# Conversational Agents Won’t Replace Training — They’ll Make It Work By [Kevin Alster](/content/blog/authors/kevin-alster/index.html) Strategic Advisor at Synthesia Create AI videos with 240+ avatars in 160+ languages ## **1. A Move Towards Capability Over Content** Most organizations optimize for content production...which is why the AI tools boom created so much hype. More courses. More decks. More playbooks. Expertise research shows that performance does not improve through exposure alone. Psychologist Anders Ericsson’s foundational research on deliberate practice demonstrates that expert performance emerges from highly structured, effortful practice with immediate feedback and clear performance goals — not from repetition or content consumption alone ( [_Deliberate Practice and the Acquisition of Expert Performance_](https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1553-2712.2008.00227.x?utm_source=chatgpt.com)). Deliberate practice requires: - Specific improvement goals - Work at the edge of current ability - Frequent feedback - Ongoing self-monitoring That is fundamentally different from completing a module. Conversational agents create the missing layer: repeatable, feedback-rich rehearsal. ## **2. The Training–Coaching Gap Is Now Measurable** The [Salesforce State of Sales (7th Edition)](https://www.salesforce.com/en-us/wp-content/uploads/sites/4/documents/reports/sales/salesforce-state-of-sales-report-2026.pdf) quantifies what many enablement leaders already know: - 52% of reps say traditional enablement doesn’t provide the skills they need - 41% don’t get enough opportunities to roleplay before customer calls - 46% rarely receive feedback on sales conversations - 40% say managers lack time to coach them properly This is the structural bottleneck. Reps want rehearsal. Managers lack capacity. Content alone cannot close the gap. Conversational agents are not a novelty feature. ## **3. Practice Must Be Embedded — Not Optional** A simulation sitting outside the learning journey will not change behavior. Deliberate practice works when it is: - Integrated into structured pathways - Tied to defined performance standards - Required before advancement - Measured against outcomes We see this across domains. Medical schools are now using AI-driven virtual patients to let trainees rehearse clinical interviews repeatedly with structured feedback ( [_Digital Docs: How AI is Helping Train Medical Students, Cornell University_](https://alumni.cornell.edu/cornellians/ai-med-students/#:~:text=At%20Weill%20Cornell%20Medicine%2C%20students,that%20approximates%20a%20telehealth%20visit.)). Students can iterate until they demonstrate empathy, diagnostic thoroughness, and clarity. When rehearsal becomes routine — not optional — skill accelerates. ## **4. A New 70–20–10 for AI Investment in HR** The temptation with AI is to start with tooling as we did in 2024/5. But research on AI transformation from Boston Consulting Group shows that organizations generating real value invest roughly: - 70% in people and process redesign - 20% in technology and integration - 10% in algorithms Agents multiply clarity. They do not create it. If performance standards are vague, AI will scale vagueness. Redesign before deployment. ## **5. The Strategic Implication** Conversational agents will not replace traditional training. They complete it. They create the bridge between: - Structured instruction and Live human coaching - Scenario-based rehearsal - Immediate, objective feedback - Purposeful iteration - Workflow-embedded reinforcement ### References 1. [Ericsson, K. A., Krampe, R. T., & Tesch-Römer, C. (1993). _Deliberate Practice and the Acquisition of Expert Performance._ Psychological Review, 100(3), 363–406.](https://onlinelibrary.wiley.com/doi/10.1111/j.1553-2712.2008.00227.x) 2. [Synthesia x Cornerstone Webinar: AI for Upskilling with Conversational Agents](https://hubs.ly/Q041yDYh0) 3. [Harvard Graduate School of Education. _The Value of Self Reflection_](https://www.gse.harvard.edu/ideas/usable-knowledge/14/10/value-self-reflection) 4. [Salesforce Research. (2026). _State of Sales, 7th Edition_](https://www.salesforce.com/en-us/wp-content/uploads/sites/4/documents/reports/sales/salesforce-state-of-sales-report-2026.pdf) 5. [Cornell University. _Digital Docs: How AI is Helping Train Medical Students_](https://alumni.cornell.edu/cornellians/ai-med-students/) 6. [Boston Consulting Group. _The Leader’s Guide to Transforming with AI_](https://www.bcg.com/featured-insights/the-leaders-guide-to-transforming-with-ai) ## Kevin Alster Kevin Alster is a Strategic Advisor at Synthesia, helping enterprises apply generative AI to learning, communication, and performance. With over a decade in education and media, he’s built programs for General Assembly, NYT School, and Sotheby’s.