bigger teams. They’re impressive because of what they free their people up to do. The hours that used to evaporate into repetitive work are now going into strategy, relationships, and building the next system. That’s growth, not compression.

The leverage was never the AI seat. It’s the expertise behind it, the systems built on top of it, and the leaders willing to invest in both. 

The companies that win the next few years won’t be the ones with the most subscriptions or the leanest org charts. They’ll be the ones whose people stopped typing prompts and started building machines, and whose leaders gave them the room to do it.

Slingshot