Autonomous, Where It Counts.
The Five Levels of Earned Decision Rights.
Nobody in your company holds authority permanently — it’s delegated on evidence and withdrawn when conditions change. Your planning software should be held to the same bar.
- Tuesday, August 25, 2026
- 8:00 AM PT
- 11:00 AM ET
- 5:00 PM CET
- 8:30 PM IST
- 60 minutes, live on Zoom
- Recording sent to all registrants
June 4: Resilient. July 14: Integrated. This session completes the arc.
Today’s AI-assisted planning produces recommendations your planners may freely ignore — and when a good one is ignored, nobody books the cost. Your systems can tell you who ignored it and when. They cannot tell you what it cost. Meanwhile, the policy that was safe in March is dangerous in July, and the governance that would revise it meets quarterly. Two holes, one conclusion: decision rights can’t be assigned once, in a workshop or an org chart. They have to be computed — continuously, per decision, against the futures your business might actually face.
of supply chain disruptions resolved without human intervention by 2031
Gartner press release, Mar 2026 →
Agentic AI SCM software by 2030
Gartner press release, Apr 2026 →
Autonomous planning: past the Peak of Inflated Expectations
Gartner press release, Nov 2025 →
The destination is no longer in question. The mechanism is.
- 01How to measure what ignored recommendations actually cost you — the economic accountability hole in AI-assisted planning.
- 02The Decision Resilience Score: how robustness across futures, consequence, and reversibility decide — at runtime — whether the machine or the human holds the rights.
- 03The five delegation grades (L1–L5), and why one enterprise legitimately runs three of them at once.
- 04What the top level looks like: the machine improving its own policies while humans own the constitution.
- 05Live: one decision traveling three levels — and a machine that attacked its own plan overnight.
- 0:00Recommendation theater: the cost of being ignored
- 0:14The claim: decision rights, computed
- 0:20The score and the five levels
- 0:33Live: one decision, three levels, one engine
- 0:42Live: the machine attacks its own plan
- 0:45Who Should Decide? — open Q&A
Ashutosh Bansal
Founder & CEO · VYAN
Twenty-plus years building enterprise decision systems for supply-chain-heavy businesses. Now building the System of Intelligence — a Decision Twin that scores every decision before commit and delegates authority only when it’s earned.
Free to attend. Recording, deck, and community invite sent to every registrant.
Zoom collects first & last name, work email, company, and title — plus one custom question: “If you could automate one decision tomorrow, which one?”