Business objectives, knowledge boundaries and continuous evaluation determine what AI can take on.
OUR PERSPECTIVE
As AI becomes easier to obtain, business judgement matters more.
Model capability is improving while access becomes less expensive. That does not erase differences between enterprises. We see the next AI advantage emerging from task selection, knowledge quality and delegation. Which judgements should involve a machine, which actions can be entrusted to it, and how will the organisation know it still works?
THE SIGNAL. OUR READING.
The shifts that matter. Our interpretation.
>280×
reduction in inference cost at a specified capability level1
2025 · Nov 2022–Oct 2024 · GPT-3.5-level performance
THE ZHICE VIEW
As models become cheaper, worthwhile problems become the scarce resource.
Falling inference prices repeatedly redraw the economic boundary of tasks. They do not remove the costs of process integration, human review or error handling at the same pace. We believe AI investment should move from a fixed tool list to a revisable task portfolio, reassessing automation and human control as capability, prices and consequences change.
task success reported for agents on the OSWorld computer-use benchmark2
2026 · OSWorld benchmark
THE ZHICE VIEW
Agents create value within a controllable scope of delegation.
Task completion has improved substantially, yet failure remains. The enterprise question extends beyond intelligence: how long a chain can be delegated, what consequences can it trigger, and who takes over when it fails? In energy businesses, authority, knowledge grounding and recovery will jointly determine usable capability alongside the model.
of surveyed data leaders are early in defining how to scale and govern agents3
2025 · 1,700 senior data and analytics leaders globally
THE ZHICE VIEW
Governance capability will determine how far AI can go.
Agents reading information, organising workflows and initiating actions across systems challenge existing application boundaries. We believe scale means more than additional users and calls: permissions, evaluation, exception handling and accountability must expand together. Governance belongs in product design so the organisation knows when to delegate and when to intervene.
Allocate AI resources around business value rather than model popularity. Assess task value, data readiness, consequences of failure and organisational readiness together, building a portfolio that can be validated progressively and explaining what evidence supports further investment.
Moving from answering questions to helping complete tasks requires business knowledge, tool boundaries and clear authority. ZHICE connects traceable knowledge, reviewable plans and authorised actions around energy-sector work, enabling people to inspect, take over and correct.
Make release and scaling decisions reviewable. Evaluate task quality, risk, cost and adoption across model, knowledge and workflow versions, bringing governance into daily operations rather than leaving it in a policy or pre-release checklist.
Look beyond capability to judge what comes next in AI.
We track technical progress and test whether organisations can absorb it. The business boundary of AI changes with task economics, failure modes and accountability; one procurement decision cannot define it permanently.
01
Cost per valid task, not price per call
What we watch
Track calls, review time and exception handling needed for a usable result. Separate cheaper inference from expanding consumption.
How it changes the judgement
Falling review effort can justify revisiting the task portfolio. Cheaper calls with more exception work point to a need to redraw task boundaries.
02
Recovery across longer task chains
What we watch
Watch whether agents recognise deviation, stop inappropriate action and recover from interruption to an acceptable business state.
How it changes the judgement
Maturing recovery supports discussion of longer delegation. Failures that are hard to locate warrant narrower authority.
03
Business owners take responsibility for AI
What we watch
Look for owners who can explain outcome changes, adjust scope and jointly decide model and knowledge updates with technical teams.
How it changes the judgement
Sustained business ownership supports scale. Accountability left with a temporary project team signals unfinished operating arrangements.
A VIEW THAT CAN BE PUT TO WORK
A perspective must stand up in practice.
Five stages—task analysis, portfolio choices, controlled design, evidence-based validation and scaled operations—connect business investment decisions with governance across the AI lifecycle. Drawing on published management research, the NIST risk management framework and agent engineering practice, the method brings evaluation datasets, permissions, human review and exit mechanisms into the design stage. Task requirements guide model selection; validation evidence determines deployment scope. Each capability upgrade prompts a fresh review of business outcomes and risk boundaries.