Stone Arch Collective

Management Briefing

Thursday, October 1, 2026

Last updated October 1, 2026 at 6:45 AM CT

The Stone Arch Bridge spanning the Mississippi River in Minneapolis
Stone Arch Bridge, Minneapolis — via Wikimedia Commons (placeholder — swap for one of ours).

Today's Leadership Read

A PwC survey of nearly 50,000 workers across 48 countries, published Tuesday, found that workers who use AI daily said they were confident in their job security, and were also 15 points more likely to trust top management than infrequent users, while the largest group, the "engine room" (56%), has neither scarce skills nor much AI experience, and only two in five of these workers said they could access the learning they need — which means the AI-coaching conversation you have or skip with each direct report this week is no longer just a skills decision, it's now a measurable trust-and-retention one.

Today

Management & Org News

  1. 01

    PwC's new 2026 Global Workforce Hopes and Fears Survey, based on almost 50,000 workers across 48 countries and 29 sectors between May and June 2026, splits employees into four groups and finds access to needed learning and development resources fell from 59% to 51% even as AI use climbed; PwC's workforce leader warned the gap risks making a large share of staff operationally irrelevant. PwC / Business Insider

  2. 02

    A new workforce-research roundup out this week highlights that Littler's 2026 Labor Survey found that 67% of respondents expect increased AI use to reshape job responsibilities, while 42% expect work to be redistributed across teams and 30% anticipate reductions in entry-level roles, yet only 9% said their organizations feel very prepared to manage the labor and employee-relations issues AI adoption may create — a direct readiness gap for frontline managers. CTR HCM (citing Littler's 2026 Labor Survey)

  3. 03

    This week OpenAI began bringing voice-based agentic features to the ChatGPT mobile app, letting users trigger workflows such as drafting documents, emails, or summarizing Slack messages directly from their phones, which turns the ChatGPT app into a pocket workflow engine, lowering friction for teams that already use ChatGPT Work but rely on laptops for most automation — meaning the oversight question just moved off the desktop and into employees' pockets. AI Agent Store News (Week of Sept. 25, 2026 roundup)

  4. 04

    A report this week argues your next coworker might well be an AI agent — and will require a whole new model of workplace interactions, noting one researcher said "companies are just starting to wrap their heads around what guardrails" are even applicable for AI agents that don't fit existing HR or IT categories. Wired (via Google News)

Past 7 days

September 30, 2026

  1. 01

    Apple considered cutting about 5,000 AppleCare phone and web support jobs on the theory that AI agents could handle the work, but has now put those plans on indefinite hold, according to Bloomberg's Mark Gurman — even though Apple already runs a generative AI assistant on its support line that walks callers through troubleshooting before transferring to a human. AI Weekly (via Bloomberg)

  2. 02

    A Harvard Gazette explainer published Sept. 29 notes that despite rising AI-attributed layoff headlines and Anthropic CEO Dario Amodei's 2025 warning that AI could eliminate half of entry-level white-collar jobs within five years, the U.S. unemployment rate is still hovering around 4.1%, largely buoyed by healthcare-sector growth — a reminder for managers that the 'AI jobs apocalypse' narrative and the actual labor market aren't moving in lockstep yet. Harvard Gazette

  3. 03

    California's legislature sent three AI-employment bills to Governor Newsom's desk at the end of August, including the revised 'No Robo Bosses Act' (SB 947), and today, Sept. 30, is his statutory deadline to sign or veto them — the second consecutive year he's faced this exact decision after vetoing the bill's predecessor (SB 7) in October 2025. Global Policy Watch

  4. 04

    Separate from SB 947, a California disclosure law is moving on a faster track, requiring covered employers to disclose starting October 1, 2026, whether AI contributed to certain mass layoffs — meaning any manager planning a reorg announcement in California needs a straight answer to the 'was this AI?' question ready by this Thursday, not eventually. Epstein Becker Green (Workforce Bulletin)

  5. 05

    Mark Zuckerberg launched the Meta Enterprise Platform this week, a new business unit selling Meta's AI models and agent tools to companies, and hired MongoDB's CEO Chirantan 'CJ' Desai as Chief Enterprise Platform Officer reporting directly to him — MongoDB's stock fell more than 20% on the news, a concrete example of the same companies flattening middle management simultaneously building new C-suite layers around AI products. VentureBeat / AI Weekly

