TD Bank Group Cites $370M Cumulative AI Value Across 75 Use Cases

TD Bank Group's enterprise AI portfolio is a company-wide program of roughly 75 production use cases across customer acquisition, customer insights and risk management that, according to group president and CEO Raymond Chun, generated $170 million in value in fiscal 2025 and is targeted to add $200 million in incremental value in fiscal 2026.

Executive Summary

TD Bank Group said its fiscal 2025 AI portfolio delivered $170 million in value from approximately 75 use cases, and its fiscal 2026 pipeline is expected to generate $200 million in incremental value. Group president and CEO Raymond Chun disclosed the figures Dec. 4 on the bank's Q4 earnings call, saying the program spans customer acquisition, customer insights and risk management.

The bank's disclosures also include a production deployment of a generative AI virtual assistant in its U.S. contact center. TD Bank told CIO Dive the assistant, built on a Knowledge Management System, handles questions from agents to reduce client hold time on calls, resolve issues faster and lower escalations to supervisors.

For enterprise buyers, TD's numbers matter because they come from a regulated financial institution that ties the figure to specific use cases, not from a vendor benchmark. The bank says it is now moving from discrete use cases toward end-to-end process reimagination, a shift that executives at other large banks are likely to watch closely.

What does TD Bank's $170 million AI value figure actually measure?

The figure is a management-reported aggregate, not a GAAP financial metric, and TD has not published a formula for how each use case's value is calculated. What the Q4 transcript makes clear is that the value is attributed to use cases that shipped in fiscal 2025, including transformed loan underwriting, intelligent lead generation and client relationship deepening.

Chun described the portfolio as concrete plans that are already delivering clear outcomes. The bank's investor presentation, published alongside the earnings call, states that in fiscal 2025 TD implemented about 75 use cases generating $170 million in value and expects fiscal 2026 use cases to generate $200 million in value.

The $370 million cumulative figure is an expectation, not a guarantee. The 2026 number is framed by TD as expected and incremental, meaning it is intended to stack on top of the 2025 result rather than replace it.

How did TD Bank get from 75 use cases to a $200 million target?

TD's 2026 target is not simply more of the same. Chun told investors the coming year would include use cases that reimagine end-to-end processes, which suggests the bank is shifting from point automation to workflow-level redesign. The 2026 pipeline is organized around the same three priority categories: customer acquisition, customer insights and risk management.

The progression is consistent with how enterprise AI programs typically mature: early deployments prove value in isolated functions, then the data and governance infrastructure built for those pilots is reused across adjacent processes. TD's fiscal 2025 portfolio already spanned underwriting, leads and relationship deepening, giving the bank a base of production experience to scale from.

The contact center deployment is the clearest example. TD Bank told CIO Dive the Knowledge Management System provides consistent responses to agents, which is designed to reduce client hold time and supervisor escalations. That is a customer-experience use case with operational-cost implications, and it sits alongside risk and acquisition use cases in the same portfolio.

What is the generative AI virtual assistant in TD's U.S. contact center?

The virtual assistant is a generative AI system deployed as a Knowledge Management System in TD's U.S. contact center, according to CIO Dive. It is aimed at contact center agents rather than customers directly: when an agent needs an answer during a live call, the assistant retrieves and synthesizes policy or product information from the knowledge base.

TD Bank told CIO Dive the assistant was added to reduce client hold time on calls and resolve issues faster, while lowering the number of times problems are escalated to supervisors. For a retail bank running large contact center operations, those two metrics are direct drivers of both customer satisfaction and labor cost.

The deployment is notable because it targets agent augmentation rather than full deflection. Instead of replacing the agent with a customer-facing chatbot, TD is using generative AI to make the human agent faster and more accurate, a pattern that is easier to deploy in a regulated environment because a human remains accountable for the interaction.

CIO Dive's report, headlined 'TD Bank US head of AI weighs in on agents,' linked the contact center deployment to TD's broader approach to AI agents. The bank did not disclose usage volumes, latency targets or error rates for the assistant in the research materials.

Where is TD applying AI across the bank?

Metric	Fiscal 2025 actual	Fiscal 2026 target	Source
AI use cases in production	Approximately 75	Not disclosed; includes end-to-end process reimagination	TD Bank Group Q4 2025 earnings call transcript
Value generated	$170 million	$200 million incremental	TD Bank Group Q4 2025 earnings call transcript; Q4 2025 investor presentation
Priority categories	Customer acquisition, customer insights, risk management	Same categories plus end-to-end process reimagination	TD Bank Group Q4 2025 earnings call transcript
U.S. contact center AI	Generative AI virtual assistant added	Assistant in production with stated goals of lower hold times and fewer escalations	CIO Dive

The three priority categories are not equally weighted in TD's disclosures. The bank did not break out the $170 million by category, and it did not say how many of the 75 use cases sit in each bucket. What the disclosures do show is that the portfolio is broad enough to span revenue, cost and risk outcomes.

The underwriting transformation referenced by Chun is particularly significant for risk management. Loan underwriting is a core, high-volume process where AI can compress cycle times and improve consistency, but it is also one of the most tightly regulated. TD's inclusion of underwriting in its 2025 value count suggests the bank has moved beyond experimentation into production in a controlled risk function.

