Singtel is using its latest creative agency review to redesign how brand, engagement and loyalty work gets done. The Singapore telecommunications company is considering traditional networks, independent shops, embedded agency providers, AI-first production partners and technology-enabled specialists for a model that brings external capability closer to its internal team.
The tender is notable because AI is not framed as a separate service line. It sits inside planning, production, personalization, training and customer experience, with governance and human oversight built into the brief. That makes the review less about choosing a supplier and more about choosing an operating model for an AI-enabled marketing function.
Table of contents
Jump to each section:
- Why Singtel is redesigning the agency relationship
- AI changes where agency value sits
- The embedded model trades handoffs for context
- What marketers should know about AI-enabled agency design
Why Singtel is redesigning the agency relationship
The conventional agency process is built around a sequence. A client briefs, an agency interprets, production begins, review follows, and revisions travel back through the chain. Singtel is challenging that structure by asking prospective partners to work alongside its brand, engagement and loyalty team and help reshape how work is planned, produced and scaled.
The scope connects creative and strategic planning with content production, AI-supported workflows, capability building and coordination across internal and external teams. It also calls for a combination of specialist expertise and Singtel’s own institutional knowledge. The partner is expected to contribute to campaigns and customer experiences while helping the internal team become more capable.
This is where the review becomes strategically interesting. AI can compress routine production, but speed alone does not remove the need for customer context, brand judgment or organizational coordination. If anything, faster output makes weak handoffs more visible because teams can generate work more quickly than they can align on what the work should accomplish.
When automation shortens production, context becomes the scarce resource.
AI changes where agency value sits
Singtel’s brief suggests that routine production will carry less of the value in the relationship as AI handles more repetitive work. The harder contribution shifts toward connecting business priorities, creative judgment, customer data and technical execution. That is why the tender is open to a wider range of partner types than a conventional creative review might attract.
The common assumption is that AI makes outside agencies less necessary because internal teams can produce more themselves. The contrasting reality is that automation creates new integration work across data, systems, approvals and brand standards. The strategic implication is that agencies may remain essential, but the basis of their value changes from supplying capacity to designing and governing capability.
Singtel is making that change visible in the responsibilities it has set out. Prospective partners must support AI-enabled ways of working, strengthen creative capacity and help streamline campaigns while preserving human oversight. Training is part of the assignment, which means success cannot be measured only by what the partner delivers. It must also include what the client team learns to operate.
The most durable agency relationship may be the one designed to reduce dependency on the agency.
That sounds contradictory, but it reflects a more mature model. A partner that transfers methods, documents workflows and builds internal judgment can become more valuable over time because it helps the client tackle harder problems. The relationship earns its place through continued capability creation, not through keeping routine knowledge outside the organization.
The embedded model trades handoffs for context
An embedded structure can reduce the distance between insight and execution. Working inside the brand, engagement and loyalty team gives specialists faster access to business context, customer needs and the people responsible for approvals. It can also make creative, technology and data expertise part of the same conversation before ideas harden into deliverables.
For personalization, this proximity matters. Singtel already uses AI in customer service and applies richer data intelligence to shape more relevant interactions. Extending that approach into marketing requires more than generating variants. Teams need a shared understanding of which customer signals are appropriate to use, how a recommendation becomes a message, and where human judgment must interrupt automation.
Embedding a partner does not automatically solve those questions. It can simply move unclear ownership closer to the client. The model works only when decision rights are explicit: who can approve a workflow, who owns the customer outcome, who evaluates the quality of an AI output, and who acts when the system produces something off-brand or mistimed.
Proximity improves collaboration only when accountability travels with it.
This is why governance belongs inside the creative operating model rather than beside it. In an AI-enabled workflow, a creative decision may also be a data decision, a customer-experience decision and a technology decision. Treating governance as a final review step would preserve the sequential structure Singtel is trying to replace.
What marketers should know about AI-enabled agency design
Singtel’s review gives marketing leaders a useful way to evaluate partner models as AI moves from isolated tools into the daily flow of campaign and customer work.
Define the capability being built. A brief should state what the internal team will be able to plan, operate or evaluate after the engagement, not only which assets the partner will produce.
Reward context, not just speed. Faster production has limited value when customer knowledge, brand standards and business priorities remain fragmented. Partner access to the right context becomes part of performance.
Make training part of delivery. Singtel includes capability building in the scope, treating knowledge transfer as real work rather than an informal benefit at the end of a project.
Keep human authority specific. Human oversight is meaningful only when teams know which decisions require approval, which risks trigger intervention and who is responsible for the outcome.
The wider shift is not simply that agencies are adopting AI. Brand organizations are reconsidering what they need an agency to be. Production capacity still matters, but it is increasingly joined by workflow design, technical fluency, governance and the ability to make internal teams stronger.
That changes how marketers should run reviews. Credentials and creative reputation remain relevant, but they do not reveal whether a partner can work inside a connected operating model. Leaders also need to examine how a prospective agency handles data context, transfers knowledge, documents decisions and collaborates across functions.
Singtel’s tender does not settle which structure will win. It does clarify the question now facing many marketing teams: not who can use AI, but who can help the organization use it without losing judgment, accountability or customer understanding.
The agency model will not be transformed by faster output alone. It will be transformed by who learns to connect speed with responsibility.