- Key insight: An AI assistant is shortening traders' idea generation time from hours to minutes.
- What's at stake: Firms risk losing client edge if AI-driven research access isn't operationalized securely.
- Forward look: Prepare for agentic AI deployments across global markets in 2026, reshaping workflows.
Source: Bullets generated by AI with editorial review
TD Securities launched an AI virtual assistant in June that lets traders query the bank's own equity research, and the early results look positive.
"It's a massive time save," Dan Bosman, chief information officer at TD Securities, told American Banker. "It's not unusual for people in the capital markets group to receive ten 20-page PDFs simultaneously in their email inbox.
"Even if you're the fastest reader and you're really skim reading, you have to take that first 30 minutes in the day to pore over that before you can make your first call," Bosman said. "Now with the tool, you're able to get those insights and make those calls within minutes, as opposed to waiting half an hour."
Some people will still read the actual reports when they have time, he said.
"The real need is, we put out so much content and folks on a trade floor are constantly bombarded with news and signals," Bosman said. "Part of what we do is try to reduce that signal-to-noise ratio, getting the right signals to our sales traders and ultimately out to our clients as quickly as possible."
In the broader picture, this initiative is one of many at TD Bank Group, whose CEO Raymond Chun has set ambitious goals for AI at the firm.
"Across TD, we're deploying the capabilities needed to drive speed, such as AI-powered virtual assistants, AI-enabled adjudication, predictive tools and new applications," Chun said at a recent investor day. "These new capabilities are already driving strong outcomes. We're approving mortgages in hours instead of days. We're pre-approving credit cards with data-driven insights for millions of clients. We're producing reports in minutes versus hours or days, and we're responding to clients in just a few seconds, significantly shortening call and wait times."
The bank is aiming to get $1 billion in annual value from AI, half through revenue increases and half through cost savings.
TD now has 2,500 data scientists, engineers, data analysts and experts building proprietary platforms and applications.
"This is huge," Chun said. "AI is fast becoming fundamental to business and to client experience."
Elsewhere in TD Securities, Bosman's team has given software developers coding assistants.
"A year ago, folks were saying, we need this, and now seeing it's being woven into their careers and what they're doing," he said.
Many Wall Street firms are looking to get an edge through AI adoption.
"TD made an early move into AI, with its acquisition of Layer 6, that has helped it create value across its front-office workflows," Brad Bailey, research director at Burton-Taylor Consulting, told American Banker. "However, they are not alone. Getting AI and generative AI right is a key initiative across broker dealers and investment banks across the Street. Certainly, the largest brokers and investment banks have invested vast resources into the selection of the right offering for their stakeholders and have made numerous announcements on their generative AI leverage."
Many banks are realizing that a huge amount of knowledge — such as research and client insights — "sits buried in internal systems and is rarely easily accessed when needed," said Sumeet Chabria, CEO of ThoughtLinks. "Generative AI tools like ChatGPT are proving very effective at surfacing this information in a natural, conversational way. Creating an access layer that makes research and data more discoverable and usable is becoming a common goal for banks."
Inside the trading bot
TD Securities chose OpenAI's enterprise version of GPT for its virtual assistant. It was a natural extension of TD's cloud partnership with Microsoft and Azure.
"With Microsoft's partnership with OpenAI, there's an enterprise-grade platform," Bosman said. "We've invested heavily on our data journey and securing our infrastructure. So it is a logical choice."
Bosman's team also liked the data security and privacy built into the platform.
The virtual assistant was co-developed with Bosman's team, TD Securities' sales and trading team and Layer 6, the AI company TD Bank Group acquired in 2018. One thing the Layer 6 team did was coax ChatGPT to speak the language of researchers and traders.
As a result, "it sort of sounds like an equity researcher and it sounds like a salesperson," Bosman said. "And how you prompt it is also very straightforward. We take a lot of time to make sure that we've got the right citations, we've got the right ability to get back into the underlying research."
Once upon a time, TD Securities' technology department worked to optimize the screen real estate on traders' multiple monitors.
"Now with that little white ChatGPT input field, you just enter your prompt and you have the world's knowledge behind it that you can bring forward into something they can understand," Bosman said. "It's quite an effective way to synthesize a lot of data."
The bot can translate text to structured query language, a programming language that can interact with databases, and produce not only written answers to questions, but also structured data in tables, graphs and other visualizations.
"Before, you had to have a business intelligence tool, and a data analyst to do the work," Bosman said. "Now, you can literally just ask for what you want to see, and you get it."
Traders and salespeople use the virtual assistant to ask questions like, why a certain stock is moving up or down. Meeting prep is another strong use case, Bosman said.
"Say I'm meeting with the CEO of Oracle, I can prompt, 'Give me five trade ideas,'" he said.
Some traders start their day with the virtual assistant, asking for five things they need to know as they start their workday, for instance.
"Anywhere we can say this could create an edge for our clients, that's where we're focused," Bosman said.
Another place this gen AI model has been useful is in drawing out macro industry themes.
"As you're starting to look at things like the government shutdown and tariffs, and you want to see, 'Well, what's the sector impact?' And maybe there's a correlation between what's happening in the sector that you cover and those names that you cover, and something else, it'd be really hard to pull that out" manually, Bosman said. "This really helps to glean some of those insights that are probably a little broader than what you would have by yourself."
"We're always trying to give an edge to our team, to our clients," Bosman said. "So this is about trying to get insights out, both to our sales and trading staff as well as to our clients. It's something a human can't keep up with."
TD uses retrieval augmented generation to limit the data sources for ChatGPT to the bank's own research, including daily reports, and market data.
"One thing we've been really clear on is making sure this thing is not hallucinating," Bosman said. "It is accessing this corpus of knowledge and we're validating it."
While research content is not as real time as live market data, "the key risk is overreliance on AI summaries without verification," Chabria said. "The best implementations build in controls within new AI enabled processes so traders can validate key facts and assumptions, and some firms are adding automated cross-checks to confirm the accuracy of what's generated. These are essential safeguards, including the ability to trace generated information back to its source."
Next steps
Today, about 1,000 people are using the virtual assistant at TD Securities. Bosman plans to deploy it across TD's entire global markets business over the next month. Other business lines in the bank are interested in using it, too.
The virtual assistant has been "one of the most impactful" AI projects because "the group that we're working with is all falling in love with the product," Bosman said. "We've gotten a lot of traction from it."
"I'm hoping 2026 is the year of agentic," he said. "I'm hopeful we'll start to see even more material value being delivered in these solutions, both those that we build internally and those that we partner on."
Agentic AI will be used to reimagine processes to make them faster.
"I think there's immense value to be unlocked with operations and tech looking at how we can reimagine processes using this technology," Bosman said.
