Scaling AI in Learning: Behind-the-Scenes on Guardrails and Outcomes

March is stacked. 🔥 Our LIT Spotlight series brings three practical, forward-thinking conversations designed for learning leaders who want to build smarter ecosystems. 🗓️ 𝗠𝗮𝗿𝗰𝗵 𝟭𝟮, 𝟮𝟬𝟮𝟲  Melissa Milloway   𝘜𝘴𝘪𝘯𝘨 𝘎𝘦𝘯𝘈𝘐 𝘪𝘯 𝘓𝘦𝘢𝘳𝘯𝘪𝘯𝘨 𝘗𝘳𝘰𝘫𝘦𝘤𝘵𝘴 𝘞𝘪𝘵𝘩𝘰𝘶𝘵 𝘓𝘰𝘴𝘪𝘯𝘨 𝘊𝘰𝘯𝘵𝘳𝘰𝘭  Real examples. Real guardrails. Real talk about designing AI-powered learning that actually works. 🗓️ 𝗠𝗮𝗿𝗰𝗵 𝟭𝟵, 𝟮𝟬𝟮𝟲  Shelly Blake-Plock 𝘞𝘩𝘢𝘵 𝘪𝘴 𝘐2𝘐𝘋𝘓 𝘢𝘯𝘥 𝘞𝘩𝘢𝘵’𝘴 𝘐𝘵 𝘈𝘭𝘭 𝘈𝘣𝘰𝘶𝘵?  In an age of AI proliferation, it is more crucial than ever to ensure the promise of data interoperability to increase transparency, verifiability, and accessibility. Introducing I2IDL, the Institute for Infrastructure and Interoperable Data in Learning. Come find out what they're all about and explore new opportunities for innovation and accessibility in learning data and open source systems built to support better outcomes. 🗓️ 𝗠𝗮𝗿𝗰𝗵 𝟮𝟲, 𝟮𝟬𝟮𝟲  Jason Haag 𝘐𝘴 𝘠𝘰𝘶𝘳 𝘓𝘔𝘚 𝘏𝘰𝘭𝘥𝘪𝘯𝘨 𝘠𝘰𝘶𝘳 𝘋𝘢𝘵𝘢 𝘏𝘰𝘴𝘵𝘢𝘨𝘦? 𝘜𝘴𝘦 𝘋𝘦𝘭𝘦𝘨𝘢𝘵𝘦𝘴 𝘵𝘰 𝘗𝘭𝘢𝘺 𝘹𝘈𝘗𝘐 𝘊𝘰𝘯𝘵𝘦𝘯𝘵 & 𝘍𝘳𝘦𝘦 𝘠𝘰𝘶𝘳 𝘋𝘢𝘵𝘢  If you’ve ever felt limited by your LMS, this one’s for you.  Whether you're exploring GenAI, rethinking instructional design, or trying to unlock better data from your learning tech stack, there’s something here for you. 👉 Registration links for each session are in the comments. Stay curious! #LITSpotlight #LearningInnovationTechnology 

Educators: Empowered, Not Replaced "The tools are thoughtful... our teachers feel empowered." ✨ That is a direct quote from the schools we partner with. There is a common fear that AI will replace the human element in education. At BloomBridge, we’re proving the opposite. Our goal is to Reduce Burden & Strengthen Connection. By automating the "heavy lifting" of data entry and scheduling, we give you the most precious gift of all: Time. Time to mentor, time to inspire, and time to lead. Ready to see how our tools integrate with your current LMS? Let Us Talk And Visit Our Website : bloombridgecorp.com #TeacherEmpowerment #SustainableTech #EdReform #BloomBridge #HumanCentered

There’s a lot of conversation right now about AI, learning platforms, and new tools. But one question keeps coming up for me. Are our learning systems actually built for the world we’re moving into? Many organisations are investing heavily in new technologies, yet the real challenge often isn’t the tools themselves. It’s the infrastructure behind them — how systems connect, how data flows, and how learning experiences move beyond a single platform. That’s exactly what we’ll be exploring in our next xAPI Collective session. We’re joined by Shelly Blake-Plock, from Institute for Infrastructure and Interoperable Data in Learning, for a conversation on: Beyond the LMS: The Future of Learning Infrastructure This isn’t a technical deep dive. It’s a forward-looking conversation about how learning ecosystems are evolving, and why infrastructure and governance are becoming some of the most important conversations in learning and development. 📅 27 March 2026 ⏰ 11:00 AEDT If you’re curious about where learning is heading and how infrastructure quietly shapes everything we do, you might enjoy this conversation. Register here: https://lnkd.in/gxZkr3zd #xAPI #LearningData #LearningInfrastructure #LearningAnalytics #EdTech #FutureOfWork #DigitalLearning #ThinkOutsideTheLMS

