Trust, speed and scale: TD's blueprint for AI-driven marketing
Marketers in financial services are under pressure to deliver experiences that feel seamless and personal. TD Bank Group (TD) is tackling that head-on with TD AI Prism, a predictive foundation model built to anticipate customer needs and shape more relevant engagement across channels.
As TD's VP of growth marketing, Akif Unal frames it simply: AI isn't just about efficiency. It's about making things simpler and faster so teams can get better outcomes for clients and colleagues.
What TD AI Prism actually does
TD AI Prism ingests hundreds of data signals and consolidates many separate propensity models into one system. That shift lets marketing move from broad segments to engagement that feels specific and timely.
"The marketing teams can now look at all the granular predictions from TD AI Prism and further personalize customer engagement with highly relevant communication," says Unal. It's already informing high-value programs like targeted credit card offers and identifying customers in the market for a mortgage.
Beyond immediate offers, the model helps TD understand each client's potential value and needs, guiding both real-time decisions and long-term relationship strategy.
Speed is a growth lever
Unal is clear: precision matters, but speed wins momentum. By moving from many small models to a single predictive system, TD can develop, deploy and govern models faster, cutting manual work and freeing teams to focus on deeper customer relationships.
More insights, shipped faster, create proof points you can scale. That feedback loop compounds learning and lifts performance across programs.
Scaling responsibly starts with trust
Trust is the center of TD's approach. The bank has expanded its patent portfolio to more than 2,500, including over 800 related to AI-signaling sustained work in this space.
"We've been building in a way that is responsible, that is trustworthy and transparent," says Unal. TD applies generative AI through its Trustworthy AI Framework, requiring each use case to meet principles of safety, fairness and transparency, and to align with client expectations.
If you're formalizing your own guidelines, the NIST AI Risk Management Framework is a helpful reference for governance and risk controls.
What marketers can steal from this playbook
- Unify models where it makes sense. A single foundation with flexible predictions reduces technical debt and speeds deployment.
- Personalize beyond static segments. Use granular predictions to time offers, channels and creative to individual needs.
- Build a tight test-and-learn loop. Track uplift by audience and moment, then redeploy learnings across programs quickly.
- Simplify governance. Standardize features, documentation and approvals so compliance moves at the pace of experiments.
- Balance now and later. Optimize for immediate response and long-term value with shared metrics across marketing and product.
- Reallocate resources with intent. Put more budget and talent behind models and messages that prove lift.
- Measure speed, not just accuracy. Add time-to-decision, model coverage and opt-out rates to your scorecard.
- Lead with transparency. Publish plain-language explanations, provide easy opt-outs and test for bias routinely.
- Upskill your team on AI ops. Make model literacy, prompt fluency and experiment design part of everyone's toolkit.
The bar TD is setting
TD has set a target to generate $1 billion in value from AI. With TD AI Prism now embedded in marketing and more AI initiatives underway, the strategy blends speed, responsibility and creativity-anchored in trust and focused on outcomes that feel human and simple.
Next steps for marketing leaders
- Inventory your existing models and identify consolidation opportunities.
- Define a minimal, practical governance workflow tied to business risk.
- Stand up a weekly rhythm of experiments powered by predictive signals.
- Publish a clear, public-facing statement on how you use AI in marketing.
If you're building skills across your team, explore focused training for marketers: AI Certification for Marketing Specialists.
Your membership also unlocks:
