Skip to content
← All issues
THE H.O.T. TAKE · ISSUE 02 · NOVEMBER 13, 2026

Keep the judgment.
Where does the judgment sit?

Each weekday we read one AI story where people, organization and technology meet, and end with one thing to try on a single workflow on Monday. Once a week, the best of them arrive here.

Three stories this week, one pattern. Fortune reports that firms which cut roles for AI are asking people back for the judgment the systems couldn't supply. In HBR, four Deloitte leaders show how tighter AI rules push people into workarounds, and argue for giving them authority instead. CrowdStrike put the stop button for a rogue agent in a named person's hands. The agent can take on the work; the judgment, and the authority to use it, stay with people.

Below them: the week in five numbers, six more reads across Human, Organization and Technology, the latest Moonshots episode, and one question to take into Monday.

THIS WEEK IN NUMBERS

Where the judgment went missing.

1 in 5

US workers replaced by AI were later asked back, most often for judgment, relationships and oversight. FORTUNE →

6%

of more than 1,300 leaders think their top executives are fully ready to lead AI-driven change. KORN FERRY →

9%

of large firms have completed a significant operating model change, though 79% run agents in production. GENPACT →

Under 1 in 10

firms name a person accountable for what their agents do. Only 6% can revoke a rogue agent's access within an hour. GRAVITEE →

46%

of frontline workers surveyed had no training, briefing or written rules on AI, though 65% had used it at work. HUMANFORCE →

FIELD NOTE · HUMAN · 2 OF 5

The structure has changed. The humans haven’t.

Researchers gave 515 startups the same AI tools, training and support, then showed half of them how other firms had reorganized their work around AI. That half earned 1.9 times the revenue.

Same software. Same budget. What changed was how the people pictured the work.

John Longhill and Jani Ashmore take the series one layer deeper, to the person being asked to think, decide and relate differently on Monday morning, and to why leaders have to go first.

Read the field note →JOHN LONGHILL WITH JANI ASHMORE · HUMAN · EXO 3.0
THE H.O.T. TAKE · THIS WEEK

One episode, three takes. One move each for Monday.

THE MOONSHOTS READ · EPISODE #309
Watch Moonshots #309: As AI turns political, write down who decides in each workflow

Moonshots #309 covers the midterms' turn against AI, a 5% Nasdaq drop and faster recursive self-improvement, with a segment on AI, politics and the future of decision-making. This week's move starts from that conversation.

OUR READThe rules around AI are now set in campaigns as well as in labs, and they can change faster than any process built on top of them. An organization that has written down who decides what can absorb a new rule by editing a page; one that hasn't has to rediscover its own decision rights under pressure. So start with the organization: make decision rights explicit before the next policy swing makes them urgent.

ON MONDAYWhen the rules around AI can shift with an election, pick one workflow and write down which decisions AI informs and which a named person makes, so a policy change means editing one page instead of rebuilding the process.

WATCH THE EPISODE →Moonshots with Peter Diamandis · Nov 7, 2026 ↗

HUMAN
where judgment goes

Firms that cut roles for AI are rehiring for the judgment it couldn't supply

Fortune reports SHL survey data showing about one in five US workers replaced by AI were later asked back, and a Robert Half survey found a third of hiring managers who cut roles for AI had to reinstate them, most often for institutional knowledge, relationships and oversight.

ON MONDAYBefore automating any role, list the judgment, relationships and checks the person in it actually carries, and design the workflow so those stay with a named human.

Fortune · Nov 9, 2026 ↗

ORGANIZATION
who holds the authority

Tighter AI rules push people into workarounds; give them authority instead

In HBR, four Deloitte leaders describe an "AI control trap": when adoption lags, more monitoring and mandates push employees to hide their AI use, which leaders then answer with still tighter rules.

ON MONDAYPick one team's workflow and give its people clear authority over how they use AI there, with the training and the room to admit mistakes, and judge it by what the team learns rather than by rule compliance.

Harvard Business Review · Nov 9, 2026 ↗

TECHNOLOGY
where the limits hold

CrowdStrike puts the stop button for a rogue agent in a named person's hands

CrowdStrike added an AI Kill Switch to Falcon Guardian that lets an authorized administrator halt a misbehaving agent, block it from relaunching and cut its network access, with every action logged under a human-in-control banner.

ON MONDAYFor one workflow where an agent acts, name the person who can stop it, write down when they should, and test that the stop works before you need it.