Most training programs focus on course completion, but real impact comes from performance improvement. If your e-learning strategy isn’t aligned with clear business outcomes, you’re missing the bigger opportunity. Start with measurable goals. Build learning for action, not just information. Track performance metrics, not just completions. Blend digital learning with practical application. Leverage AI to personalize the experience. When learning is designed with purpose, it doesn’t just educate, it drives measurable business growth. #Elearning #LearningAndDevelopment #CorporateTraining #BusinessPerformance #DigitalLearning #AIinLearning

𝐖𝐡𝐞𝐧 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠 𝐃𝐞𝐬𝐢𝐠𝐧 𝐒𝐭𝐨𝐩𝐬 𝐁𝐞𝐢𝐧𝐠 𝐀𝐛𝐨𝐮𝐭 𝐂𝐨𝐮𝐫𝐬𝐞𝐬 Many organizations still approach learning as if the primary task were to produce courses. When a new need appears, the instinctive response is to build another module, another training program, another set of slides. The workflow is familiar: content is created, delivered, completed, and recorded. But beneath that routine, something more structural is beginning to shift. In many workplaces, the real challenge is not the absence of courses. It is the absence of well-designed learning processes. People do not struggle because information is unavailable; they struggle because the path from understanding to performance has not been designed clearly enough. When learning is viewed as a course, the work of the instructional designer revolves around content. When learning is viewed as a process, the focus changes. The designer begins to think about decisions, practice, feedback, and the sequence of experiences that help someone move from a starting point to competent performance. This distinction becomes even more visible as organizations experiment with AI tools. Technology can accelerate the production of content, but it cannot compensate for the absence of a coherent learning process. Faster slides or scripts do not necessarily produce better learning outcomes. In that sense, the most interesting shift happening in learning design today may not be technological at all. It may be conceptual. The question quietly emerging in many teams is no longer how to create better courses, but how to design learning processes that actually lead to performance. If you are working at the intersection of AI and learning design, I’d love to hear how you are approaching it in your organization. 𝐈𝐧 𝐁𝐥𝐚𝐜𝐤 & 𝐖𝐡𝐢𝐭𝐞 — 1/6 Reflections on learning systems. More thoughts on AI, learning ecosystems, and human-first design in the next post. #InstructionalDesign #LearningExperienceDesign #AIinLearning #FutureOfLearning #HumanCenteredLearning

This is the warm fuzzies I needed on a Friday morning. With the way AI is going, it’s hard for anyone to keep up. Prepping for each AI Accelerator cohort means there’s not much time to step back and hear the difference that we’ve made. But taking these intentional pauses to recalibrate and learn what was effective is just part of effective learning design. Hear Chelsea Nelson speak about her experience with the AI Accelerator Certificate program. Chelsea was incredibly engaged, laughed at all of our facilitators jokes (bonus points!), and came ready to experiment. She walked away with prompting skills and strategies that she can wield across different tasks, and to hear her tell it, she’s been able to save 10 hours a week. Those 10 hours can be put to high-priority work that we’ve classified in our human-AI delegation framework. You can read more about that in the comments. — If you or your team are looking for guided AI training that enriches your skillset and gives you free reign to explore solutions relevant to your organization, consider the AI Accelerator Certificate!

“This saved me ~10 hours (and it was just lesson one).” 🤩 One of our AI Accelerator Certificate alumni Chelsea Nelson talked about a remarkable practical win: She had a recurring pain point: pulling reports from her LMS: time-consuming, inconsistent, and always rebuilt from scratch. 😵💫 After our first prompt-engineering lesson, she created a standardized prompt that reliably produced the same reporting output every time, making reporting dramatically faster (she estimates ~10 hours saved). She also highlighted something we care deeply about: a safe, supportive learning community to ask questions and learn together. 🪄 If you work in L&D / instructional design, what’s the one workflow you’d love AI to standardize for you? Follow Artha Learning’s AI Accelerator Certificate page and stay updated with insightful posts and cohort announcements! (If you want the program details, comment “AI” or your email and we’ll DM you what to expect + the next cohort updates.) #LearningAndDevelopment #InstructionalDesign #LMS #WorkplaceLearning #AIinLND #PromptEngineering #ReportingChallenges #AICommunity #WorkflowAutomation

