Connecting Client Engagement to Commercial Outcomes: Singletrack’s AI-First Approach

Capital markets firms generate enormous volumes of information through their interactions with clients, yet much of that data remains fragmented across research, sales and trading, corporate access and investment banking. Bringing those activities together offers opportunities to improve client engagement, identify commercial opportunities and gain a clearer understanding of the value delivered by different parts of the business.

TradingTech Insight spoke to Brijesh Malkan, CEO of Singletrack, about the company’s approach to embedding AI into front-office workflows, the challenges of establishing reliable data foundations and the potential for more proactive, agentic capabilities.

TTI: Singletrack describes itself as an AI-first rather than an AI-enabled platform. What does that distinction mean in practice, and where is AI embedded most deeply within the day-to-day workflows of capital markets users?

BM: At its heart, being AI-first means embedding AI across the entire platform as a core capability, rather than bolting AI features on top of existing functionality. We’ve been incorporating AI techniques into our product since 2017, so we have extensive experience of integrating AI capabilities directly into our users’ day-to-day workflows.

In practical terms, that means 3 things. First, identifying key workflows that can be handed over to AI, freeing users from low-value activity so they can focus on high-value work. Second, using large datasets to generate suggested next actions that would otherwise be missed. Third, providing insights that demonstrate proven ROI, so our clients can understand more deeply the value of the services they provide.

For sales and trading, AI automates the capture of revenue-generating interactions, rather than requiring users to log them manually, and surfaces “Guided Next Action” suggestions. For research, it provides insight into report composition, suggesting changes to readability, the use of text versus charts and subjectivity to help improve engagement. For corporate access, it automatically produces a suggested target list for an upcoming event, based on indicators of high relevance such as an individual’s interests, previous activity and consumption patterns.

TTI: What do you see as the biggest challenges in creating sufficiently complete, accurate and contextual client engagement data across research, sales and trading, corporate access and banking?

BM: The primary challenge is overcoming the legacy siloed architecture across capital markets desks. Research readership, corporate access roadshows, trade flows and banking deal pitches remain trapped in disconnected systems, rather than forming a unified 360-degree view of the client.

Manual data capture creates further friction and severe gaps in completeness. Time-poor front-office professionals prioritise high-touch client execution over CRM entry, which results in patchy, delayed or lost records of billable activity.

Even when activity is captured, standard CRM tools record basic metrics (that a touchpoint occurred) but miss the nuance, depth and client intent needed to distinguish a quick execution confirmation from a high-value, bespoke analyst interaction or C-suite meeting.

Strict regulatory compliance and information barriers pose a constant hurdle. Aggregating engagement metrics safely, without exposing sensitive deal data or breaching ethical walls, requires domain-specific governance and granular entitlement models.

Finally, transforming raw, high-volume interaction logs into commercial value remains difficult without specialised AI that can clean, contextualise and link engagement data directly to trade revenue, research monetisation and broker vote outcomes.

TTI: When automatically capturing interactions and enriching them with tags and metadata, how does the technology determine the relevant context from emails and meetings, and what controls are in place to ensure that the resulting data is accurate enough to support downstream decisions?

BM: Our “one-click” interaction capture uses AI and machine learning techniques to identify and extract entities from email bodies and meeting descriptions. These are then applied to the interaction as structured tags and metadata, so sales teams don’t have to enter this information manually. The capability has logged more than 25,000 interactions, saving our clients roughly 835 hours.

High-quality data is what keeps our AI capabilities effective and relevant, so accuracy is critical. That’s why we enforce a human-in-the-loop approach that allows users to verify the details of each interaction before it’s captured and stored.

TTI: Your “One Bank” model allows you to connect research consumption with subsequent engagement and trading activity. How do you establish whether a piece of research genuinely influenced a client’s behaviour rather than simply identifying a correlation between events?

BM: Moving beyond basic temporal correlation requires analysing the full, multi-touch client journey: tracking how a specific research insight cascades into direct follow-on interactions, analyst queries, corporate access requests and trade executions.

Our platform tracks deep engagement signals across all advisory services. Corporate roadshow attendance, research readership, sales and trading activity, and revenue and vote data come together to create the deepest understanding of intent and value.

AI-driven analytics then evaluate time-to-action metrics and intent signals across front-office channels. This proves causality by demonstrating that client outreach or execution was preceded by specific, high-intent research consumption.

We also capture qualitative feedback and follow-up notes directly from sales and trading desks in real time. The platform links that front-office commentary to the original research trigger, validating the research’s role in driving the client’s decision.

Continuous feedback loops tie these engagement journeys directly to broker vote results and commission allocations, providing concrete, quantifiable proof of research ROI and genuine advisory value across the “One Bank” model.

