The decisions that make enterprise AI useful
I speak about the choices behind AI adoption: what a useful result costs, when an agent should act, and how people retain judgment and responsibility as the technology changes.
We can work through the cost of one task, including retries and review, or follow a proposed action from recommendation to approval and a confirmed result. The articles and worksheets on this site provide concrete examples your audience can examine and use.
Topics we can work through
- The cost of a useful AI result. Compare models at the workflow level, including retries, human review and an accepted outcome. Use a worked cost model to see which assumptions change the decision. Read the analysis and use the workbook.
- Agent security and human control. Follow a proposed action through access, approval and execution. Examine where review can prevent a mistake, and where it becomes a routine click. Read the argument.
- The context behind a decision. Work through a case where a real policy is still the wrong policy for the date or person involved. Identify the evidence, authority and current state the system needs. Explore the example.
- Living and working with personal AI. The story of Maya opens a more personal conversation about continuity, memory and the choices a person keeps as AI becomes part of daily life.
Selected recognition
MIT CISR · Listed in “In the Media” · Enterprise Digital Twin (4 January 2026). Read the listed article.
VentureBeat · Quoted analysis · Chain of Draft
DNV · Research reference 74 · Test-time compute. Read the original article.
World Journal of Radiology · Journal reference 228 · Large Concept Models (PDF). Read the cited article.
HackerNoon · Reference 16 · Vibe coding. Read the referenced article.
AussieAI · Research bibliography · Chain of Draft. Read the original analysis.
Biography for organizers and editors
Ajith Vallath Prabhakar writes about enterprise AI at ajithp.com and builds Maya, a personal agentic OS. His work includes the Control Tower Reasoning System (CTRS), research deep dives and practical guides. He focuses on the economics of useful AI work, the context behind decisions, and the relationship between human judgment and agent action.
Talks, briefings and media requests
Keynotes, panels and interviews can explore the broader argument; smaller briefings can work through a particular decision.
For an event, send the topic, audience, format and date you have in mind. For an interview or comment, include the question, publication or programme, and deadline. Links to a brief or an article you want to discuss are useful.
You must be logged in to post a comment.