September 30, 2026

  1. 01

    Uber said it is cutting about 3,300 jobs, roughly 10% of its workforce and its largest reduction since 2020, in a restructuring the company says is meant to flatten management layers and simplify operations as competition intensifies around autonomous driving and transport technology — a direct, named example of AI-era org redesign rather than a routine cost-cutting round. Quest Review Center (citing LiveMint/Reuters)

  2. 02

    A new IBM Institute for Business Value study of 1,500 chief HR officers (CHROs) and 8,800 employees finds a sharp perception gap: seventy-one percent of CHROs say supervising, validating, and overriding AI outputs is the most essential workforce skill, yet only 29% of employees rank that judgment work as important, while sixty percent of employees worry AI is eroding their own critical thinking — and nearly half of CHROs are being excluded from AI strategy development in the first place. MarketingProfs (citing IBM Institute for Business Value)

  3. 03

    HCLSoftware announced Tuesday its intent to acquire Robotiq.ai, a Zagreb-based enterprise RPA (robotic process automation, software that automates repetitive digital tasks) platform used by large banks, insurers, and telecoms, extending HCL's agent-orchestration tools from planning into actual task execution — a sign the enterprise tooling market for managing fleets of AI agents is consolidating fast, meaning more of these bundled tools will likely arrive as IT-mandated defaults rather than opt-in pilots. AI Weekly

September 28, 2026

  1. 01

    As of September 27, 2026, tech layoffs for the year have hit 225,122 workers across 519 tracked events, according to data compiled by Skillsyncer's layoffs tracker, and 2026 is the first year the industry has openly blamed AI for the job losses. For managers, this is the number employees now cite back to you in one-on-ones. Startup Fortune / Skillsyncer

  2. 02

    California lawmakers have passed state Senate Bill 947, which would limit employers' use of AI tools in workplace decision-making. The measure would ban businesses from relying solely on automated decision-making systems for terminations and impose restrictions and notice rules when decisions rely primarily on them. The bill is now on Governor Newsom's desk, and he has until September 30 to sign or veto it. Bloomberg Law / Inside Privacy

  3. 03

    A new Workera survey finds the share of companies offering AI-specific skills training rose from 25% to 58% in a year, but more than eight in 10 employees (84%) say they spend five hours or less per week on skill development or training — a gap between stated investment and actual time protected for learning. PR Newswire / Workera

  4. 04

    On September 22, 2026, Governor Wes Moore announced a statewide artificial intelligence framework focused on protecting Maryland residents, supporting workers, and keeping children safe, and it is particularly relevant to organizations using AI in employment, though the announcement itself does not create new requirements for private businesses, and several initiatives will require legislation or further agency action. The CommLaw Group

  5. 05

    States considered 155 bills addressing AI in the workplace during the 2026 legislative session, with proposals focused on transparency, accountability, worker privacy, and workforce protection. Colorado and Connecticut enacted laws requiring employers to notify applicants when automated hiring tools are used and provide explanations for adverse decisions, with Colorado also granting a right to request human review. MultiState Insider

September 25, 2026

  1. 01

    Microsoft unveiled a major redesign of Copilot into a workplace 'super app,' combining Office collaboration, a natural-language app-builder called Code, and Autopilot — a feature that brings agentic capabilities, AI that can act on tasks on its own, to the company's flagship AI app. The rollout also adds cost-tracking tools so businesses can see what AI use is actually costing them. VentureBeat / Fortune

  2. 02

    A new PayrollOrg survey released Sept. 22 found more than 46% of respondents say they are uncomfortable with artificial intelligence being used in payroll processes, with only about a quarter saying they're comfortable — a reminder that trust gaps aren't limited to scheduling or hiring tools. PayrollOrg via PR Newswire

  3. 03

    Enterprise AI vendor Dataiku announced Agent Management on Sept. 24 — a standalone product that inventories AI agents across platforms, tracks business KPIs and technical performance, and tiers agents by risk, with general availability planned for October, signaling that 'which AI agents does my team even have' is becoming its own management category. AI Agent Store news roundup

