Evolving HR Tech

AI Workforce Planning Requires More Than Task Analysis—It Requires System Analysis Most AI workforce planning still starts with task analysis: “What activities can AI automate, and how much time will that save?” This is necessary—but incomplete. A more accurate lens comes from systems engineering: Amdahl’s Law. It explains why accelerating parts of a system does not translate into proportional end-to-end improvement. Unlike diminishing returns, it introduces a structural ceiling on total performance driven by the non-accelerable portion of work. Formula: Overall Speedup = 1 / [(1 − P) + (P / S)] Where: * P = proportion of work AI can accelerate * S = speedup factor of AI on those tasks * (1 − P) = human core (judgment, coordination, accountability, decision-making) What this means in practice: If a knowledge worker’s execution tasks (25%) are accelerated 10x: Speedup = 1 / [0.75 + (0.25 / 10)] Speedup = 1 / 0.775 ≈ 1.29x Even with extreme task-level acceleration, total role productivity increases by ~29%, not 10x. This is where workforce assumptions often diverge from system reality. Many organizations implicitly assume: “If AI speeds up tasks, we can reduce headcount proportionally.” But Amdahl’s Law shows the constraint is not execution speed—it is the human core of the system. When HR focuses only on task analysis, it misses system constraints: 1. Review bottlenecks. AI increases output faster than it can be validated, shifting load to senior roles and governance functions. 2. Workforce imbalance. Reducing roles based purely on automation potential can weaken coordination, oversight, and decision capacity. 3. Capability erosion. Over-automation of execution can reduce experiential learning pathways for future senior talent. Implication for HR and workforce planning: The focus must expand from task mapping to system mapping: * What work is execution vs. judgment? * Where are the real end-to-end bottlenecks? * How does work actually flow through humans and AI together? This shifts workforce design from activity automation to system throughput. AI improves local task speed. But organizational performance is constrained by system structure. Pic Gene Amdahl (November 16, 1922 – November 10, 2015)

Every LMS vendor is slapping “AI-powered” on their platform these days, but the capabilities behind that label vary widely. I wrote this article to help learning leaders evaluate those capabilities. It covers five questions that I would ask about personalization, responsible AI, administrative controls, content quality, and long-term performance. These questions can help your team understand how a platform makes decisions, what happens when its AI produces poor results, and how well the system continues to perform after the initial implementation period. So, if you're in the market for a new LMS, hopefully, this gives you a better idea of what to look for before you sign that multi-year contract 😉 https://lnkd.in/e5s2XicM

I have made and saved a lot of money using remote teams across all of my companies.  Here’s how you do it: Almost every business could use at least some remote talent. It’s a great way to access a broader talent pool than your local area. You can also lower overhead costs — less office space, lower bills, and even hire talent from other countries. So how do you get the most out of a team that you don’t see face to face? Step 1: Define your objectives and needs Nail down your biggest reason for building a remote team. Broaden your hiring pool? More flexibility? Lower costs? Your main goal guides your future decisions. Then, assess which of your positions are suitable for remote or hybrid work. — Step 2: Develop a remote work policy A solid policy sets the tone and expectations for your team. Try to answer all questions ahead of time. Clarify Scope and Purpose: •  Who is eligible to work remotely? • For hybrid, how many days? • Is there a distance requirement? Set Communication Standards: • When should people be online and available? • What communication tools should they use? Security Protocols: Password manager?  VPN? Are you providing work equipment or expecting BYOD? — Step 3: Update your hiring process Build remote-specific job descriptions: Highlight skills like self-discipline and communication. Use diverse recruitment channels: Remote-specific job boards and communities. Tailor interviews for remote readiness: Include video calls and assess their home office setup. — Step 4: Find the right tools & technology Equip your team with tools that support collaboration and productivity. You’ll probably need: • An async communication hub (like Slack) • A video call platform (Google Meet) • A project management tool (Asana or Trello) • Hardware/software support Provide equipment or offer a stipend. — Step 5: Establish clear communication guidelines Effective communication is the backbone of remote work. Do you need people to: • Set online statuses? • Post daily updates? • Follow a response time rule? • When do you need people available for video calls? Make sure to set regular meetings and check-ins. Weekly stand-ups and monthly all-hands help keep everyone aligned. — Step 6: Build a strong team culture Strong remote teams thrive on culture and connection. Start with thorough virtual onboarding. Set up meet and greets and mentoring sessions. Add regular team activities: • Virtual coffee breaks • Game time • Casual Slack channels Celebrate everything: • Individual and team wins • Holidays • Company milestones — Step 7: Keep tabs on performance Address concerns head-on with clear goals and regular feedback. Set SMART goals: Specific, Measurable, Achievable, Relevant, and Time-bound. Schedule quarterly reviews. Focus on outcomes — not hours worked. — If you’re interested in remote staff for your teams. Comment below or message me and I’ll get you connected.

