4 min read
Why AI Makes Data Governance Essential
Three hallmarks of data quality in the AI era
Key Takeaways
- As AI accelerates both the volume and accessibility of data to investors, quality will become the critical issue and focus of governance efforts
- Data quality standards should include not only accuracy and timeliness, but complete coverage, accessibility through workflows, and actionability
- Companies will face increased data governance burdens; trusting the quality of their data providers is essential
The acceleration of AI is showing that while the volume and accessibility of data will increase, not all of it will be of the quality needed to be useful for investors.
We all know the “garbage in, garbage out” refrain. It won’t be long before the AI revolution exacerbates the gap between the good and the garbage, laying bare the ramifications of poor data governance. If the promise of AI is greater analytic power in more hands, the peril is unsteady foundations leading to poor decisions.
I started my career as a data analyst at Morningstar, so I know the power of data. Data can create a common language that empowers all investors. Data can bring order to chaos. Data can unlock decision-making.
Morningstar has always been in the business of data that speaks, and today our data universe covers well over 800,000 managed investments, 55,000 public companies, and 4.2 million privately held companies—canvassing nearly every investment vehicle in the financial landscape.
With the advent of AI, our goal is to raise the bar on data quality and timeliness, while increasing breadth and depth. Doing so means that investors will have access to data that’s complete, accessible, and actionable.
Complete: You would never build a house from partial blueprints; it’s the same with data that tells only part of a story. High-quality data is comprehensive in exploring upsides and downsides, and it is enriched by people and processes that shape it into useful, comparable signals. That’s where Morningstar’s deep research capability infuses data with our unique intellectual property. The difference is a table of fund returns versus star ratings that bring intuitive visual comparisons across Morningstar Categories. Voila, the data is more valuable to investors.
Accessible: All the best data doesn’t matter if people can’t use it when it’s timely and relevant. That’s why we’ve built an infrastructure that processes an average of 23,000 data points each minute. It’s why we use technology that easily plugs into the workflows of our clients, allowing them to efficiently build portfolios and accurately benchmark performance. Clients can choose to ingest our data directly, incorporate an API into their own tools, access it via Morningstar-branded software, or even get it from third-party software.
Actionable: Good tools, reporting, and applications transform data into meaningful insights, empowering investors to make informed, confident decisions in any market around the world. We’ve made it possible for investors to X-Ray their portfolios on Morningstar.com for years. As another example, using data from over half of U.S. advisors, Enterprise Analytics offers home offices dashboards that provide actionable insights around firmwide activities, including compliance and benchmarking.
We’ve already seen how machine learning and large language models (LLMs) can be used in collecting, analyzing, and delivering data in new ways. Companies have an increased data governance burden to build AI solutions that speak the truth, and relying on trusted data providers to support that effort is essential. It also means we all must be more discerning in the data that runs our lives.
At Morningstar, we’re building an infrastructure that is comprehensive, transparent, and accessible so that Morningstar data always speaks investor success.