NYX / 01
Enterprise AI agents: Which workflows are worth a pilot?
A strong starting point combines recurring effort, verifiable outcomes and manageable consequences. Four questions help identify workflows worth testing.
Read insightNYXAI / INSIGHTS
Think clearly.
See further.
Perspectives on AI agents, sovereign systems and decisions under uncertainty.
In focus
Enterprise AI agents: Which workflows are worth a pilot?
A strong starting point combines recurring effort, verifiable outcomes and manageable consequences. Four questions help identify workflows worth testing.
Read insightHUMAN AUTHORITY / MACHINE CAPABILITYA clear mandate. Controlled impact.
NYX / 02
What does an AI agent pilot cost?
Data access, integrations, evaluation and handover shape a pilot budget. A defined scope makes the commercial decision easier to assess.
Read insightNYX / 03
Local AI or cloud: Which architecture fits?
Model location answers only part of the deployment question. Data sources, identities, logs and backups belong in the same decision.
Read insightNYX / 04
Decision intelligence: From data to decisions
How organisations connect information, possible actions and operating constraints in a decision process that can be examined.
Read insightNYX / 05
Controlled AI agents need clear authority
What makes an agent usable in operations: limited permissions, specific approvals, useful handovers and behaviour that can be evaluated.
Read insightNYX / 06
AI in logistics: comparing responses to a supply shortage
An illustrative factory supply example shows how forecasting, optimisation and human approval work together.
Read insightNYX / 07
Defence AI: Clarity under uncertainty
How provenance, uncertainty and human responsibility shape the use of AI in defence.
Read insightNYX / 08
DeepSquare: what chess reveals about decisions under time pressure
Chess makes search, evaluation and time budgets unusually testable. Which lessons extend beyond the board?
Read insightNYX / 09
European decision platforms: assessing alternatives to Palantir
A procurement framework for evaluating data models, actions, governance, operations and exit through a concrete workflow.
Read insightNYX / 10
AI strategy: from ambition to a first decision
An AI strategy earns its name when it names the decisions it will change, the workflows it starts with and the evidence required before scaling. Four building blocks keep it concrete.
Read insightNYX / 11
AI proof of concept (PoC): what it should prove before you invest
A PoC is useful when it answers a named question with representative data and agreed criteria. What counts as success, and what counts as a reason to stop, matters more than the demo.
Read insightFrom perspective to practice