MOVE 38FRONTIER LABS

How is the world changing?

From pilots to production

Access to authorised AI tools grew from below 40% to almost 60% in one year. A quarter of organisations have moved at least 40% of experiments into production; 54% expect to reach that threshold within the next three to six months.

  1. 01Now
  2. 02Next
  3. 03Horizon
The lab’s direction across three horizons.

Deloitte, State of AI in the Enterprise 2026, pp. 8–10.

Efficiency today. Reinvention tomorrow.

66% report efficiency or productivity gains today, while only 20% report revenue growth against 74% hoping to achieve it. 34% are deeply transforming products, processes or business models; 30% are redesigning key processes.

  • Data boundaries
  • Provenance
  • Evaluation and criteria
  • Human oversight
  • Observability
  • Access
  • Ownership
  • Deployment and dependencies
Governance dimensions to define according to risk—not a sequence or a certification.

Deloitte, State of AI in the Enterprise 2026, pp. 10–11.

Agents and physical AI

Agentic AI is considered at least moderately important by 23% today and 74% expect it within two years. Physical AI is already used at least in limited form by 58%, with 80% expected within two years; leading applications include collaborative robotics and digital twins.

  1. 01Proceed
  2. 02Do not proceed
  3. 03Test
  4. 04Wait
  5. 05Scale
  6. 06Retire
  7. 07Reassess
The decisions available in a partner Capability Roadmap.

Deloitte, State of AI in the Enterprise 2026, pp. 17–25.

Capability must be governed

Only 21% report mature governance for agentic AI. The leading risks are privacy and data security (73%), legal/IP/regulatory exposure (50%), governance and oversight (46%), and quality or explainability (46%). 83% see sovereign AI as at least moderately important.

  • Data boundaries
  • Provenance
  • Evaluation and criteria
  • Human oversight
  • Observability
  • Access
  • Ownership
  • Deployment and dependencies
Governance dimensions to define according to risk—not a sequence or a certification.

Deloitte, State of AI in the Enterprise 2026, pp. 15–21.

Where we intervene

Deloitte’s 2026 manufacturing research estimates around 20% potential improvement on core KPIs, yet only about one in five use cases is scaled consistently across sites or the enterprise. 38 Frontier Labs closes that gap: from a real process to a measured, governable capability.

Deloitte, AI in Manufacturing Survey 2026.

The future takes shape in today’s decisions.

Bring us a process, a constraint or a possibility. Together we build the capability that can hold up in the real world.

Start a conversation
Contact · Let’s talk