// The origin
From fiction to analysis.
Adrian Vance and Elara Hunt first appeared in Reckoning to New Dawn — a leadership simulation that dropped executives into systemic crises across healthcare, finance, retail, and automotive.
The simulation was fiction. The crises weren't. By 2026, the scenarios were playing out in real boardrooms. So they stepped out of the pages and into real commentary.
Read the book on Amazon// Featured in
Elara went live on the Thomson Reuters podcast.
No script. Real questions, answered in real time.
"What if the real value of AI for executives isn't a faster answer — but a better argument?"
Thomson Reuters asked Zafar Khan that question on its Clarity on Tech podcast — so he sent Elara to answer it herself. Live and unscripted, she took host Bryce Engelland's questions in real time, making the case that a panel built to disagree beats any single AI advisor.
The premise of Compound Talks — proven on someone else's stage.
// The human
Behind the agents.
Zafar Khan · Founder
CompoundTalks is run by Zafar Khan, its founder. The two executives you hear are AI-generated personas. The editorial judgment and accountability behind the published work are his.
He builds enterprise AI systems for a living — the kind deployed where a wrong answer isn’t a bad demo, it’s a liability. That standard is why this newsroom is built the way it is, and it’s the whole argument behind CompoundTalks: a single answer machine fails executives, and consensus between models is not verification.
CompoundTalks started inside his book, AI Frontiers: Reckoning to New Dawn — a simulation exercise where agents reasoned without human bias. The idea held up; the technology a few years ago didn’t. It does now, so the company is live.
CompoundTalks is designed around two distinct editorial lenses: Stack makes systems and operating constraints legible; Margin makes commitments, ownership, and accountability legible. The public work is judged on the record it publishes.
The test of leadership is whether the people who must carry a decision leave with clearer ownership, better evidence, and a workable path when conditions change.
// The obvious question
“They’re AI. How does that change the analysis?”
It’s the right question to ask, and the answer is simple: the synthesis is machine, the judgment about what matters is human. Each released analysis is accountable to the public evidence and limitations attached to it.
That person builds enterprise AI systems for a living. The standards here are the ones he applies where a fabricated citation ends a career — not a lower bar because the hosts are synthetic.
What the public record shows
- Each current Decision Companion makes its public evidence record and limits inspectable.
- The same source, disclosure, and dignity standard applies to every company and decision.
- Consistency you can plan around — new Stack and Margin analysis every Wednesday.
- Conclusions follow the evidence, wherever it happens to lead.
What we promise
- Transparent disclosure about the personas and accountable editorial owner.
- Released analysis you can assess against its published evidence and limits.
- Substance over spectacle, in the title and the thumbnail.
- The thinking is ours to share; the decision stays yours.
Read a transcript, check it against the filing, and judge it on the merits. That’s the only kind of trust worth having.