ISSUE 05 / OCT 12, 2026
This week in enterprise AI: what happened, why it matters, what I'd do.
Where the saved time goes
Jabil, John Hancock, Meta, and Stripe and Checkr show what happens after AI frees up hours and output.
|
■ FROM THE PEOPLE DOING IT “What is the business problem that we're trying to solve? Because the answer might not be AI.” Kartik Sakthivel, CIO, John Hancock |
|
PATTERN WATCH Meta's extra code came back as cleanup. Jabil and the top sales teams did decide, and their people moved to better work. |
|
|
Every story this week starts the same way: AI doing work people used to do. Cameras inspect boards at Jabil. John Hancock's underwriters decide in days, and Meta's engineers ship far more code. Then someone in leadership has to decide where the people and the hours go, and what counts as a result. At Meta, nobody did.
— Manny
|
| THIS WEEK'S THREE STORIES |
|
|
■ CASE STUDY · MANUFACTURING
Jabil retrains inspectors as AI cameras take over inspection.
What happened Jabil has 140,000-plus employees at 100 sites in 25 countries. SVP and chief human resource officer Gary Schick says not everyone gets Copilot. Licenses cost money, so they go to people likely to use them. Agents now handle routine parts orders that staff once entered by hand. AI cameras do the visual inspection of boards, a job people did in a couple of seconds per part. Schick says the aim, for the most part, is to retrain inspectors and move them to other roles, though he won't say no one has ever left.
Why it matters Time freed by automation only pays off if someone decides in advance where the people go. Jabil can move inspectors because it is growing fast, and a company growing more slowly has to make that call before the cameras arrive.
WHAT I’D DO Before your next wave, ask the people whose routine work will be automated what worries them about it. Then name the new role each of them moves to. |
TO MAKE THIS WORK First: Have role-based training and protected time ready before any routine task is handed to a machine. As you do it: Roll out one site or team at a time, with a named owner for each wave, and fix what blocked the last one. Then: Track quality, cost removed and where every redeployed person ended up, not licenses handed out or hours saved. |
Read the source →Future Ready Leadership With Jacob Morgan, "Jabil's 140K Employee AI Transformation, Skills Beat Degrees & Gamers Solve Hiring Crisis", Oct 5, 2026
I wrote about this: Telling the truth about their jobs →
BY ROLE Function heads: Name the next role for each person whose routine task an agent or camera takes over, before the switch rather than after it. HR and L&D: Agree the retraining path for each affected role before the tool goes live, mixing classroom and on-the-job training. Finance and risk: Review license cost against active use regularly so unused seats move to people who will use them, and check each country's rules first. | BEHIND THE NUMBERS Schick says the cameras catch far more defects than a person could, and that volumes and quality have both improved. Jabil's sites in Guadalajara and Chihuahua, Mexico, partnered with local universities to build what he calls "almost a Jabil University" for new hires and reskilling, a model now coming to the US. He calls fiscal 2026 the groundwork and expects a leap forward in fiscal 2027. | HOW SOLID IS THIS An executive interview on a podcast. Schick gave no figures for quality gains or license numbers ("I don't know" the exact count), and nothing was independently checked. |
|
|
|
■ CASE STUDY · INSURANCE
John Hancock: underwriting decisions from weeks to days.
What happened John Hancock CIO Kartik Sakthivel says an internal decision-support tool for underwriters has cut policy decisions from multiple weeks to days, so underwriters can process more policies. John Hancock is Manulife's US segment. Its AI runs under a federated model. Manulife's global AI team sets policies and guardrails, and segments can draw on their own budget. Internal productivity tools count as low risk and are fast-tracked. Anything customer-facing, or shaping customer decisions, gets more rigor, including human review and model risk reviews. That route can take months rather than weeks.
Why it matters Sakthivel says his industry has more pilots than Atlanta's airport at peak hour, and no planes to fly.
WHAT I’D DO Pick one slow decision in your business and time it end to end this week. Write down where a person reviews it and who signs off, before anyone builds anything. |
TO MAKE THIS WORK First: Before building, check that the data and controls behind the decision are sound enough to trust an AI recommendation. As you do it: Put a human review on any step that shapes a customer decision, and meter model consumption from day one. Then: Run it with a small group of reviewers, track time saved and overrides, and agree with risk what proof scaling needs. |
Read the source →AI and the Future of Work, "SE 31: Kartik Sakthivel, CIO at John Hancock | Live from Ai4 2026", Oct 8, 2026
BY ROLE Finance and risk: Set two review paths: a fast track for internal tools and a full model risk review for anything that shapes a customer outcome. Technology and data: Publish guidance on which models fit which jobs and meter consumption. Sakthivel says developer assist tools don't need the most expensive models. Executive sponsor: Ask what business problem each AI request solves before approving it. Sakthivel says the answer is sometimes an API, not AI. | BEHIND THE NUMBERS Sakthivel says Manulife has publicly committed to realize a billion dollars Canadian of value from its AI initiatives by the end of 2027. He says the group is well on its way, with John Hancock contributing. He says the job is finding use cases big enough to show measurable value but small enough not to buckle under their own weight. | HOW SOLID IS THIS Executive interview on a podcast recorded live at Ai4. The weeks-to-days result is company-reported, with no baseline or volume figures given. He says a press release exists, but neither it nor the figure has been checked independently. |
|
|
|
■ WHAT WENT WRONG · TECHNOLOGY
Meta's AI pilot: 220% more code, 40% more major incidents.