  4. 04

    Workplace messaging startup Ando came out of stealth Sept. 24 with a team messaging app built to let AI agents participate as first-class members of conversations, backed by a $20M raise — an early sign that 'who's in this channel' may soon include software colleagues. AI Agent Store news roundup

  5. 05

    Insurance-tech firm Strada announced a browser-automation capability Sept. 24 that lets its agents record and replay workflows inside web-based carrier portals and legacy systems, with full run-time logging for audit — useful for small and mid-size operators stuck with old software, but it raises new questions about who's watching those automated logins. AI Agent Store news roundup

September 21, 2026

  1. 01

    A CNBC feature published today reports that as algorithmic management spreads, companies may need a new 'robot relations' or 'worker-automation relations' function distinct from HR, citing an OECD survey finding that nine in ten managers say their firm has adopted at least one tool to instruct, monitor, or evaluate workers. CNBC

  2. 02

    A third wave of the Konstanz AI study, based on 1,105 employees surveyed in May 2026, found workplace AI use ticked up only slightly (35% to 38%) year over year, and that adoption is mostly employee-driven rather than company-managed — only 55% of AI users say the tool they use most was officially rolled out by their employer. phys.org (University of Konstanz study)

  3. 03

    In a Truth Social post Friday, President Trump said he would appoint an AI czar and stand up a new 'AI Force' modeled on the Space Force to oversee the industry, dismissing AI safety concerns as a 'hoax' — a signal that federal AI oversight structure remains in flux even as states press ahead with their own workplace-AI rules. AI Weekly

September 18, 2026

  1. 01

    Futuresource Consulting's Evolution of Work Survey, published September 17 and based on 3,075 employees across six countries including the U.S., found two-thirds of workers used an AI assistant in the past week and 78% report AI is used somewhere in their organization — but 40% say they've received no employer training, guidance, or policy support for it. Installation International

  2. 02

    A workplace-communication analysis published this week using Gallup's latest indicator data found that as of May 2026, 47% of U.S. employees say their organization has integrated AI tools, but only 30% use it a few times a week or more — the rollout is stalling at the manager level, who is 'the one interpreting the work for the team,' with manager support identified as Gallup's strongest adoption lever. Internal Comms Pro

  3. 03

    Toki launched "Toki Coordination" on Product Hunt September 17 — an autonomous AI executive assistant that proactively emails attendees, negotiates meeting availability across participants, and dispatches calendar invites, offered as a free alternative to human scheduling support, a small sign of how fast agentic tools are moving into manager-adjacent coordination work. AIToolly

  4. 04

    Insurance broker Gallagher's 2026 AI Adoption and Risk Benchmarking report found 63% of businesses have now either fully operationalized AI or implemented it within parts of their business, up sharply from 45% in 2025, even as more than half of respondents cite skills gaps and recruitment challenges as the main barrier to going further. Gallagher (AJG)

Regional

Upper Midwest Watch

  1. 01

    Indiana's statewide AI-adoption initiative said this week that IN AI, Indiana's statewide initiative to accelerate the practical adoption of artificial intelligence, today announced the addition of six peer learning networks, expanding opportunities for Hoosier businesses and professionals to learn from one another and put AI to work, with the next regional roadshow set for Southwest - October 2, 2026. CICP Indiana / IN AI

  2. 02

    A Minneapolis Fed dispatch on Ninth District states reports South Dakota's labor department is now providing training grants to businesses... In South Dakota, the grants accommodate businesses at different stages of AI adoption, from "we don't even know what we need" to "we have begun to implement AI into our workflows", and that when Northern State University in Aberdeen, South Dakota, offered free introductory AI workshops in two nearby towns this past spring, attendance was more than double what organizers expected — with North Dakota still trailing at the bottom of the district's adoption range. Federal Reserve Bank of Minneapolis

