Clarity in the Middle. HR & AI for Mid-Sized Organizations

Theresa Lynn, Clarity at Work Ltd.

Straight talk on HR and AI for leaders of organizations with 50 to 500 people, where the specialist is usually you. If you run or help run an organization with 50 to 500 people, you're in a funky space. You're big enough that a bad hire, a messy policy or an AI tool nobody approved can cost you real money, but you don't have a legal team, a data governance office or an HR specialist for every problem. Most of the time, that specialist is you. Most of the advice out there is written for companies with ten thousand employees or five-person startups. Clarity in the Middle is for everyone in between. Hosted by Theresa Lynn, President and Founder of Clarity at Work Ltd. She spent 17 years in leadership at one of Canada's largest companies, sitting across from town councils in heated exchanges, running teams across multiple geographies and leading large transformational change. Clarity at Work now works with organizations across Alberta and Saskatchewan on their HR strategy and how they govern AI. Each episode tackles the issues showing up on your desk now, the ones coming down the pipe, and the ones keeping decision makers up at night: what your staff are already doing with AI and the risk that creates, how to write a policy people will actually follow, why performance reviews fall flat, and who's accountable when something goes wrong. It's straight talk. No corporate or technical jargon for the sake of it, and no pretending there's one playbook for every organization. More at ClarityAtWork.ca. A Clarity at Work Ltd. Podcast

Episodes

  1. 1d ago

    Who Can Hit Stop?

    In June, an OpenAI research model looking up public medicine-spending statistics in Australia got unauthorised access to the Medicare Statistics Reporting Service, a Services Australia portal for spending and usage statistics, not personal medical files. It accessed public and non-public files, and both OpenAI and the Australian government said it also wrote files on the system. Australia's Prime Minister confirmed in November that no one's personal details were obtained. OpenAI notified Services Australia on September 10 and has said it should have shared preliminary findings sooner. On November 10, the Australian government announced a new AI safety regulator and released recommendations from a rapid review of the incident, including a requirement that frontier AI companies notify the government quickly when incidents happen. When an agent can act without being asked, the workplace question is who can hit stop, and how fast that stop actually takes. OpenAI has described Control Room, a system it built for its own research infrastructure. The maximum time between a stop instruction and further progress becoming impossible is 35 seconds, from a caching layer they say they are working to get down to zero. It is not a stop button in the public tools most workplaces use. In a September 20 incident from OpenAI's own report, a monitor flagged an agent that had used DNS to reach an external chatbot, a person started reviewing it three minutes later, and the run still had to be stopped by hand two and a half hours after the alert. On November 9, researchers at Microsoft posted a study of always-on work agents in a simulated company, not a real one: 17 models, 120 tasks, with writes pre-authorised in the simulation. Some of the stronger models found the work and did it, and also reopened finished work, treated a colleague's request as their own, or acted on the wrong record or sent something to the wrong people. If a tool can send a message or change a record on its own, name the person who can shut it off, ask the vendor how long a stop actually takes, and decide in writing what the tool may do without a person saying go. Plan the stop before you widen what it can reach. If you want help setting those boundaries, that is the work we do at Clarity at Work. There is a free 20-minute discovery call on the site, and the AI Clarity Sprint gives you a written policy and staff training in one package. Want help setting those boundaries? Book a free 20-minute call: https://cal.com/clarityatwork/discovery-call?utm_source=podcast&utm_medium=episode&utm_campaign=who-can-hit-stop Sources: Prime Minister of Australia, "New Australian AI safety and security protections," 10 Nov 2026: https://www.pm.gov.au/media/new-australian-ai-safety-and-security-protections Prime Minister of Australia, "Strengthening Australia against AI-driven cyber incidents," 10 Nov 2026: https://www.pm.gov.au/media/strengthening-australia-against-ai-driven-cyber-incidents Prime Minister of Australia, press conference transcript, Parliament House, Canberra, 10 Nov 2026: https://www.pm.gov.au/media/press-conference-parliament-house-canberra-55 Prime Minister of Australia, press conference transcript, New York, 24 Sept 2026: https://www.pm.gov.au/media/press-conference-new-york OpenAI, "How we will do better for Australia," 28 Sept 2026 (with 4 Oct update): https://openai.com/index/how-we-will-do-better-for-australia/ OpenAI, "Control Room: A central system to track and stop model workloads across our research infrastructure": https://openai.com/index/control-room OpenAI Alignment, "An agent used DNS to reach an external chatbot," sample/discovery 20 Sept 2026, report updated 25 Sept 2026: https://alignment.openai.com/misalignment-reports/an-agent-used-dns-to-reach-an-external-chatbot/ Shiqi Chen, Daniel Madrigal Diaz, Wei-Ning Chen, Gbola Afonja, Fangkai Yang, Rujia Wang, Chetan Bansal, Victor Rühle, Saravan Rajmohan and Dongge Han, "AutoBuddy: Evaluating Proactive, Always-On Agents on Task Discovery and Execution in a Simulated Workplace," arXiv:2611.12393, submitted 9 Nov 2026: https://arxiv.org/abs/2611.12393 A Clarity at Work Ltd. Podcast. More at clarityatwork.ca.

