CELab: The Customer Education Lab

In this kickoff to Season 8 of the CELab podcast, host Dave Derington sits down with Anya Eydman to explore a provocative idea: our biggest AI problem in Customer Education isn’t a skills gap or a technology gap - it’s a belief gap. Anya draws on her background in corporate finance, small business and startup operations, coaching, municipal and nonprofit work, and energy/civic tech to show how AI is becoming a new learning style and a genuine learning partner, not just another productivity tool. Anya connects the dots between AI, underserved communities, workforce shifts, and what it means to design learning experiences that people actually want to engage with.

Across the conversation, they unpack the role of prompt engineering as a core literacy, not just a niche technical skill, and examine how AI can level the playing field while also exposing new inequities. Anya shares how she’s using AI with interns, cities, and universities (including a forthcoming “AI Without Hype” course at Vassar College) to build trust infrastructure, redesign workflows, and bring more people into complex conversations about energy, affordability, and technology. If you’re in Customer Education or Customer Success and wrestling with how AI changes your strategy, your stakeholders, and your learners, this episode will challenge how you think about outcomes, belief systems, and what “meeting learners where they are” really means in 2026.

Key Topics:

The “belief gap” vs. skills gap

Why resistance to AI is often about belief systems, not capability.

How this parallels learner motivation and adoption challenges in CE/CS.

AI as a new learning style and learning partner

Treating AI like a sharp intern or co-worker rather than a magic answer machine.

Designing learning journeys that assume AI is in the loop.

Prompt engineering as a core literacy

Why prompt engineering should be taught like a fundamental communication skill.

How better questions drive better learning outcomes and better customer outcomes.

Non-linear, outcome-focused learning journeys

Moving beyond “I took a course” to “we changed a workflow and a behavior.”

Aligning CE programs with measurable participation and impact, not just completions.

Cross-pollination beyond SaaS/B2B silos

Lessons from civic tech, energy, and community engagement that apply directly to CE/CS.

How to bring community-style thinking into customer communities and academies.

Underserved learners and access to AI

What “underserved” means when AI accelerates knowledge but not everyone can use it well.

How CE can help customers navigate fear, overload, and trust issues with AI.

Trust infrastructure and stakeholder mapping

Building trust frameworks inside organizations and with customer stakeholders.

Rethinking stakeholder mapping now that decision-making paths and gatekeepers are changing.

Team dynamics, titles, and shared ownership of knowledge

Shifting from heavy titles to everyone as a “lead” owning a slice of the missio...

In this episode of the Customer Education Lab, Georg Gottschalk, VP of Global Education at Acumatica, to unpack what it really takes to build a skill infrastructure around a highly complex ERP platform. Georg traces his journey from early work with SAP and manufacturing optimization, through leadership roles at IBM, Salesforce, and MuleSoft, to his current mandate: Enabling a fast-growing, partner-led ecosystem in the small and mid-sized business ERP space. Georg shares why ERP is one of the ultimate “stress tests” for Customer Education and why the old excuse of “our product is too complex for AI” no longer holds up.

The conversation dives deep into AI-assisted learning, micro-vertical partner strategies, and what it means to certify companies instead of just individuals. Georg explains how Acumatica leverages tools like LearnExperts’ LEAi to transform rich but dense documentation into fun, engaging, and increasingly personalized education at scale—without tripling headcount. You’ll hear how his team thinks about partner enablement, customer onboarding that massively exceeded adoption targets, weaving education into the product workflow, and even “teaching machines” by producing content optimized for LLMs. If you work in Customer Education or Customer Success and are wrestling with complexity, scale, or AI, this is an episode you won’t want to miss.

Here are some of the key insights from this episode!

AI works for complex products too — the "too technical for AI" excuse no longer holds; AI acts as a force multiplier, handling grunt work (first-pass quizzes, structuring, transformations) so SMEs can focus on quality.

Skill infrastructure at ecosystem scale — enabling customers, partners, consultants, and developers across a complex ERP, with tools like Leah powering engaging, personalized education.

Certifying companies, not just individuals — partner-level certification builds trust and signals capability, supporting a partner-led growth and micro-vertical strategy.

Closing the value gap — bridging what the platform can do and what users actually achieve through in-workflow, contextual guidance rather than course-based learning alone.

Hyper-personalization with telemetry — using product usage data to spot skill gaps and deliver "what users need tomorrow, not today."

Teaching machines and fostering experimentation — writing content readable by LLMs and skill files for AI agents, while leaders create psychological safety for teams to experiment, keeping humans in the loop.

In this episode of CELab, Dave Derington sits down with Tim Devine, who built Optro's (formerly AuditBoard) live training program from a solo effort in 2021 into a three-person team running roughly 250 custom training engagements a year. Tim's thesis is deceptively simple and a little provocative: you don't need to hire subject-matter experts to deliver expert training. You need people who are experts in the product and the path — and the discipline to stay there.