Remote work challenge: How do you build a connected culture when teams are miles apart? At Bunny Studio we’ve discovered that intentional connection is the foundation of our remote culture. This means consistently reinforcing our values while creating spaces where every team member feels seen and valued. Four initiatives that have transformed our remote culture: 🔸 Weekly Town Halls where teams showcase their impact, creating visibility across departments. 🔸 Digital Recognition through our dedicated Slack “kudos” channel, celebrating wins both big and small. 🔸 Random Coffee Connections via Donut, pairing colleagues for 15-minute conversations that break down silos. 🔸 Strategic Bonding Events that pull us away from routines to build genuine connections. Beyond these programs, we’ve learned two critical lessons: 1. Hiring people who thrive in collaborative environments is non-negotiable. 2. Avoiding rigid specialization prevents isolation and encourages cross-functional thinking. The strongest organizational cultures aren’t imposed from above—they’re co-created by everyone. In a remote environment, this co-creation requires deliberate, consistent effort. 🤝 What’s working in your remote culture? I’d love to hear your strategies.

Building High-Performance Remote Engineering Teams is not just about video calls.... I’ve worked with teams across the UK, Europe, and the US, and one thing is clear: remote work isn’t inherently slower. But a lot of engineering teams fail because they try to run distributed teams like co-located ones. Here’s what really makes a remote engineering team high-performing: 1️⃣ Communication by Design, Not by Chance Async-first: Chat isn’t enough. Document decisions, architectural diagrams, and API contracts in a place everyone can access. Structured updates: Daily standups are optional; status tracking through PR reviews, automated CI pipelines, and project boards is mandatory. 2️⃣ Ownership & Clear Boundaries Each engineer owns services, APIs, or modules end-to-end. Service contracts are explicit. Teams don’t block each other because ownership is clear and dependencies are well-documented. 3️⃣ CI/CD Is Non-Negotiable Remote teams must trust that pushing code won’t break production. Automated testing, linting, and deployment pipelines reduce friction and async bottlenecks. Feature flags and incremental rollouts are your best friend. 4️⃣ Knowledge Visibility Remote teams fail when knowledge lives in heads. Maintain internal wikis, architecture maps, and runbooks. Code reviews aren’t just for QA—they’re the primary async learning tool. 5️⃣ Metrics That Actually Matter Velocity in story points? Fine. But measure deploy frequency, mean time to recovery, bug escape rate, and codebase health metrics. These metrics highlight systemic issues instead of punishing individuals. 6️⃣ Tech Stack Choices Matter Prefer tools that support async collaboration: GitOps, Slack with integrated threads, Jira/Trello boards, distributed logging, observability dashboards. Avoid systems that require constant synchronous attention or centralised knowledge bottlenecks. 7️⃣ Culture Is Explicit, Not Implicit High-performing remote teams share principles in writing: “We merge only green builds,” “We document before we ship,” “We pair when ownership overlaps.” Bottom line: Remote engineering success is built on process, ownership, tooling, and visibility, not on heroic effort or long hours. If your team is still treating async work like a co-located office, you’re leaving productivity and sanity on the table.

