Agency
Agency explained
Start with the problem to solve. Do you need a specific capability, additional capacity or a connecting perspective across disciplines? This determines the expertise required and the work that stays internal. Size, reputation or a long client list alone does not establish whether a team suits the specific task.
A good brief describes the reason, audience, starting point, objectives, constraints and open questions. Discussion matters as much as the document. Ask how the proposed team thinks, works and reviews quality. Work examples should make its role, scope and substantiated results clear. Not every project requires an extensive competitive pitch.
Agree concrete deliverables, responsibilities and decision routes. Clarify who supplies information, consolidates feedback and approves versions. A transparent cost framework, rules for scope changes and eventual handover of content, access and agreed usage rights matter too. An inexpensive line item can create substantial follow-on work when scope is unclear.
At Davies Meyer, Creative Engineering connects ideas, design and technical delivery. AI can support steps, simplify variants and create room for better development. Its value must show in usable results: quality, suitable delivery and total effort including review and corrections. We take responsibility for the concept and quality. A blanket savings percentage does not adequately describe this collaboration.
Examples
Hypothetical application
A company seeks support for a new digital service. Joint reveals that it needs offer clarification and an operable enquiry process alongside a website. The team agrees outputs, internal contacts, reviews and handover. An initial bounded implementation shows whether the approach and collaboration fit the task.
Key Points
- Choose expertise and people for the specific task.
- Agree scope, feedback and responsibility clearly.
- Compare quality and total effort alongside individual prices.
Practical application
Describe the task and internal prerequisites. Assess relevant skills through concrete work and conversations. Agree a clear scope with a cost framework, decision routes and handover.
Useful measures
Output quality
Check agreed requirements, usability and factual reliability.
Collaboration
Monitor unresolved decisions, clear feedback and fulfilment of agreements.
Total effort
Include internal work, rework and operation in evaluation.
Common mistakes
- Using prestige or a low hourly rate as the sole selection basis.
- Compensating for unclear requirements with more review rounds.
- Equating AI use with automatically better or cheaper work.
Sources and context
- IPA: Briefing an agency
Public introduction to the relationship between a written brief and discussion.
- WFA & VoxComm: Agency selection principles
Public summary on selection, partnership and transition; the full guide has restricted access.
Frequently Asked Questions about Agency
When a specific task requires external skills, capacity or collaboration. Which work benefits from external support also depends on existing internal capabilities.
No. Depending on the task, conversations, relevant work examples or paid, bounded collaboration may provide a suitable basis. Choose a proportionate selection process.
No. AI may simplify individual steps but also requires setup, review and correction. Evaluate the quality of the usable result and total costs for the agreed scope.
Related links
Loading related terms…
All TermsArticles about Agency

Agentic Marketing: How Autonomous AI Agents Are Transforming Marketing in 2026
Autonomous agents take over research, campaign ops and real-time budget allocation. What agentic marketing means for CMOs, which use cases work today, and how to set up governance and ROI.

In-house vs. Agency: The New Marketing Operating Model in the Age of Agents
The era of AI agents demands a radical rethink of marketing structures. Learn how to build the perfect hybrid operating model to stay agile, scalable, and future-proof.

Meta Ads AI Connectors: Why Your Ad Account Just Became an Agent Interface
With Ads AI Connectors (MCP server + Ads CLI), Meta opens its advertising stack to external AI tools: reporting, campaign management, catalog work and signal diagnostics in natural language. Here is what it means for governance, agency setups and your 2026 marketing operations.