Wellington Management

Wellington Management uses data analytics — including machine learning techniques — internally to help identify and mitigate behavioral biases in its portfolio managers' investment decision-making.

Wellington's Investment Science Group, which has grown to over 68 specialists in data science, trading, and quantitative investing, applies scientific techniques to 'developing professional investors,' which the firm describes as including uncovering and mitigating the downside of behavioral biases. The group works alongside portfolio management teams both temporarily and permanently, with the stated goal of preserving the art of investing while bringing in computing capabilities. This practice involves analyzing patterns in portfolio managers' buy and sell decisions.

Wellington Management uses AI-powered alternative data dashboards that give approximately 200 members of its investment teams access to non-traditional data sets — such as credit card spending and job opening figures — to help anticipate company growth trends and consumer patterns.

Wellington has publicly described a system in which roughly 200 investment team members use dashboards built on alternative data (e.g., credit card transactions, job postings) to help anticipate company growth, hiring trends, consumer behavior, and product changes. This supports the idea-generation phase of the investment process. The firm's Investment Science Group, which has grown to more than 68 specialists, oversees these data science capabilities alongside 100+ portfolio managers who have been trained to code in Python to analyze data and implement machine learning algorithms.

Wellington Management uses internal AI tools that allow portfolio managers and analysts to query the firm's entire research archive, earnings call transcripts, and public commentary, returning detailed answers to specific investment questions in hours rather than days.

In a November 2025 publication, Wellington Global Industry Analyst Brian Barbetta described internal AI tools that can search across all of the firm's research, earnings call transcripts, and public commentary and return detailed answers to specific queries. He gave the example of asking the system what every company in health care is saying about AI — a task that previously required reading extensive research, talking to analysts, and attending company meetings. The firm has confirmed it favors using third-party AI models rather than building its own, and takes a staged implementation approach.