TTI: You cite examples of previously dormant investors re-engaging with firms and subsequently consuming more research, interacting with other teams or executing trades. What have you learned from analysing those patterns about the relationship between content, client engagement and revenue?

BM: Analysing a sample of our customer base over the last 12 months, we identified 1,119 research reports across 9 customers that resulted in dormant accounts resuming engagement, or new accounts initiating it. Those reports led investors to consume 4,594 additional research reports and substantive email content, and to engage with sales and trading, research, corporate access and banking teams 5,221 times. Across 4 customers, 773 research reports led new or previously dormant investors to execute 1,620 trades.

The first lesson is that content provides a signal of potential opportunity. A triggering action, such as a research report that contains valuable insight, prompts deeper interest, which results in follow-up activity. Identifying this behaviour means firms can provide higher-value services to their clients, such as analyst meetings, corporate access or bespoke access. Engagement data provides an early indication of potential revenue.

The second is that anomalous activity is a strong indicator of potential revenue. A single email open of a recent report doesn’t provide a signal. Sustained or repeat activity does, especially when that behaviour is outside the norm, such as reading an old piece of research.

The third is that speed of follow-up is crucial. The sooner a signal is acted upon, the higher the likelihood of a client consuming more high-value services or acting on a trade opportunity.

TTI: Your AI-powered talent dashboard analyses characteristics such as readability, use of text versus charts and subjectivity alongside subsequent engagement. How far can AI go in helping research teams improve their output without creating incentives to optimise content for engagement metrics at the expense of analytical quality or independence?

BM: AI’s role in the research process is to improve how written research is communicated, rather than to provide conclusions or optimise for engagement. Our talent dashboard correlates engagement with document characteristics, giving heads of research metrics such as readability, chart use and structure. These provide insight into how readers prefer to consume content and analysis.

However, the quality of the analysis is still paramount, and it’s ultimately what drives more revenue. A report with low-quality analysis that has been optimised for high engagement is unlikely to generate high-value, monetisable follow-on activity, such as a meeting with the authoring analyst or a model request for the covered company.

TTI: Capital markets client engagement involves sensitive information and significant regulatory obligations. How have governance, explainability and human oversight shaped the way you deploy AI, particularly as you introduce more sophisticated models into client-facing and revenue-related workflows?

BM: Regulated capital markets workflows require strict guardrails, transparency, explainability and error prevention. Breaches can have significant ramifications for our clients, so ensuring our AI capabilities work in a compliant way is a core tenet of our development philosophy. Since we started integrating AI in 2017, we’ve prioritised clear returns and strict regulatory compliance.

As we introduce AI capabilities, we ensure there’s a human in the loop for any high-risk action, and none of these actions are carried out without explicit approval. The same principle runs through our interaction capture, where users verify the details of an interaction before it’s stored.

TTI: Looking ahead, where do you see the next significant step for AI within the platform? Do you expect it primarily to become better at capturing and analysing activity, or to move further towards proactively recommending actions, identifying opportunities and orchestrating agentic workflows across different parts of the firm?

BM: The next evolution lies in moving beyond passive activity capture and analytics towards proactive, semi-agentic workflows that orchestrate high-value tasks across sales, trading, research, corporate access and investment banking.

By deploying our platform as an enterprise Model Context Protocol (MCP) server, we securely expose rich capital markets context, data entitlements and tools to AI models. That allows AI agents to operate safely within strictly controlled compliance guardrails.

Rather than relying on isolated productivity tools, our vision is built on an end-to-end platform that connects the entire advisory workflow, automatically surfacing actionable client opportunities, identifying churn risk and initiating targeted follow-ups. Agentic workflows will transform how front-office teams collaborate across the “One Bank” model, enabling multi-step orchestration such as automatically triggering corporate access outreach or research follow-ups when high-intent trading signals are detected.

By unifying deep contextual data with standardised, protocol-driven AI, we ensure recommendations aren’t just predictive, but immediately execution-ready within the daily tools where front-office professionals already work.

Singletrack’s approach reflects the growing emphasis on applying AI to established business workflows where its contribution can be measured in operational efficiencies, improved client engagement and commercial outcomes. Its experience also highlights the importance of data quality, contextual understanding and human oversight, particularly when AI-generated insights are used to inform decisions across multiple business functions.

Looking ahead, the company’s plans to introduce more proactive, semi-agentic workflows through Model Context Protocol (MCP) could extend AI’s role from capturing and analysing client activity to identifying opportunities and coordinating follow-up actions. The challenge will be to maintain the governance and information barriers required in regulated capital markets environments while giving front-office teams greater scope to act on the intelligence generated across their organisations.