I use the word “LMS” a lot. And sometimes I realise not everyone knows what that actually stands for. An LMS is a Learning Management System. At its simplest, it’s a digital classroom: a central place to create, deliver, and track online training. But what does it actually do? In practice, it usually follows a simple workflow: Upload: Bring your content in. Videos, PDFs, quizzes, structured into modules. Assign: Make sure the right people get the right training. Deliver: Let learners access it anywhere, on any device. Track: See completion rates, quiz scores, time spent. Improve: Use that data to refine what’s not working. Organizations adopt an LMS because traditional training gets messy. Different versions of materials. No proof of compliance. High delivery costs. Inconsistent learning experiences. An LMS brings structure and consistency. What’s changing now is the shift from traditional LMS platforms to AI-powered systems. In older systems, everything is manual. You build the content. You update it. You interpret basic reports. With AI, you can generate structured courses from documents in minutes, support learners with AI tutors, and actually understand where knowledge gaps exist. That said, the technology alone doesn’t fix everything. Engagement still matters. Relevance still matters. Simplicity still matters. So I’m curious, when you hear “LMS,” what do you think of? A compliance tool? A content library? Or something more strategic?

  𝗬𝗼𝘂𝗿 𝗔𝘁𝘁𝗿𝗶𝘁𝗶𝗼𝗻 𝗠𝗼𝗱𝗲𝗹 𝗔𝗻𝘀𝘄𝗲𝗿𝘀 𝘁𝗵𝗲 𝗪𝗿𝗼𝗻𝗴 𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻 Continental Airlines saved $40M in a single year by optimizing crew scheduling [INFORMS Edelman Award, 2002]. The US Census Bureau saved $2.5B optimizing field worker assignments [Edelman Award, 2022]. Both used constrained mathematical optimization - linear programming, multi-objective analysis. Standard in supply chain and logistics for decades. Absent from the People Analytics toolkit entirely.   Your attrition model predicts who will leave. Meanwhile, your CHRO is asking: "Given my budget constraints, pay equity requirements, and headcount limits, what combination of salary adjustments, role changes, and development investments across which employee segments produces the best overall retention and productivity outcome?" That is a constrained optimization problem. The techniques to solve it have existed for fifty years. PA has never adopted them.   The reason is structural. PA inherited its entire analytical toolkit from two parent disciplines. I/O psychology contributed psychometrics, regression, and experimental design. Data science contributed ML classification, NLP, and clustering. These are formidable for measuring, predicting, and classifying. The discipline that solves "what should we do given competing constraints" — Operations Research — developed in a completely separate institutional world: different journals, different conferences, different departments. The techniques work for workforce problems. They just live in a different silo.   GenAI highlights this gap. Prediction is becoming commoditized - a competent PA team can now build a serviceable attrition model using GenAI tools in an afternoon. Optimization requires domain expertise to specify the right objective function, the right constraints, and the right trade-offs. That capability remains scarce.   If you framed your last major workforce decision as "predict what will happen," ask yourself: what would have changed if you had framed it as "optimize across competing constraints"?   Full article below.   Dave Millner, Nicole Lettich, Abid Hamid, John Boudreau, Colby Kennedy Nesbitt, Ph.D., Oliver Kasper, Amy Armitage, Igor Menezes, Tilman Sheets   #peopleanalytics #operationsresearch #optimization #decisionscience

What if upskilling your workforce on AI tools is making burnout worse, not better? New systematic review challenges conventional wisdom. A Cardiff University analysis of 201 studies (218,637 employees) reveals digital competence alone provides zero protection against technostress-induced burnout. Researchers identified two primary culprits destroying well-being: techno-overload (forced to work faster and longer through technology) and techno-invasion (constant connectivity bleeding into personal life). Sound familiar? The damage manifests as emotional exhaustion, burnout, and plummeting job satisfaction, even among highly digitally competent employees. 🔥 Why this matters for AI transformation leaders: Without organisational support structures in your AI rollout strategy, you're accelerating towards a well-being crisis. AI training increases digital capability but does nothing to protect psychological capacity. Sustainable transformation requires measuring technostress alongside adoption metrics. The question isn't "Can your people use AI?" It's "Can they use AI without breaking?" 💡 Evidence-based intervention strategies: → Organisational support trumps individual resilience. The meta-finding across 201 studies: training matters, but organisational support is the critical buffer. Give people permission, and systems, to disconnect. Make "strategic unavailability" a core value, not a career liability. Reward sustainable performance, not constant availability. → Diagnose technostress before it becomes burnout. Deploy validated diagnostic tools before and during digital transformations. Brief, single-item measures work brilliantly in fast-paced environments. You need real-time intelligence. → Target the actual stressors, not generic "wellness" The research is unambiguous: focus interventions specifically on techno-overload and techno-invasion. Different role types have different stressors. Create explicit digital boundaries (no-meeting blocks, async-first communication, mandatory shutdown protocols) modelled from leadership. 🧠 The organisations succeeding at AI adoption aren't just deploying the most sophisticated tools, they're protecting human capacity AND scaling digital capability. ---- 👋 Hi I'm Keith. I activate change and transform culture, leadership, and organisations, using behavioural science. Hit Follow for more on human-centred AI adoption strategies.

