according to the study, employees at more than 90% of companies are using personal chatbot accounts for daily tasks—largely without it approval. in contrast, only 40% of companies have purchased official large language model (llm) subscriptions.
this trend underscores what mit calls the “genai divide.” despite $30–$40 billion in enterprise investments, just 5% of organizations report transformative results from official initiatives. the vast majority—95%—see no measurable impact on their bottom line. yet beneath the surface, widespread use of personal ai tools suggests far higher adoption rates than executives realize.
instead of waiting for corporate genai projects bogged down by technical hurdles and bureaucracy, employees are turning to personal chatgpt accounts, claude subscriptions, and other consumer-grade tools to streamline their work. these shadow practices often fly under it’s radar but deliver immediate value.
“employees are already bridging the genai divide through personal ai tools. this ‘shadow ai’ frequently generates better roi than formal initiatives and highlights what actually works,” the report notes.
the 40/90 split
the findings are based on an analysis of 300+ corporate ai projects, interviews with 52 organizations, and survey data from 153 senior leaders.
key reasons behind the divide include:
-
flexibility & speed: consumer ai tools like chatgpt and copilot are easy to adopt, adaptable, and deliver instant results—unlike many rigid enterprise solutions.
-
workflow fit: employees can tailor consumer tools to their needs without waiting for approval cycles or complex integrations.
-
accessibility: with few barriers to entry, shadow ai enables experimentation and faster adoption.
by contrast, official genai deployments often stumble due to slow rollouts, clunky interfaces, and technical limitations, leaving a gap between pilot projects and real-world use.
as mit concludes: “organizations that recognize and build on these grassroots adoption patterns represent the future of enterprise ai.”