Past 7 days

September 30, 2026

  1. 01

    A Wisconsin Watch guest opinion published today argues the state should turn AI's efficiency gains into a formal apprenticeship fund rather than a thinner career ladder, urging employers to use AI to strip out repetitive prep work while giving experienced workers explicit coaching duties so beginners still build judgment — a direct response to the same entry-level hollowing-out concern showing up in national data. Wisconsin Watch

  2. 02

    The Federal Reserve Bank of Minneapolis reports Minnesota leads its nine-state district in AI adoption, with about 20% of businesses integrating AI into at least one function, and finds that where AI does change work, it mostly supplements rather than replaces employee tasks — though 10% of AI-using firms say the technology is now doing work previously done by people. Federal Reserve Bank of Minneapolis

September 30, 2026

  1. 01

    A Wisconsin Watch guest opinion published Tuesday argues the state should build on its apprenticeship tradition — which in 2025 reached a record 18,524 apprentices across more than 3,095 employers — to prevent AI from thinning entry-level career ladders, proposing employers track how long it takes a new worker to reach independent competence, not just how much time AI saves on a dashboard. Wisconsin Watch

  2. 02

    Illinois employers are still working through compliance with amendments to the Illinois Human Rights Act that took effect January 1, 2026 and represent a unique expansion of the state's effort to address AI-driven discrimination risk in hiring and employment activities; current employees must receive notice annually and within 30 days of any new or substantially updated AI system used in employment decisions, with the state's Department of Human Rights still finalizing enforcement rules. Ogletree Deakins / Hinshaw & Culbertson

September 28, 2026

  1. 01

    A Minnesota House bill, the Safeguarding Human Intelligence and Employment in Labor Displacement Act (HF4369), seeks to "protect workers from AI-driven job loss," according to a statement from the Minnesota House, with failure to comply disqualifying an employer from state contracts and risking a fine of up to $10,000 per employee. As of the committee hearing, no action was taken, and more than 31% of Minnesota jobs (813,000 workers) have "high exposure" to generative AI, according to North Star Policy Action and the University of St. Thomas. Minnesota House Session Daily / HR Dive

  2. 02

    Federal Reserve Bank of Minneapolis data show approximately 20% of Minnesota businesses reported integrating artificial intelligence into at least one business function in early 2026, but businesses with 250 or more workers adopt AI at nearly double the rate of firms with fewer than 20 employees. Approximately 96% of AI-using businesses reported no net change in workforce headcount attributable to AI over the prior six months. Bassford Remele

  3. 03

    Gov. Tony Evers and the Wisconsin Department of Workforce Development announced applications are now open to give workers job training in advanced manufacturing or artificial intelligence, with the WisTRAIN program offering more than $7 million in employer job training grants. WisTRAIN is open to employers of any size operating in Wisconsin. WBAY (Madison, WI)

September 25, 2026

  1. 01

    Indiana's statewide IN AI initiative, run with the CEOs of Indiana Corporate Partnership, brought its roadshow to Notre Dame on Sept. 18, connecting South Bend–Elkhart employers with practical AI applications, proven regional results and resources to help organizations move from exploration to implementation — one of the more concrete examples of a state actively coaching mid-size employers on adoption rather than just funding pilots. CICP Indiana AI Adoption Hub

September 21, 2026

  1. 01

    Minnesota's HF4369, the Safeguarding Human Intelligence and Employment in Labor Displacement Act, would require employers with 50 or more full-time Minnesota employees to give 90 days' notice before deploying AI that could displace jobs, with additional notice to labor representatives and the Department of Labor and Industry if 25 or more workers (or 25% of the workforce) are affected. Minnesota House of Representatives, Session Daily

  2. 02

    A MinnPost commentary argues Minnesota missed a chance to pass stronger AI worker protections this session and urges the state to prioritize three things next session: displacement notice and support, limits on digital surveillance/scoring of workers, and guardrails on AI's role in hiring, firing, and pay decisions. MinnPost

  3. 03

    Des Moines-based Business Record reports Iowa employers are navigating a fast-shifting legal landscape as AI spreads into recruiting, hiring, performance evaluation, and termination decisions, with one Iowa employment attorney calling AI regulation 'almost the elephant in the room' for HR compliance. Business Record (Des Moines)