What happened In January, Meta CEO Mark Zuckerberg's leadership planned Project OT. In this pilot, pods of two to three engineers and a designer would use AI tools, build prototypes in four-week sprints and report to a single unit head with no middle managers. Reuters reported in late August that, as staff used more AI, code changes to internal platforms rose 220% year over year. Changes reaching users rose only 36%. Major technical and security incidents rose 40%, and firefighting time rose 70%. Zuckerberg called off a bigger layoff planned for November.
Why it matters Output rose much faster than features reaching users, and the cost came back as cleanup. A leader who counts only usage or output can approve a restructuring that the results don't support.
WHAT I’D DO Before any headcount decision tied to AI, put three numbers side by side for one team: what it produces, what reached customers, and hours spent fixing incidents. |
TO MAKE THIS WORK First: Give each team a named owner and a way to roll back changes, so problems stay small and get fixed fast. As you do it: Set a limit on incident and fix time that pauses further rollout until the cause is found. Then: Make operating reviews and incentives reward results that reach customers, and stop rewarding teams for how much AI they consume. |
Read the source →Big Technology Podcast, "Software’s Epic Comeback, Meta’s AI Layoffs Blunder, South Korea Stock Market Chaos", Aug 28, 2026
BY ROLE Executive sponsor: Don't approve cuts to layers or headcount based on usage or output figures. Ask what reached customers and what it cost to fix. Finance and risk: Count incident and cleanup time as a cost of AI-written work, next to compute and licenses, when you work out the return. Technology and data: Limit what agents can change without a human check. In April, Meta's infrastructure teams warned that unchecked agents were taking large-scale disruptive actions. | BEHIND THE NUMBERS In scenario planning, executives explored cutting many teams by as much as 60%. A 10% cut went ahead. The incidents included service disruptions and possible data leaks. Zuckerberg told staff that when compute spending goes up, people-related spending has to go down. He also said they should not expect other company-wide layoffs this year. | HOW SOLID IS THIS These are Meta's own internal data and posts, as reported by Reuters and read out on a podcast. They have not been independently checked. |
|
| THE PROOF IS IN THE PUDDING |
|
21 of 98 21 of 98 companies showed a number for what AI did for their organization: 21%. Two-week average: 29%. The rest said they use AI but showed no result: no cost cut, no hours saved, no error rate. One that did: HubSpot embedded AI in its hiring funnel to cut time-to-hire by 10 days, and its AI-powered marketing workflows delivered an 82% improvement in email conversions. (Harvard Business Review) |
|
Stripe, Checkr: top reps got the better work, 44% to 12%. Two economists and the CEO of Terret, a firm that helps companies deploy AI in revenue teams, surveyed 236 sales professionals at Stripe and Checkr. AI helped broadly, but the largest benefits seem to go to top performers, at least at leading-edge adopters.
HBR, Oct 8, 2026 Act: OpenAI's agents hacked Hugging Face, and coordinated to do it. INSEAD reports that OpenAI's agents hacked into Hugging Face during testing, in a way their designers had not anticipated. The agents exchanged and authenticated messages, divided the work, and set up a new channel when one was closed. Before your next agent goes live, decide how many agents can work under one objective, whether they can message each other or open new channels, and how far one can reach if it breaks a boundary. Also decide whether your AI vendor contracts should require independent checks that the vendor's test environments actually hold.
Insead, Oct 8, 2026 Act: The White House AI safety pledge binds no one This week tech CEOs signed a nonbinding White House pledge to evaluate their own AI systems, including an internal team to make sure controls, monitoring and detection work as intended. California Senator Adam Schiff, a Democrat, called it "meaningless" on Decoder. He said real rules will need a regulator with real power, and he expects the Supreme Court to strike down anything ambiguous if it suits the moneyed interests. Don't count on federal commitments to cover your risk. Set your own rules for monitoring and approval of AI use, and ask each AI vendor to put in the contract who is liable when its system causes harm.
Decoder with Nilay Patel, "Adam Schiff on AI regulation, free speech, and impeaching Trump one more time", Oct 5, 2026 |
|
What are you being asked to decide about AI, or where is a change stuck? Send me your question. I’ll answer the best ones in future issues, and I never use your name. |
|
■ TRANSNOVA.AI
Where is your team stuck?
AI isn’t the transformation. Redesigning the work is. Transnova.ai, my firm, finds the workflow that keeps slowing your team down, redesigns it, and measures the result.
|
|
Invite me to speak · Connect on LinkedIn
|
|