The best thing I saw at Training Magazine Network’s annual conference was not AI. It was generosity. A vendor offered me temporary sandbox access to their LMS. Not a guided tour. Not a controlled demo. Real access. That changed how I evaluate vendors. Partnership over pitch tells me you believe in your product. It says, “Go explore. See how it actually works.” That kind of access also lets you evaluate something most demos hide, whether the experience builds confidence or just tracks completion. The AI conversation was refreshing too. No one treated it like the savior of learning and development. The tone was grounded. AI can assist. Humans still design. Judgment still matters. One session demonstrated staged knowledge checks between sections of new content. You could not scroll and forget. You had to respond. And whether the answer was right or wrong, you received clear feedback before moving forward. You could feel the difference. Learners were progressing with confidence, not hesitation. Completion is easy to track. Confidence is harder to measure. Adoption depends on it. Senior L&D leaders, What is one red flag that tells you adoption will struggle? #LearningLeadership #LearningStrategy #LearningTechnology #PerformanceImprovement #LearningAndDevelopment

Your L&D function has a budget problem it doesn't know about yet. Not a cost problem. A relevance problem. ---- Recently, a research team built something called OpenMAIC — a Massive AI-powered Course system deploying multiple AI agents simultaneously in a single learning environment. Not one AI assistant. A full architecture. One agent facilitates. One challenges assumptions. One assesses quality of reasoning. One plays peer. All running in parallel, calibrated in real time to where each participant actually is — not where the curriculum assumes they are. The performance delta over single-expert delivery: +34%. Not from better content. From better interaction architecture. ---- Every L&D budget currently structured around facilitated programmes, cohort journeys, and completion-rate metrics is funding a design assumption that was already theoretically obsolete in 1934 — when Vygotsky demonstrated that learning doesn't flow from expert to receiver. It emerges from productive friction between a person and their zone of proximal challenge. The facilitator was never the centre of gravity. The interaction was. Your L&D function has been optimising for the wrong variable for two decades. ---- The shadow curriculum problem: Every L&D system runs two curricula simultaneously. The declared one: critical thinking, adaptability, complex judgment. The operating one: complete the module, pass the assessment, hit the rate. The operating curriculum always wins — because it's what the system measures. Multi-agent environments don't just change the technology. They force a reckoning with the measurement frame. When the system can scaffold every participant through their own zone of challenge simultaneously — with or without a facilitator — the question becomes: what exactly are you paying the facilitator to do? ---- The question CHROs should be asking their L&D heads this quarter: "Have you mapped our interaction architecture?" Not our content library. Not our LMS utilisation rates. Our interaction architecture — the conditions we've engineered for productive challenge, peer friction, assumption testing, and applied transfer. If the answer is a blank stare, you have a legacy system with a budget attached to it. ---- The organisations that will compound on learning investment in the next five years are not building better content. They are redesigning the conditions in which capability emerges. Content is infrastructure. Conditions are strategy.

Last week, I watched a teacher close her laptop… and sigh. Not because she didn’t love teaching. But because she had just spent hours uploading content to an LMS no one really uses. The system was “working.” But nothing was actually working. Students weren’t engaging.   Parents weren’t checking updates.   Reports were still being done manually.  And the question she asked stayed with me: “What exactly is this system helping me do?” That moment says everything about where we are in learning today. We have more technology than ever. More platforms.   More tools.   Even AI now sitting in the middle of it all.  But somehow… The experience still feels heavy. Because here’s the truth most people don’t talk about: You can have the best LMS.   The most advanced AI tools.   The cleanest dashboards.  And still have a broken learning experience. Why? Because the system was never designed for the people using it. So teachers work *around* it.   Students disengage *from* it.   Administrators struggle *with* it.  And over time… The platform becomes a place where content goes to sit. Not where learning actually happens. This is the shift I believe we need to make: From “What tool are we using?”   To “What experience are we designing?” Because when the system is right: Teachers feel supported   Students stay engaged   Data actually means something  And learning finally works the way it should. Now with AI entering the picture, this gap is becoming even more obvious. Because AI won’t fix a broken system. It will only expose it faster. So here’s the real question: Are we building systems people use…   or systems people avoid? I’d really like to hear this— Have you ever used a platform that looked perfect… but didn’t actually work for you? #EdTech #AIinEducation #LearningSystems #DigitalTransformation #InstructionalDesign #LMS #FutureOfLearning #EducationInnovation