September 18, 2026

  1. 01

    Today, September 18, the CEOs of Indiana Corporate Partnership, the Indiana Chamber of Commerce, the South Bend–Elkhart Regional Partnership, and Notre Dame's iNDustry Labs are bringing the statewide 'IN AI Roadshow Series' to the Morris Inn at Notre Dame, connecting regional businesses and community leaders with practical AI adoption resources as part of Gov. Mike Braun's IN AI initiative. CICP Indiana

  2. 02

    Wisconsin Manufacturers and Commerce's CEO Bauer told the Daily Reporter that AI is "revolutionizing manufacturing the way the assembly line did," describing hybrid "dark factories" where humans work one shift and robots run the others, while acknowledging members are still weighing job impact and data-security risk. The Daily Reporter (Wisconsin)

  3. 03

    The Federal Reserve Bank of Minneapolis reports Minnesota leads the Ninth Federal Reserve District in AI adoption, with businesses in more labor-intensive sectors like agriculture, transportation, and construction lagging well behind at under 10% adoption. Federal Reserve Bank of Minneapolis

  4. 04

    Minnesota law firm Bassford Remele notes that while 96% of AI-using Minnesota businesses reported no net headcount change tied to AI, three bills now before the state legislature would make it an unfair employment practice to deploy AI producing discriminatory hiring, promotion, or termination effects, and would require employers to disclose AI use in those decisions. Bassford Remele

Practice

Try This With Your Team

The Engine Room Check-In

The 10-minute versionThis week, ask each direct report one question in a 1:1: "What's one AI tool or skill you wish you had more training on?" Just listen — don't problem-solve yet.

Level upCompile the answers across your whole team into a one-page 'training wish list' and bring it to your next skip-level or L&D conversation — turning PwC's 'engine room' data point into an actual action item instead of a stat you read.

Learn more — See PwC's 2026 Global Workforce Hopes and Fears Survey

Shadow Agent Amnesty

The 10-minute versionIn your next team meeting, ask plainly: "Is anyone using an AI tool or agent for work that IT doesn't officially know about?" Frame it as a no-blame fact-find, not an audit.

Level upTurn whatever surfaces into a short list and send it to IT/security as a heads-up before it becomes an incident — most organizations have no formal registration requirement for this yet, so you'd be ahead of policy, not behind it.

Learn more — Read the AI Agent Management Gap report

Past 7 days

September 30, 2026

The 'Was This AI?' One-Liner

The 10-minute versionBefore your next team or all-hands meeting, write one honest, plain-language sentence answering the question 'did AI play any role in this decision/reorg/headcount change?' — even if the answer is 'not yet, but we're watching it.' Say it out loud once before you need it live.

Level upShare your draft sentence with your own manager or peer managers this week so the answer is consistent if the same question lands on three desks at once instead of just yours.

Learn more — Why this matters: AI-attributed tech layoffs this year

Shadow AI Show-and-Tell

The 10-minute versionIn your next team meeting, go around and ask each person to name one AI tool (like ChatGPT or Copilot — software that can draft, summarize, or analyze on request) they used this week and for what, with zero judgment attached. The goal is surfacing use, not policing it.

Level upTurn it into a recurring 15-minute monthly slot where someone demos one thing AI helped them do well — this builds the skills-coaching habit that most employer AI training is currently missing.

Learn more — Background: the training gap behind weekly AI use

September 30, 2026

The Skills-Gap Check-In

The 10-minute versionIn your next 1:1 or team huddle, ask each person one question: 'What's one task you've started letting AI do, and what's one task you'd never trust it with yet?' Just listen — no scoring, no follow-up assignment this week.

Level upCompile the answers (anonymized) into a one-page list and share it back with the team, explicitly naming where your own sense of 'AI should handle this' does or doesn't match theirs — new research shows most CHROs and employees disagree sharply on which AI-judgment skills matter most.

Learn more — Read the IBM CHRO/employee gap findings

Name Your Own Layer

The 10-minute versionSend your team a one-question anonymous poll: 'If AI took over half of my routine coordination work, what should I spend the freed-up time on instead?' Read the answers alone; don't respond yet.