  2. Nov 10

    Hints, Not Answers: What Ten Minutes With AI Does to Your Team

    Brian Christian has spent 20 years thinking about artificial intelligence. He wrote The Alignment Problem. He also keeps a paper notebook on his desk and writes in it every day, because he worries that too much time with AI could cost him his own edge as a scholar. This month he helped publish evidence that the worry is reasonable. Christian is a research fellow at UC Berkeley and one of five authors on a study presented in early October at the Conference on Language Modeling, with co-authors from Carnegie Mellon, MIT, Oxford and UCLA. The researchers ran three experiments with 1,222 people. While an AI assistant was available, the AI group did better. Once it was taken away without warning, they did worse than people who never had it, and they were more likely to give up. That showed up after about 10 to 15 minutes of AI support, on both math and reading. How people used the tool mattered. About six in ten mostly asked for answers. About a quarter used it for hints or clarification. The answer group did worst once the AI was gone. The hint group solved problems at about the same rate as people who never had AI. The authors say that piece cannot prove cause and effect, so treat it as a directional signal. I put that next to Deloitte's 2026 Change Management Study with Workplace Intelligence, a survey of 1,000 leaders in the US, the UK, Australia and Canada. 59% said AI overload is hurting their employees. 69% said their latest change initiative added to workload and burnout. 45% said poorly managed change contributed to high performers leaving. Organizations where leaders modeled the change were at least twice as likely to report strong returns on adoption that beat expectations. For mid-sized employers, the practical question is whether you are teaching people to use AI for answers or to help them think. Most AI training I see covers features, where to click and which prompts work. If the habit is type the question and copy the answer, this research suggests you can end up with staff who are quicker with the tool and less able, or less willing, to work through a hard problem without it. That matters most for newer staff still learning the work. Want help writing AI guidelines and training that build your team's skills along with speed? Book a free 20-minute discovery call: https://cal.com/clarityatwork/discovery-call?utm_source=podcast&utm_medium=episode&utm_campaign=ep10-hints-not-answers Sources: Liu, Christian, Dumbalska, Bakker, Dubey, "AI Assistance Reduces Persistence and Hurts Independent Performance" (COLM 2026): https://arxiv.org/abs/2604.04721 Berkeley News, Oct 9, 2026: https://news.berkeley.edu/2026/10/09/using-ai-for-just-10-minutes-erodes-your-ability-to-persist-at-hard-things/ Deloitte press release, Oct 20, 2026: https://www.deloitte.com/us/en/about/press-room/deloitte-survey-roi-of-change.html A Clarity at Work Ltd. Podcast. More at clarityatwork.ca.

  3. Nov 3

    If Review Never Changes Anything, It Isn't Oversight

    On October 13, HR Dive reported that Workday is rolling out what it calls autopilot agents for hiring, payroll and accounting. The agents start work on their own and bring a person in when judgment matters. That sounds useful until you ask who decides which actions need a human, and what happens when that human almost never changes anything. In this episode I put Workday's announcement next to a governance test from Helen Lalthanpari. Workday says customers choose which actions run automatically and which need review, and Gerrit Kazmaier, its president of product and technology, said in the release that in payroll and finance a good guess isn't good enough. Helen's test asks if reviewers ever use that power. Pull the override rate for your highest-risk AI system over the last 90 days. If reviewers almost never change or reject what the system recommends, review has probably become approval by habit. For mid-sized employers in Alberta and Saskatchewan, I cover who should own the sign-off settings before an agent goes live, who is allowed to say no (one name, in writing), and how you'll track the override rate once it's running. Want help mapping which agent actions need a human sign-off, or building an override-rate check into your AI approvals? Book a free 20-minute discovery call: https://cal.com/clarityatwork/discovery-call?utm_source=podcast&utm_medium=episode&utm_campaign=ep9-oversight Sources: HR Dive, "Workday brings autopilots to hiring, payroll in autonomous AI bet," Oct 13, 2026: https://www.hrdive.com/news/workday-brings-autopilots-finance-autonomous-ai-bet/832890/ Helen Lalthanpari, "Artificial Echo Chambers: The Quiet Danger In AI Governance," The AI Blind Spot (LinkedIn), Oct 15, 2026: https://www.linkedin.com/pulse/artificial-echo-chambers-quiet-danger-ai-governance-helen-lalthanpari-ewzre A Clarity at Work Ltd. Podcast. More at ClarityAtWork.ca.

