Franklin Templeton has been running AI across live operations for long enough that it can now measure results. Its Intelligence Hub, a platform built on Microsoft Azure, pulls together fragmented data sources, research materials, and more than 15 separate workflow tools into a single interface for its sales team. Salespeople get AI-generated recommendations on which financial advisers to visit, what to focus on in each conversation, and which clients are worth prioritising based on geography and relationship history. After a year-long pilot, it was rolled out to the full sales force in early 2026.
The back office got the same treatment. The firm has AI-enabled platforms in production for automated reconciliation of trades, which is the often-invisible but expensive process of matching up records across different systems. Its CEO noted that every large financial institution spends heavily on this kind of work, and AI can do much of it faster and with fewer errors.
On the investment research side, the firm built an AI system called Gromit that can study complex topics independently, check whether human analysts have their facts right, and deliberately argue the opposing view. This is not about replacing analysts; it is about making them harder to fool, including by their own thinking. A separate platform called MosaiQ combines portfolio construction and manager research. An AI assistant within it, called Pixel, responds to plain-language questions and can carry out multi-step tasks on behalf of the user. The stated philosophy is "copilot, not autopilot."
This is a useful frame for any business operator to hold onto. Across a 2026 global survey of 131 asset managers, 74% said they use AI for operational efficiency and 69% use it as a support tool for analysis, but only 6% said AI is used for actual decision-making. The gains showing up in real numbers are in operations and speed of insight, not in improved investment returns. Only 8% of firms in that survey reported measurable improvement in returns from AI.
The honest picture is that AI is restructuring the cost base and the speed of work, not yet producing magic investment outcomes. BCG's 2026 global asset management report found that AI-driven workflows can reduce operational costs by around 40% and free up 35% to 50% of distribution capacity for firms that go far enough with implementation. That is a structural cost advantage, not a marginal productivity improvement.
There is a competitive dynamic worth noting. A large firm like Franklin Templeton can invest in bespoke platforms, partnerships with Microsoft and specialist AI vendors, and a dedicated chief AI officer with a team underneath him. But a 2026 survey of financial advisers found that 63% believe AI will allow smaller advisory practices to compete with much larger firms. The tools are becoming cheap enough and widely available enough that the moat created by scale is narrowing.
The broader industry data confirms the direction. A SimCorp report found that 70% of investment firms now use AI in their front-office operations, compared to roughly one in ten the previous year. Large firms plan to spend an average of $101 million on AI in 2026 alone, according to KPMG. And 94% of organisations say they will keep investing even if results do not show up this year.
For business operators in any sector, the Franklin Templeton story is less about investment management and more about what a serious, company-wide AI deployment actually looks like in practice. It requires someone at a senior level who owns it, product teams that combine business thinking and engineering in the same unit, an adoption team whose job is to get employees actually using the tools, and a governance structure that keeps regulators and auditors comfortable. The technology is the easy part. The organisation design is the hard part, and most companies have not started it yet.