Competition and Markets Authority (CMA): using Glass.AI to identify growth signals across millions of UK firms.
• Public Sector Organisations
Regulator Uses AI to Identify and Understand High-Growth Firms
The Competition and Markets Authority (CMA) researched millions of UK firms, creating a new dataset of web-based growth signals to complement traditional measures of business growth.
↗ Key Impact
New approach to identify ‘hidden’ growth firms
Impossible manual research task reduced to weeks
38,000 high growth firms identified
23 different growth signals across hiring, investment, innovation and collaboration.
Firms displaying growth signals grew employment around twice as fast as firms without signals.
Background
Traditional approaches to identifying high-growth companies typically rely on measures such as revenue and employment, but these measures can be incomplete or become available long after growth has already occurred. Companies leave other signals as they change and evolve. They hire people, raise investment, launch products, form partnerships, enter new markets and expand internationally. These activities create digital footprints across the open web that we collect and analyse to understand more deeply the direction of a company.
The challenge is scale. Finding these signals for one company is relatively straightforward. Finding, classifying and validating them consistently across millions of companies is not. The Competition and Markets Authority's (CMA) Microeconomics Unit (MU) commissioned Glass.AI to explore whether these observable company growth signals could complement traditional firm-level data and provide a richer view of high-growth businesses.
Solution
Using its proprietary Transparent AI research technology, Glass.AI conducted deep research across the public web on millions of UK firms, identifying and structuring evidence of company activity between 2020 and 2025.
The AI-driven web research identified 23 different signals across four broad categories:
Hiring: hiring plans, job advertisements, key hires and other evidence of workforce expansion.
Entry & innovation: new products and services, entry into new industries or geographic markets, new premises and other expansion activity.
Collaboration: new customers, suppliers, partners, joint ventures and research collaborations.
Investment: venture capital, private equity, debt financing, public funding, M&A and other investment activity.
Glass.AI classified and structured the underlying evidence and removed duplicate signals so that individual announcements appearing across multiple online sources did not distort the results.
The results were matched with Companies House and ONS data, allowing the CMA's economists to compare evidence discovered across the web with conventional measures of firm characteristics and growth.
Impact
The research provides evidence that web-derived company signals contain meaningful information about business growth and can complement traditional data with more timely, granular market intelligence.
Results included:
38,000 firms with growth signals identified, representing 81,000 signals between 2020 and 2025.
Signalling firms grew employment around twice as fast as non-signaling firms - 20% p.a. on average.
Investment and hiring signals showed particularly strong relationships with growth, with firms displaying these signals recording average employment growth of around 24%.
Companies displaying a broader variety of signals tended to grow faster, with firms exhibiting three or four signal types showing higher average employment growth than those with only one or two.
Traditional high-growth firms were more likely to display signals.
The CMA concluded that web-based growth signals can complement traditional sources and help identify fast-growing businesses in a timely way, including companies that are less visible in traditional sources.
The research demonstrates how evidence-based AI can turn the public web into a new source of economic evidence by researching millions of businesses consistently while retaining the underlying evidence and provenance behind every result.
For policymakers, analysts, economists and researchers, this creates opportunities to understand not only which companies are growing, but also the activities associated with that growth: from hiring and investment to innovation, market expansion and new commercial relationships.
The opportunity also extends well beyond high-growth firms. The same evidence-led approach can be used to map emerging technologies and sectors, identify innovation and R&D activity, track investment and international expansion, understand skills demand, map supply chains and business ecosystems, monitor technology adoption, identify new market entrants, and detect business expansion or distress.
Want insights no one else can see?
Discover how Glass.AI can transform your research process and unlock intelligence hidden across the web. Let's discuss your specific needs.