Breaking: A 27-year-old Atlanta agency makes itself the test client 40% of internal workflows automated, Creaxion says Human-controlled guidance becomes the product

Company profile / Atlanta / Brand innovation

Creaxion Found AI Hiding in Its Own Name

After 27 years of selling attention, the Atlanta boutique turned itself into the first case study for its AI-era reinvention. The clever part is not the machine. It is the deliberately small roster of humans standing between the machine and the public.

In the summer of 2025, Mark Pettit announced that his 27-year-old marketing firm was changing course. The slogan was pleasingly compact: “Don’t panic. Pivot.” The agency would combine artificial intelligence with the sort of judgment earned from product launches, reputation emergencies, executive positioning, and the occasional phone call no client ever wants to make. Then Creaxion did something smarter than writing a manifesto. It made Creaxion the first client.

The company says the experiment took less than 60 days, automated 40 percent of its internal workflows, and tripled its content output. Its current site makes an even tighter claim: 30 days from assessment to market for the repositioning itself. These are company-reported figures, not an audited productivity study. Still, the sequence is worth noticing. The firm did not begin by asking the public to admire a chatbot. It began by using new tools to change its own work.

27Years before the AI repositioning
40%Internal workflows automated, company-reported
Content output in under 60 days, company-reported

The pivot before the pivot

Creaxion was never a technology startup waiting for its model release. Pettit founded it in Atlanta in 1998 after a career that moved through television journalism and corporate communications. The agency learned the classic trades: brand identity, public relations, media strategy, advertising, events, sponsorships, and crisis work. Its announced client history includes AT&T, Arby’s, Coca-Cola, Marina Bay Sands, Porsche, Pizza Hut, and KFC. In 2015, the company said its first 17 years had generated billions of media impressions.

Creaxion founder and CEO Mark Pettit
MARK, WITH A PLAN. Founder Mark Pettit has been a reporter, agency chief, author, actor, and producer. Apparently one lane felt needlessly restful.

But the decisive move came before the current AI rush. In 2021, Pettit said he had converted Creaxion from a full-service agency into a specialist brand consultancy taking only three clients at a time, partly so he could give more attention to acting and producing. That sounds like a lifestyle choice. It was also a sharp business constraint. A firm that promises senior judgment cannot behave like a content mill. Fewer accounts makes the promise believable.

“Our experience is the part AI can’t replicate.”Creaxion’s current positioning

Four years later, the constraint gave the AI pivot a shape. Creaxion is not trying to become a software company. It sells customized projects and retainers, keeps the roster deliberately short, and wraps general-purpose AI capabilities in a service layer it calls the Pivot Lab. There is no public rate card. The trade is explicit: fewer logos, more attention.

Six old agency jobs, one faster engine

The Pivot Lab begins with research, but it ends in familiar places. A 10-point scorecard examines identity, visibility, and readiness. The agency then maps 30-, 60-, and 90-day actions. Around that framework sit six practices, each aimed at a recognizable problem: a stale position, an executive nobody hears from, a content operation that cannot keep pace, a visual identity that no longer fits, a campaign missing its audience, or a reputation in danger.

The difference from a conventional agency is speed around the craft. AI can scan trends, organize research, propose variations, and remove repetition from campaign production. The difference from a do-it-yourself AI subscription is accountability. Creaxion’s phrase is “human-controlled guidance.” Every recommendation, it says, passes through an experienced person before it reaches an audience.

That positioning serves companies and executives in the awkward middle: large enough to have a reputation worth protecting, but not necessarily staffed with a full internal brand studio. It also suits organizations in transition. The offer is less “make me an ad” than “tell me what has become dated, show me a better position, and help my team move without embarrassing itself.”

The client that stayed

Agencies adore launch announcements because beginnings photograph well. Longevity is duller and more persuasive. Creaxion began working with payment processor Segpay in 2011. In 2014, Segpay signed a five-year extension and credited Creaxion with brand identity and strategic guidance, saying its business had grown 40 percent since January 2012. Financial terms were not released. In 2025, the companies signed another three-year extension focused on media relations, executive thought leadership, brand development, and AI guidance.

1998

Creaxion opens in Atlanta as a marketing and communications company.

2011

The Segpay relationship begins.

2021

The full-service agency narrows itself to a three-client consultancy.

2025

Pivot Lab launches and Segpay extends again.

2026

Creaxion develops identity, video, and media relations for Assisted Response.

The relationship reveals the business model better than the AI vocabulary does. Creaxion sells continuity through change. For Segpay that has included repositioning, launches, press material, executive visibility, and an owned news channel. For newer client Assisted Response, Creaxion says it developed the identity, produced an explainer video, and took on media relations while Brady Mills Agency helped build the website. The deliverables change. The role stays close to the decision-maker.

A guardrail with history behind it

“Human-controlled” also carries a less comfortable meaning. In 2019, the Federal Trade Commission finalized a consent order settling allegations that Creaxion, founder Mark Pettit, a publisher, and its principal had presented paid endorsements as independent consumer opinion and advertising as editorial content during a mosquito-repellent campaign tied to Zika concerns before the 2016 Olympics. Creaxion and Pettit neither admitted nor denied the allegations, except as necessary to establish jurisdiction.

The first thing to fail: disclosure

The order required clearer identification of paid relationships, monitoring of endorsers, and records demonstrating compliance. In an era of synthetic media and automated persuasion, those are not antique rules. They are the operating manual.

The incident matters because Creaxion now sells judgment around a technology that can blur authorship at industrial speed. A human in the loop is useful only if that human asks who made the claim, who paid for it, whether the audience can tell, and who will take responsibility. Oversight is not decorative. It is the product’s most valuable part.

The part worth copying

Creaxion’s play is available to any experienced small firm, even if its name contains no conveniently discoverable initials. The useful idea is to convert history into a filter for new tools. Do not begin with every possible application. Begin with the work your team repeats, the decisions clients already trust you to make, and the places where a mistake would be expensive.

01

Audit the drag

Map the slow research, approval, and production steps before choosing tools.

02

Be your own test account

Change your positioning and workflow first. Keep the receipts and the awkward lessons.

03

Package a path

Turn advice into a scorecard and a 30/60/90-day sequence clients can understand.

04

Name the responsible human

Review claims, provenance, disclosure, tone, and reputational risk before publishing.

What changed Pettit’s mind was not one dramatic model demo. The public trail shows two stages: first, a decision to stop being everything to everyone; then months of AI training, outside guidance, internal workshops, and testing. Creaxion retrained experienced communications staff for AI brand strategy instead of presenting automation as a substitute for them. The resulting pitch is conservative in a useful way. The machine widens the funnel. A small group of people narrows the answer.

The useful limit

This model is a poor match for buyers who want cheap self-serve software, endless high-volume output, or execution with no senior review. It also weakens when a firm has no deep craft to place above the tools. AI can accelerate a judgment system; it cannot quietly supply the missing judgment.

There is a nice joke buried in the reinvention: Creaxion says AI has been in its name from day one. The typography may be new, but the more serious point is old. A company can survive several eras if it knows which part of itself should move quickly and which part should refuse. Creaxion has chosen speed for the machinery and friction for the final decision. For a business built on public words, that is a sensible place to put the brake.