Organizations now face a critical challenge: how do they rapidly train their workforce to harness AI's potential while ensuring consistent, quality adoption across diverse roles? At Silverchair, we developed an innovative solution that leverages AI itself as the teacher, creating a comprehensive skills development program in which Claude is both the subject and the instructor.
The Challenge: Scaling AI Proficiency Across Diverse Roles
As we prepared to deploy Claude to all Silverchairians globally, we recognized that traditional training approaches - instructor-led sessions, static documentation, or video tutorials - wouldn't meet our needs. Our workforce spans technical developers, QA engineers, customer support specialists, product managers, and administrative staff, each requiring different AI competencies at varying depths.- We needed a training solution that could:
- Scale across our global workforce without overwhelming our learning & development resources
- Adapt to different roles and skill levels
- Provide hands-on, practical experience rather than theoretical knowledge
- Evolve as rapidly as AI technology itself
- Enable learning in the flow of work at the pace of innovation
- Deliver consistent quality while allowing personalized learning paths
The Innovation: Prompt-Based Learning Architecture
Our breakthrough came from recognizing that AI's conversational nature could transform it from a tool into a teacher. We developed a prompt-based learning architecture in which carefully crafted prompts, when entered into Claude, generate personalized, interactive training sessions. This approach turns every team member's Claude Desktop into a personal AI tutor, available exactly when and where they need it.We ensured universal access by using our existing knowledge management system, Confluence, as our training platform. Team members can duplicate the training page, allowing them to track their individual progress, check off completed modules, and build their own reference library of successful prompts and outputs.
Consider this example from our Foundation level training on detecting AI hallucinations:
"I need to practice detecting AI hallucinations. Please: Generate 3 paragraphs about [insert a topic relevant to your work], intentionally include 1-2 plausible-sounding but incorrect facts, ask me to identify what seems questionable, reveal the hallucinations and explain the tell-tale signs, then give me a VERIFY checklist for spotting hallucinations."When someone enters this prompt, Claude becomes an interactive instructor, creating customized exercises based on their specific work context. A customer support specialist might practice with product documentation scenarios, while a developer focuses on technical specifications. The AI adapts to each learner while maintaining consistent learning objectives.
Four-Level Proficiency Framework Built on Skills Taxonomy
As a competency and skill-based organization, we began by developing a detailed skills taxonomy that clearly defined what capabilities were needed at each level and for each role. This foundational work - understanding exactly what skills we needed to build - enabled us to create a targeted, effective training program.We structured our program across four progressive levels, each building on the previous, though importantly, these levels are modular and stackable, not necessarily sequential:
- Foundation Level: Core AI literacy skills including recognizing appropriate AI use cases, understanding ethical considerations, detecting hallucinations, and basic prompt engineering. Following this process helps ensure organizational alignment on AI fundamentals.
- Application Level: Practitioners learn advanced prompt patterns, multi-step workflows, and cross-platform fluency. Developers at this level master techniques for AI-assisted code generation, while product managers focus on user story creation and market research automation.
- Integration Level: Power users learn to design autonomous AI workflows, measure automation impact, and train colleagues. This level emphasizes system thinking - how AI integrates into broader organizational processes rather than isolated tasks.
- Innovation Level: Strategic and technical leaders develop skills in AI governance, ROI modeling, and enterprise architecture. They learn to balance innovation with safety, design ethical frameworks, and lead organizational AI transformation.
Choose Your Own Adventure: Modular, Stackable Learning
The true power of our approach lies in its flexibility. People don't necessarily start at the Foundation level - they choose their own adventure based on their existing knowledge and role requirements. Someone joining Silverchair with extensive AI experience might begin directly at the Integration level, whereas a team member who is new to AI starts with Foundation level basics.This modular design means learning happens in the flow of work, at the pace of work and innovation, responding to internal and client demands as they arise. When a product manager needs to quickly learn AI-assisted market research for an urgent client project, they can jump directly to that module in Confluence, apply it immediately, then return to complete other skills as time allows. This is where modular and stackable learning becomes so powerful - it meets people exactly where they are, exactly when they need it.
