Marketing has been consolidating for years, and AI is accelerating the default. Gartner’s latest figure makes that visible: 73% of marketing leaders now report their function is fully or mostly centralized, up from 67% the previous year.
The finding may sound like a straightforward win for centralization, but the reality is messier. AI does not simply reward scale—it also lowers the cost of local execution. That is why the centralization conversation has returned with a twist.
The logic for centralization is straightforward. AI performs best on shared foundations: common data, consistent infrastructure and clear governance. Specialist AI talent is scarce and expensive, so building that expertise once and sharing it across brands and regions often makes sense. Centralization also promises fewer duplicate tools, lower costs and stronger brand consistency.
AI also gives local teams new muscles
The same technology lets smaller teams create and adapt content, analyze information and handle work that used to require a central team or an agency. That autonomy helps markets respond faster to customers, competitors and regulation.
That shifts the calculation for marketing leaders. A capability that once sat at the center because it required specialist help may now be deliverable by a small market team with the right AI tools. But those tools still need shared data and guardrails to stay consistent and safe.
Push too far in either direction and the costs appear. Too centralized means slower decisions and weaker local ownership. Too decentralized means duplicated capabilities, fragmented data and inconsistent practices.
Decide capability by capability
Gartner’s counsel is not to choose one model for everything. Data, technology, creative, media, marketing operations and market execution can sit in different places.
- Centralize capabilities where consistency, governance, scarce expertise or shared infrastructure carry the most weight.
- Keep local capabilities where customer needs, regulations, channels or competitive conditions vary widely by market.
- Write down decision rights. Hybrid models fail when it is unclear who owns a capability and how much freedom local teams have.
For many organizations, the result is a hybrid: enterprise-level data, technology and specialist expertise, with business units retaining control of execution.
In practice, that often looks like a two-speed design—slower shared assets at the top and faster market execution at the edges. The failure point is rarely the structure itself; it is ambiguity. When teams cannot answer “who owns this?” or “how far can we go?”, coordination costs eat into the flexibility a hybrid model was meant to create.
Revisit the design before it hardens
Before pulling a capability into the center, ask what greater centralization would accomplish—eliminating duplicate technology, improving access to expertise, strengthening governance or cutting costs. Then ask what gets sacrificed: speed, local knowledge or business-unit ownership of commercial outcomes.
Gartner recommends reviewing the capability mix at least twice a year, and again when business strategy, market conditions, regulation, technology or AI maturity shifts. Measure performance, cost, responsiveness and risk. A review is not a signal to reorganize every six months; it is a disciplined check that the current placement still fits the evidence.
Keeping the inherited structure is fine—if leaders know why it works and what would make them rethink it. The goal is to put each capability where it works best, not to make the org chart permanent.
Source: MarTech