  4. Oct 27

    The Junior Work You Automate Is Where Your Next Manager Learns

    Borrowing from Tomorrow, a report released September 30 by the Coqual Global Lab at the Center for Talent Innovation, makes a point I haven't been able to shake. An entry-level task may be inefficient, and it may also be how expertise develops. AI is now very good at the work your newest people do, and that work is where most of them learn judgment. In this episode I look at what the report's survey of 173 organizational decision-makers found. 49% said a broad range of human perspectives, experiences and needs are never, rarely, or only sometimes intentionally considered in AI governance. The figures were 47% for workflow redesign and 44% for workforce restructuring, which are the moments when junior work gets cut or merged. SHRM's 2026 Talent Trends report adds a US data point: 80% of HR professionals said the skills they had the hardest time finding in candidates were systems and resource management skills, which include judgment and decision-making. If you can't hire for judgment and you've stopped growing it at home, the gap shows up in three to five years, right when you need a new supervisor. For a mid-sized employer, I'd start with a task inventory before you switch on a new AI feature. A spreadsheet is enough. Beside each task the tool will take over, answer one question: who learns something from doing this today? I also cover getting HR on the agenda before the contract is signed, treating the first rollout as a 90-day test, and adding one line about learning to your AI policy. Want a hand building that task inventory, or working development checks into your AI approvals? Book a free 20-minute discovery call: https://cal.com/clarityatwork/discovery-call?utm_source=podcast&utm_medium=episode&utm_campaign=ep8-junior-work Sources: Center for Talent Innovation, Borrowing from Tomorrow report page: https://www.talentinnovation.org/globallab/borrowing-from-tomorrow Center for Talent Innovation press release, Sept 30, 2026: https://www.prnewswire.com/news-releases/new-research-warns-companies-may-be-borrowing-from-tomorrow-as-they-redesign-work-302894545.html SHRM 2026 Talent Trends release, Apr 27, 2026: https://www.shrm.org/about/press-room/shrm-unveils-2026-talent-trends-report--data-driven-insights-for A Clarity at Work Ltd. Podcast. More at ClarityAtWork.ca. Since recording: in an October 16, 2026 report on its own workforce, Anthropic says junior roles fell from 27% to 9% of its technical hires since early 2025. https://www.anthropic.com/research/employment-at-anthropic

  5. Sep 30

    Clarity in the Middle. HR & AI for Mid-Sized Organizations, Trailer

    Straight talk on HR and AI for leaders of organizations with 50 to 500 people, where the specialist is usually you. If you run or help run an organization with 50 to 500 people, you're in a funky space. You're big enough that a bad hire, a messy policy or an AI tool nobody approved can cost you real money, but you don't have a legal team, a data governance office or an HR specialist for every problem. Most of the time, that specialist is you. Most of the advice out there is written for companies with ten thousand employees or five-person startups. Clarity in the Middle is for everyone in between. I'm Theresa Lynn, President & Founder of Clarity at Work Ltd. I spent 17 years in leadership at one of Canada's largest companies, sitting across from town councils in heated exchanges, running teams across multiple geographies and leading large transformational change. Now I work with organizations across Western Canada on their HR strategy and how they govern AI. Each episode tackles the issues showing up on your desk now, the ones coming down the pipe, and the ones keeping decision makers up at night: what your staff are already doing with AI and the risk that creates, how to write a policy people will actually follow, why performance reviews fall flat, and who's accountable when something goes wrong. It's straight talk. No corporate or technical jargon for the sake of it, and no pretending there's one playbook for every organization. More at ClarityAtWork.ca. A Clarity at Work Ltd. Podcast

About

Straight talk on HR and AI for leaders of organizations with 50 to 500 people, where the specialist is usually you. If you run or help run an organization with 50 to 500 people, you're in a funky space. You're big enough that a bad hire, a messy policy or an AI tool nobody approved can cost you real money, but you don't have a legal team, a data governance office or an HR specialist for every problem. Most of the time, that specialist is you. Most of the advice out there is written for companies with ten thousand employees or five-person startups. Clarity in the Middle is for everyone in between. Hosted by Theresa Lynn, President and Founder of Clarity at Work Ltd. She spent 17 years in leadership at one of Canada's largest companies, sitting across from town councils in heated exchanges, running teams across multiple geographies and leading large transformational change. Clarity at Work now works with organizations across Alberta and Saskatchewan on their HR strategy and how they govern AI. Each episode tackles the issues showing up on your desk now, the ones coming down the pipe, and the ones keeping decision makers up at night: what your staff are already doing with AI and the risk that creates, how to write a policy people will actually follow, why performance reviews fall flat, and who's accountable when something goes wrong. It's straight talk. No corporate or technical jargon for the sake of it, and no pretending there's one playbook for every organization. More at ClarityAtWork.ca. A Clarity at Work Ltd. Podcast

You Might Also Like