Level upDraft one paragraph adding to your own job description reflecting what you learned — framed around oversight and judgment, not task coordination — and share it with your own manager as a starting point for how your role should evolve as management layers get restructured elsewhere in the industry.

Learn more — See how Uber and Google are restructuring management layers

September 28, 2026

The Skills-Visibility Check-In

The 10-minute versionIn your next 1:1 or team huddle, ask each person two questions: 'What's one task you've started using AI for?' and 'What's one you haven't tried yet but wonder about?' Two minutes per person, no follow-up homework required.

Level upCompile answers into a shared one-page doc mapping who's using AI for what — turning it into an informal peer-learning directory instead of a formal training program nobody has time for.

Learn more — See the training-vs-time gap data

The Honest AI-Attribution Line

The 10-minute versionBefore your next org-change or workload conversation, write one honest sentence answering 'did AI play a role in this decision?' — even if the answer is no — so you're not improvising if someone asks.

Level upLoop in HR or your peer managers to align on the same sentence across the team, so employees hear one consistent, credible answer rather than several different ones.

Learn more — See this year's AI-layoff tracker

September 25, 2026

Map Your Team's Shadow AI

The 10-minute versionAt the start of a team meeting, go around and have each person name one AI tool they use at least weekly that was never officially rolled out by the company, plus one task they'd never hand to it. Just listen — don't correct or discourage yet.

Level upTurn the list into a simple one-page 'approved / ask-first / off-limits' tools doc and share it with HR or IT this week, so the next audit or vendor conversation starts from what people actually use, not what was purchased.

Learn more — Why adoption is running ahead of policy

The Overnight Agent Question

The 10-minute versionAsk your team: 'If an AI agent kept working on your project overnight, what's the one decision you'd want it to always stop and ask a human about first?' Collect answers on a whiteboard or shared doc — don't debate yet, just gather.

Level upPick the top 2-3 answers and turn them into an actual written norm for how your team uses AI agents or assistants, then revisit it in a month as tools like Microsoft's new Autopilot start showing up in your stack.

Learn more — See what Microsoft just launched

September 21, 2026

Surface the Shadow AI

The 10-minute versionAt your next team huddle, go around and ask each person to name one AI tool they're already using for work — approved or not — and why. No judgment, just visibility. Write the list on a shared doc.

Level upCross-check the list against your IT/HR-approved tools. For anything unsanctioned but useful, ask IT what it would take to approve it rather than banning it outright — that's usually faster than people expect.

Learn more — Read the Konstanz AI adoption study

Name the Algorithm

The 10-minute versionAsk your team directly: 'Is there any tool right now that schedules, scores, ranks, or evaluates your work automatically?' List whatever surfaces — scheduling software, performance dashboards, ticket-routing tools — on a whiteboard.

Level upFor each tool named, get a plain-language explanation from HR/IT of how it works and who to contact if someone wants to dispute its output. Share that explanation with your team before anyone has to ask.

Learn more — See the CNBC piece on 'robot relations'

September 18, 2026

The 10-Minute AI Skills Check-In

The 10-minute versionIn your next 1:1 or team huddle, ask each person two questions: 'What's one AI task you've tried this month?' and 'What's one thing you wish you knew how to do with AI but haven't tried?' Just listen and note patterns — don't solve anything yet.

Level upTurn the answers into a simple team skills map (who's ahead, who's stuck, who hasn't started) and use it to pair a confident user with someone who wants to learn, closing the training gap yourself instead of waiting for a formal program.

Learn more — See the training-gap data behind this

Invite the Pushback Before It Goes Anonymous

The 10-minute versionBefore your next team decision goes out, spend 10 minutes drafting the strongest counter-argument to your own call — out loud, in the meeting, using your own reasoning (or an AI tool, transparently, in front of the team). Ask: 'What am I missing?'

Level upMake it a standing 5-minute agenda item called 'poke holes in this' before any decision ships, so dissent has an official channel instead of becoming a Slack memo that circulates behind your back.