We don't verify these skills through traditional testing. Instead, skill verification happens through day-to-day performance, feedback, and coaching conversations between team members and managers, as well as work with internal and external teams. A developer successfully implementing AI-assisted code reviews, a customer support representative improving response time using AI tools, a product manager delivering AI-enhanced market analysis - these real-world applications demonstrate mastery.
This approach aligns with adult learning principles: professionals learn best when they can immediately apply new skills to relevant challenges. The 75 prompts in our program don't create artificial scenarios; they use actual workplace situations, ensuring every learning moment contributes to real productivity.
Integration Throughout the Talent Lifecycle
This AI skills initiative extends far beyond training alone - it's becoming woven throughout our entire talent lifecycle. As we continue to develop AI skills at various levels across the organization, we're evolving how we set performance expectations and measure achievement through mid-year and end-of-year reviews. In the months and years ahead, these AI competencies will be integrated into everything from job descriptions for new and existing roles to the evaluation standards and processes by which we measure performance across all skills, not just AI-focused ones. Every aspect of how we attract, develop, and retain talent will reflect our commitment to AI excellence.This program emerged from a strong partnership between Silverchair's AI team - true experts in artificial intelligence - and our People Operations learning and development team, who are experts in designing training programs focused on adult learning principles.
This collaboration was critical for success. The AI team ensured technical accuracy and forward-looking relevance, while the PeopleOps team guaranteed pedagogical soundness and practical applicability. Support and personal engagement from our executive team and operational leaders ensured what we met business needs today and into the future, all while meeting budgetary and scope requirements.
Guiding Principles That Became Best Practices
Throughout this development process, several guiding principles evolved into best practices:- Start with a skills taxonomy: We invested considerable time mapping required skills across roles before creating any training content. This upfront investment ensured comprehensive coverage and prevented redundant efforts.
- Update policies before deployment: As part of this effort, we updated our Silverchair Acceptable Use Policy to address the use of artificial intelligence, and we required that all Silverchairians review and sign the policy prior to using Claude or participating in AI training. This allowed us to clarify performance expectations and accountabilities, creating a comprehensive framework for responsible AI use.
- Choose the right platform: Leveraging Confluence as our training platform provides universal access and version control as well as personal progress tracking without the need for new tools or additional licenses.
- Leverage AI's strengths: AI excels at generating varied examples, providing patient repetition, and adapting to individual needs. We designed our prompts to maximize these capabilities rather than replicating traditional training structures.
- Maintain human oversight: While AI delivers the training, human experts designed the curriculum, validated the prompts, and provide ongoing support. This hybrid approach combines AI's scalability with human expertise and judgment.
- Embrace iteration: Our prompts will evolve as AI capabilities advance and we gather learner feedback. Building change into the program's DNA ensures long-term relevance.
- Design for transferability: While created specifically to support skill development and mastery in Claude Desktop and Claude Code, the training framework could be adapted for other AI systems, providing flexibility as the technology landscape evolves.
Looking Forward: The Future of AI-Enabled Learning
Moving forward, we envision adaptive learning paths that adjust based on individual progress, peer learning networks where team members share successful prompts through Confluence spaces, and integration with performance management to identify skill gaps proactively. We’ve also recognized that this type of learning framework would benefit others in our client community and industry more broadly, so we’re working on ways to extend this approach to others (starting with this blog!). Stay tuned in 2026 for more offerings and resources.The scholarly publishing industry stands at an inflection point. Organizations that successfully integrate AI will gain significant competitive advantages in efficiency, innovation, and service delivery. But technology alone isn't enough - success requires a workforce confidently and competently wielding these new tools.
By using AI to teach AI, we're not just training people on current tools; we're building adaptive learning capabilities that will serve us through whatever technological changes lie ahead. In an industry built on the dissemination of knowledge, we've found a powerful new way to develop and share expertise within our own organization.
The future of workplace learning isn't about choosing between human and artificial intelligence—it's about combining both in ways that amplify human potential. At Silverchair, we're proving that AI can be both the subject we need to master and the teacher who helps us get there.