Deloitte: Scale 'autonomous intelligence' for real growth

Enterprise leaders must progress past generative applications and scale autonomous intelligence to capture real growth, according to new guidance from Deloitte Consulting LLP. Generating text or summarising internal communications offers localised productivity improvements, yet these abilities rarely alter the core cost or revenue structure of a large organisation. Enterprises are now focused on deploying systems capable of independent execution. Prakul Sharma, principal and AI & Insights Practice Leader at Deloitte Consulting LLP, described this shift as the third stage on an intelligence maturity curve. He noted that while today’s GenAI-era abilities sit in the middle of that curve, agentic AI acts as the bridge into autonomy. The difference lies in agency: autonomous intelligence pursues an outcome by reasoning over a goal, invoking tools and data, and adapting as conditions change, with humans setting guardrails rather than driving every step. To extract actual economic value, these autonomous systems must integrate directly into revenue-generating or cost-heavy workflows. Sharma advises starting with a decision audit to identify value chains where outcomes are bottlenecked by decisions rather than tasks. Achieving this level of automation requires a forensic examination of existing operations, including verifiable identities in enterprise resource planning systems and approval thresholds formally endorsed by legal and compliance teams. Technological execution frequently stalls owing to upstream friction, specifically regarding data infrastructure. Sharma observes that frontier models have become largely interchangeable commodities, but the friction point lies in connecting these engines to legacy data architectures. Most enterprise data estates were built for human analysts, not autonomous systems, meaning reporting-grade data is inadequate for decision-grade requirements. Providing decision-grade data involves integrating autonomous agents with event stores and databases designed to manage both structured and unstructured enterprise information. Transitioning from controlled testing environments to live enterprise deployment exposes vulnerabilities known as the production gap and governance debt. Teams eager to prove a concept frequently bypass standard corporate security protocols during pilots, creating gating items that prevent future scaling. What unites all three failure modes is that each one is invisible during a well-run pilot until the system must operate in the full enterprise with real users and legal scrutiny. Prakul Sharma’s interview was conducted ahead of the AI & Big Data Expo North America, where Deloitte is a sponsor. Sharma will be sharing more of his insights during a panel session at the industry-leading event.