CrowdStrike · Nov 3, 2026 ↗

EVERY TAKE ON THE SITE →

ALSO THIS WEEK

Six more reads. One move each.

Two per strand, then the latest Moonshots episode. Each one is on the site in full, with its source.

HUMAN
where judgment goes

Korn Ferry: managers carry AI change that few leaders feel ready to lead

Korn Ferry's survey of more than 1,300 leaders finds only 6% think their top executives are fully ready to lead AI-driven change, while two in three expect managers to spend more time guiding teams through it than running the daily work.

ON MONDAYPick one workflow that is changing, brief its manager on the reasons before the team hears the news, and name one person that manager can take the hard questions to.

Fortune · Nov 7, 2026 ↗

HUMAN
where judgment goes

Supervisors trust AI users less, unless the organization says AI is welcome

Six studies from the Hebrew University of Jerusalem and WHU found supervisors trusted employees less for using AI on the same work, and the penalty largely went away when the organization explicitly encouraged AI use.

ON MONDAYPick one workflow, write down that AI is expected there and what a good result looks like, so people are judged on the work rather than on how it was made.

American Friends of the Hebrew University · Nov 9, 2026 ↗

ORGANIZATION
who holds the authority

Genpact: 79% of firms run agents in production, but 9% have changed how they operate

Genpact's survey of more than 2,000 senior executives found 79% of large firms have AI agents in production, yet only 9% have completed a significant operating model change and 15% tie AI spend to P&L results, while the leaders name owners for AI value and govern agents for their whole life.

ON MONDAYTake one workflow where an agent already works, name the person who owns its result on a P&L line, and write down where the agent acts and where a person decides.

Genpact (PR Newswire) · Nov 12, 2026 ↗

ORGANIZATION
who holds the authority

Avanade: most firms call themselves AI-mature, but few have changed how teams work

Avanade's AI Reality Report, a survey of 3,600 decision-makers in 11 markets, found only 18% have fundamentally reshaped how their teams operate around AI, and most say too much AI spending goes to incremental gains.

ON MONDAYChoose one workflow, name the person with authority to redesign it end to end, and measure it against its own baseline before adding more tools.

The Manila Times · Nov 7, 2026 ↗

TECHNOLOGY
where the limits hold

Gravitee: fewer than one in ten firms name a person accountable for their agents

Gravitee's survey of 900 US and UK technology leaders found 97% run agents that call other agents, yet fewer than one in ten have a named person accountable for what their agents do, and only 6% can revoke a rogue agent's access within an hour.

ON MONDAYTake one agent-run workflow, give it a named owner with the authority and the switch to stop it, and test that the stop works.

Gravitee (GlobeNewswire) · Nov 9, 2026 ↗

TECHNOLOGY
where the limits hold

ServiceNow puts agents on control testing and sends failures to the control owner

ServiceNow announced agents that set up and run continuous checks on a company's controls across cloud, identity and SaaS tools, route each failure to its control owner, and give internal audit full test results instead of samples, with general availability expected in Q1 2027.

ON MONDAYPick one control your team attests to each quarter, confirm who owns it by name, and decide which failures an agent may fix within policy and which wait for that owner.

ServiceNow (Business Wire, via Yahoo Finance) · Nov 10, 2026 ↗

ORGANIZATION
who holds the authority

Moonshots #310: as agents work longer alone, decide when a person checks in

ON MONDAYIn one workflow where an agent already works, write down how long it may run on its own and who it reports to when it stops, before that time horizon stretches again.

Moonshots with Peter Diamandis · Nov 11, 2026 ↗

A QUESTION FOR YOUR LEADERSHIP TEAM

Who in your organization can stop an agent today, and would they know when to?

In Gravitee's survey, fewer than one in ten firms had named a person accountable for their agents. If the answer takes you more than a minute, start there on Monday.

Where does your organization stand?

The H.O.T. Readiness Scan reads Human, Organization and Technology separately and returns a shape, not a grade. Free; your email opens it.

A sample read and free workspace: your organization's areas scored, and your next step
Assess your organization →OR CREATE YOUR FREE WORKSPACE →
Get The H.O.T. Take every Friday.

The week's best takes, one email.

The H.O.T. Take is Pegasus Collective's independent commentary. Sources are linked, not reproduced, and the words are ours. References do not imply endorsement. ExO 3.0 frameworks are by Salim Ismail and OpenExO.

HUMANORGANIZATIONTECHNOLOGY

Thrive in a world of exponential change.