We all know the "one legendary expert who doesn't scale" problem, right? Tim approach is to hire trainers with no auditing background, coach them to "learn the language and trust the audience to fill the gaps," and keep a customer subject-matter expert on deck to field methodology questions. Guardrails — a prescriptive course catalog, recommended agendas, and tightly structured virtual sessions — aren't limitations; they're what make scale possible.

The conversation digs into Optro's "do-it-with-me" methodology, where a temporary sandbox is mirrored from the customer's own UAT data so learners click along live instead of watching a demo — a direct answer to the live-attendance problem. Tim also opens up the tech stack: Amplitude to correlate learning objectives with product outcomes, Gainsight health scores, and Skilljar as the LMS — plus a 13% renewal-rate lift that reframes training as revenue protection rather than a cost center.

On AI, Tim shares two practical wins: auto-scoping tailored learning paths from sales-and-implementation handoff data, and using Zoom's auto-chapters to make session recordings instantly navigable. Honest, tactical, and full of things you can steal on Monday.

In this episode of the Customer Education Lab, Adam Avramescu sits down with Kelly Mullaney, Head of AI Guild at Juniper Square, to unpack how AI is actually changing Customer Education. Kelly traces his journey from tax tech writer at Microsoft, to leading Customer Education at Envestnet, to building AI Guild—an internal enablement function focused on helping teams 10x their productivity with AI. Along the way, he shares concrete builds: an AI‑driven help center that watches code commits and auto‑drafts release notes and docs, a 37‑agent competitive intelligence system feeding Slack, and experiments in “vibe‑coded” e‑learning that challenge what LMSs and SCORM should look like in an AI‑first world.

Adam and Kelly go deep on the questions every Customer Education and Customer Success leader is wrestling with: Will AI kill our jobs, or amplify our impact? When does it make sense to build with AI versus buy another SaaS tool? How can small, scrappy teams of one keep up when engineering output increases 10x? You’ll hear a practical breakdown of the Anthropic stack (Claude Chat, Code Desktop, Cowork, Design, Managed Agents), a simple “80% AI / 20% human‑in‑the‑loop” mental model, and a candid look at the future of help centers, LMSs, and documentation. If you’re trying to move past the hype and see what AI‑enabled customer education really looks like, this episode is for you.

We talk about:

AI as the new “electricity”: why simply bolting AI onto old workflows fails—and what it means to rewire your education stack.

Kelly’s path: tech writer → running Customer Education at a $5B company → leading AI Guild at Juniper Square.

Human‑in‑the‑loop: using AI to get to 80% done and relying on experts for the final 20% quality bar.

AI‑driven help centers: monitoring Git repos for meaningful commits, pulling Jira context, applying style guides, and auto‑drafting release notes and help docs.

Fixed docs vs. AI assistants: why Kelly layers AI chat on top of a traditional help center instead of replacing it outright.

Discovering content gaps: using AI search and chat logs to identify missing documentation and new questions customers are asking.

Build vs. buy math: when building 80% of a SaaS tool in a weekend (with Claude + Supabase + agents) beats paying $30–60K per year.

Agentic competitive intel: 37 Claude agents monitor competitors, post daily briefs to Slack, and learn from a feedback loop.

Rethinking LMS and SCORM: legacy tools’ reporting limitations vs. what AI‑enabled tracking and analysis can unlock.

Demystifying Anthropic’s stack: how Claude Chat, Code, Cowork, Design, MCPs/connectors, and Managed Agents fit together.

Tool choice vs. governance: innovating even when your company only “allows” one AI vendor like Copilot.

Why CE leaders are primed for AI: curiosity, user empathy, and great questions matter more than coding backgrounds.

Getting started advice: don’t just “insert AI into the SDLC”—reimagine your education model for a 10x‑faster product team.

In this episode of the Customer Education Lab we sit down with longtime friend and customer education leader Lila Krutel at Skilljar by Gainsight to unpack how MCP (Model Context Protocol) and modern AI tools like Claude are reshaping the role of Customer Education in B2B SaaS. Together they explore how MCP acts like “arms” for your AI assistant—connecting it to systems like your LMS and Customer Success Platform so you can finally query, explore, and analyze your own data without waiting in the CS Ops queue.  

Lila shares concrete, in-the-trenches stories from Gainsight, including how she used Gainsight’s CS MCP plus Claude to identify customers with unused training passes, analyze the business impact of admin certification, and start connecting education programs to ARR and retention - all without writing SQL or begging for a special report. If you’ve ever felt blocked by data access, stuck waiting on BI dashboards, or unsure how to “prove” the value of education, this conversation will show you the art of the possible and give you practical ways to start experimenting—safely, skeptically, and out loud—with AI in your own CE or CS org.