How a two-person ad shop chases big-agency results

BaD Mktg says it reached $1.5 million in revenue in its second year with a 38% profit margin and is on track for $2 million this year, helped by in-house AI tools that cut turnaround times from six weeks to six days. The two-person agency uses AI for strategy, content, video storyboards and internal operations while keeping senior staff directly involved. The article highlights AI as a productivity lever for small businesses, but the market impact is limited and primarily anecdotal.

Analysis

The signal here is not that AI is helping one small agency work faster; it is that AI is collapsing the minimum efficient scale of high-end professional services. That is structurally negative for the labor-arbitrage model that underpins large consultancies and agency networks, because senior oversight can now be paired with software leverage rather than junior headcount. The first-order winner is any firm able to convert domain expertise into proprietary workflows; the second-order winner is software infrastructure that sits inside those workflows rather than generic model access.

For ACN, the read-through is mixed and more about mix than demand. The market often assumes AI spend flows cleanly into large transformation budgets, but this kind of adoption pushes value toward smaller implementation stacks and away from labor-heavy advisory hours, which can pressure utilization and pricing power at the margin. Over 6-18 months, the risk is that clients use AI to shorten discovery and content cycles, compressing billable scope before they expand budgets into larger enterprise programs.

The contrarian point is that the real monetization may be in governance, integration, and workflow orchestration, not in model generation itself. That favors vendors that can embed into calendars, communications, identity, document systems, and approval chains — the “system of work” layer — while punishing firms whose differentiation is mostly human bandwidth. If the current AI enthusiasm is too broad, the underappreciated trade is to fade labor-intensive service intermediaries and lean into software or platform names that make AI operational rather than aspirational.

Tail risk: if client trust breaks down around AI-generated output quality or confidentiality, adoption slows quickly and the productivity gains get capped. But if that trust issue is solved, the margin expansion story in small teams scales far faster than most public-market service businesses are priced for, making the dislocation most relevant over the next 12-24 months rather than next quarter.

Key Decisions for Investors

• Reduce exposure to labor-heavy consulting and agency proxies versus software-enablement names over the next 3-6 months; prioritize any short against firms with weak utilization sensitivity and high wage leverage.
• Long ACN only on pullbacks if management can show AI-led delivery is expanding mix into recurring workflow/integration revenue; otherwise consider a tactical underweight for 6-12 months as pricing power gets compressed.
• Pair trade: long software/workflow enablers (MSFT or NOW) / short labor-arbitrage services exposure (ACN) for a 6-12 month horizon; thesis is margin capture shifts to the platform layer, not the services layer.
• Buy upside exposure in infrastructure names tied to enterprise AI deployment rather than generative-content tools; use 9-12 month calls where implied vol is still below realized adoption-driven dispersion.
• Watch for enterprise procurement language around confidentiality, human-in-the-loop approval, and workflow integration; if those budgets accelerate, the winner set broadens from model providers to platform vendors within 2 quarters.