You can't hire 300,000 workers who don't exist. Infrastructure Australia just confirmed what we've known was coming: workforce needs to double by 2027. Brisbane 2032, 1.2M new homes, AUKUS, energy transition. All competing for the same shrinking labour pool. But here's what the report also says: productivity improvements are the answer, not just more bodies. Every major project currently burns worker hours navigating compliance mazes, chasing approvals, and coordinating between siloed systems. That's where the doubling happens. Not in workforce size, but in what each worker can actually deliver. We've been positioning governance-integrated digital twins for exactly this moment. Not as a technology play, but as a workforce multiplier. When Brisbane City Council reviews a development, when Arup delivers Olympic infrastructure, when tier 1 contractors bid on energy projects... the teams that can automate the governance complexity will win. The procurement shift from Dubai, Abu Dhabi and now Madrid (all using building approval automation) to Australia (Brisbane 2032 pressure) isn't coincidence. It's necessity becoming very visible. The workforce crisis just made digital twins non-negotiable. What productivity constraints are you seeing in your projects right now? https://lnkd.in/gpmdwmBg

I recently posted about the big opportunities many L&Ds are missing by not integrating AI into their existing company flows and got a LOT of DMs about what this actually looks like in practice. Here’s how I’d approach it. Start by evaluating where your organization's knowledge is currently documented. In other words, build your own enterprise knowledge graph. Google coined the term knowledge graph in 2012 to find not just independent artifacts spread across the web, but to contextualize using the relationships between artifacts. Typical search works well if there's one answer, but when that answer is dependent on the context you're in, it gets difficult. By harnessing your company's knowledge graph, connecting all the knowledge specific to your company that sits inside HR systems, L&D tools, presentations, spreadsheets, documents, RFPs, intranets, emails, Slack/Teams channels, and your heads, you can truly achieve learning at scale. Let me use an example to show how your learners can use AI to cater to their own structured but also unstructured learning... Imagine that I’m an Enterprise Account Executive who's just started at a company that sells bike part. An incredibly technical product, filled with a lot of specifications. 1. ONBOARDING: I’m greeted on my first day with an interactive, personalized onboarding. Throughout my onboarding, I participate in self-paced courses, virtual and in-person sessions on Sana. A week later, I might recall something from one of the onboarding sessions I had—I can easily search "What were the 5 principles to account management from the session last week" and LMS generates an answer to my question with a link to the recap. 2. ENGAGEMENT: Fast forward my journey a little and I’m continuing to get ongoing enablement and personalized learning from AI tutors. And I can supplement my learning by chatting with the AI assistant to learn more from the best examples of proposals and demos shared by my team. 3. DEVELOPMENT: As I develop and grow in my role, I can contribute back to my team's development using Sana's AI assisted editor. I can also leverage the AI assistant to auto-complete proposals and RFPs. 4. PROGRESSION: I've reached a pivotal moment in my journey where I'm ready to advance to the next step of becoming a manager. As a result, I've been automatically enrolled in Sana's program to develop essential skills. Hopefully, this gives you a glimpse into a future where harnessing AI and your company's knowledge graph can transform how your employees develop and become even more productive. What else do you think is missing in the journey above? I’d love to hear your thoughts in the comments. #peopleops #learninganddevelopment #AI