Learn more — Why managers are the strongest adoption lever

Implications

What This Means for How You Lead

  • 01The coaching conversation you skip is now a trust problem, not just a skills gap: PwC's data ties daily AI use directly to higher trust in management and job-security confidence, so uneven coaching across your team is quietly producing uneven trust in you.
  • 02Agent oversight has become a line-management job, not just IT's: with a majority of IT leaders suspecting unauthorized AI agent use on their own teams and few formal registration policies in place, the 'who approved this bot' question is landing on managers before any policy catches up.
  • 03Vendors are moving faster than most workplace policy: OpenAI's new voice-agent features put always-on AI workflows directly in employees' pockets this week, which means the 'who reviews what the agent did overnight' question is now a mobile-app reality, not a hypothetical.
  • 04Readiness is lagging the hype: multiple surveys out this week independently show employers increasingly expect AI to reshape jobs and layer structures, but single-digit percentages of organizations call themselves genuinely prepared — meaning the honest, one-line answer managers need for 'was this AI?' questions is still mostly undrafted.

Watch

Team Adoption Pulse

  • Shadow agents are becoming an enterprise-wide governance problemGuild.ai / Morning Consult, published Sept. 22, 2026

    A new Guild.ai/Morning Consult study of IT decision-makers found 96.4% of IT decision-makers are confident their organization has a complete and accurate inventory of its AI agents, yet 66.7% of organizations with agents experienced an agent-related operational consequence in the past 12 months. Worse, 60.6% of IT decision-makers say it is very or somewhat likely that employees are deploying agents without formal approval. Only 35.9% of organizations require agents to be registered before deployment.

    GlobeNewswire

  • Business size still decides who actually benefits from AIFederal Reserve Bank of Minneapolis, Ninth District data through mid-2026

    Federal Reserve data shows 30 percent of U.S. firms with at least 250 employees said they use AI, while 17 percent of firms with fewer than 20 employees do — and nationally, 22 percent of U.S. businesses used AI as of the end of July 2026, a 4 percentage point increase since the beginning of the year, showing adoption is accelerating but still concentrated among larger employers.

    Federal Reserve Bank of Minneapolis

  • Most companies still aren't converting AI spend into productivityUC Today AI Agents coverage

    Industry coverage this week flagged that only 5% of companies are closing the AI productivity gap, with vendors like Asana now building unified human-agent work platforms specifically to address why pilots stall before they scale.

    UC Today

Weekly — updates Mondays

Workplace Policy Watch

Last updated: September 28, 2026

Legislation

SB 947 (Automated Decision Systems in Employment)

California · Passed the legislature and on Governor Newsom's desk; he has until September 30 to sign or veto.

Would ban businesses from relying solely on automated decision-making systems for terminations and impose restrictions and notice rules when employer decisions rely primarily on them; the bill would take effect July 1, 2027, and imposes a civil penalty of $500 per violation.

Read the bill

HF4369 — Safeguarding Human Intelligence and Employment in Labor Displacement Act

Minnesota · No action taken as of the committee hearing; referred to the Workforce, Labor, and Economic Development Finance and Policy committee.

Would require employers to issue a notice to employees, labor representatives, the Minnesota Department of Labor and Industry, local officials, and the regional workforce board 90 days before AI-driven layoffs. Applies to businesses with at least 50 full-time Minnesota workers planning AI layoffs affecting at least 25 workers or 25% of staff, whichever is lower. Would also require access to upskilling or retraining programs for affected staff.

Read the bill

SB 26-189 (Colorado AI Act repeal-and-reenact)

Colorado · Signed into law by Governor Polis on May 14, 2026, repealing and reenacting the CAIA; takes effect January 1, 2027.

Removes the onerous requirements of risk management programs and annual impact assessments in favor of a more targeted and procedural framework focused on notice and disclosure requirements.

Read the bill

Statewide AI Framework

Maryland · Announced September 22, 2026; creates no new requirements yet — several initiatives will require legislation or further agency action.

A statewide artificial intelligence framework focused on protecting Maryland residents, supporting workers, and keeping children safe, combining existing state initiatives and recently enacted protections with a forward-looking agenda for legislation and agency guidance.

Read the bill

Grants

Nothing here yet.

A note on accuracy: this briefing is AI-assisted research. Double-check anything you'd cite, quote, or act on — especially numbers, dates, and direct claims.