How should other banks benchmark TD's AI disclosures?

TD is one of the few large North American banks to publish a dollar-denominated, use-case-level AI value figure on an earnings call. That creates a benchmark for peer institutions, even though methodologies are unlikely to be directly comparable across banks.

1. Track value by use case rather than by total AI spend. TD's $170 million figure is tied to about 75 deployed use cases, which makes the number auditable internally and gives the bank a per-use-case value profile.
2. Prioritize by business category before selecting technology. TD says it is prioritizing customer acquisition, customer insights and risk management, meaning investment choices follow business outcomes.
3. Look for second-order operational metrics. The contact center assistant's goals, including lower hold time, faster resolution and fewer escalations, are measurable operating metrics that support the broader value figure.
4. Plan for end-to-end reimagination after the pilot phase. TD's 2026 language moves from discrete use cases to end-to-end processes, signaling that the next stage of value comes from workflow redesign.

Executives should also note what TD did not claim. The bank did not say the $170 million figure is audited GAAP revenue or profit; it said the use cases generated value, a management-defined concept. Peer banks that adopt similar disclosures should define their own methodology and state it clearly.

What are the market and stakeholder implications?

TD's disclosures arrive as the banking industry is moving from generative AI pilots to agentic deployments. Banks face pressure from investors to show that AI spending translates into earnings, and TD's decision to put a dollar figure on its portfolio is a direct response to that pressure.

The choice of customer acquisition, customer insights and risk management also signals where TD sees competitive advantage. Customer acquisition and insights are revenue-side, while risk management is a cost-side and safety-side function. Covering both sides of the P&L makes the AI program's value case more resilient to changes in interest rates or loan demand.

For technology vendors and systems integrators, TD's trajectory implies growing demand for governance tooling, model evaluation and process-reengineering services. A bank that says it will reimagine end-to-end processes in fiscal 2026 will need workflow automation, data infrastructure and change management, not just model inference.

What did Raymond Chun tell investors about TD's AI results?

Chun framed the numbers as evidence that TD's AI strategy has moved from aspiration to execution. On the Q4 call, he said the bank implemented about 75 use cases in fiscal 2025 and pointed to specific business functions where the work is concentrated.

As you heard at Investor Day, AI is a massive opportunity for TD, and we have concrete plans that are already delivering clear outcomes. This year, we implemented approximately 75 AI use cases that generated $170 million in value. These use cases span from transforming loan underwriting to creating intelligent leads to deepening relationships and meet more of our client's needs. For next year, we expect AI use cases to generate $200 million in incremental value, including use cases to reimagine end-to-end processes, as described at Investor Day. We are prioritizing our AI investments with use cases focused across categories such as customer acquisition, customer insights and risk management.Raymond Chun, Group President and CEO, TD Bank Group

Chun's phrasing ties the 2025 result to the bank's Investor Day narrative and positions 2026 as the year of process reimagination. For investors, the key sentence is the expectation of $200 million in incremental value, which implies TD believes it has a repeatable mechanism for turning AI use cases into measurable business outcomes.

What's next for TD Bank's AI roadmap?

TD's stated 2026 plan has two layers: more use cases in the existing priority categories, and a new wave of end-to-end process redesigns. The contact center assistant is likely to be a template for similar deployments elsewhere, since the pattern of augmenting employees with a knowledge-grounded assistant can be applied to underwriting, compliance and operations.

The bank's risk management focus will also keep TD close to regulatory expectations. Canadian and U.S. banking regulators have both emphasized model risk management and fair-lending considerations in AI, and TD's decision to frame risk management as a priority category suggests the bank is trying to make AI governance part of the value story rather than a constraint on it.

The Knowledge Management System also points to TD's governance approach. Because the assistant is grounded in an internal knowledge base rather than generating answers from open-ended model weights, it gives the bank a clearer audit trail for contact center decisions, which matters in a regulated industry.

The $200 million target is a management expectation, not a forecast in the accounting sense. Peer institutions and investors should watch TD's fiscal 2026 quarterly disclosures for evidence of how the end-to-end process work translates into the cumulative value figure.

Frequently asked questions

How does TD Bank measure its $170 million AI value figure? TD Bank Group reported that approximately 75 AI use cases implemented in fiscal 2025 generated $170 million in value, according to its Q4 2025 earnings call transcript. The bank has not published a uniform methodology, but CEO Raymond Chun said the use cases span loan underwriting, intelligent leads and relationship deepening.

What is the Knowledge Management System in TD's U.S. contact center? It is a generative AI virtual assistant that answers questions from contact center agents, providing consistent responses to reduce client hold time and resolve issues faster while lowering escalations to supervisors, TD Bank told CIO Dive.

Which AI categories is TD prioritizing? TD says it is prioritizing customer acquisition, customer insights and risk management, with fiscal 2026 plans that include reimagining end-to-end processes, according to the Q4 2025 earnings call transcript.

Does TD's $200 million 2026 target include the $170 million from 2025? No. Chun described the $200 million as incremental value expected in fiscal 2026, in addition to the value generated in fiscal 2025.