BONUS How Scrum Masters Can Use AI Without Becoming the Human API With Fred Deichler

BONUS: How Scrum Masters Can Use AI Without Becoming the Human API AI is already changing the practical, everyday work of Scrum Masters and Agile Coaches. In this BONUS episode, Vasco talks with Fred Deichler about moving from curiosity to real AI-supported workflows: finding team signals faster, preparing better conversations, keeping documentation in sync, and staying focused on outcomes instead of just producing more output. From Automation to an AI Sparring Partner "I have this partner I can work with to help me ideate things." Fred’s journey into AI started before the current AI wave, with automation in Jira and the practical need to surface bottlenecks without manually watching every board. The turning point came when ChatGPT stopped feeling like a search box and started acting like a thinking partner. Faced with a team being pushed toward multiple sprint goals, Fred dumped the context into AI and asked for three to five options instead of one “right answer.” That shift helped him move from a deterministic mindset into a problem-solving mindset, using AI to explore options, challenge his own assumptions, and prepare a better conversation with the team. The Monday Morning AI Context Builder "It instantly sets that context for me. It sets that tone for the whole week." Fred describes his weekly workflow as a very practical use of AI: on Monday morning, he opens Cursor and works with his own AI harness, Atlas. Atlas is built from markdown files containing persona, skills, history, meeting transcripts, notes, and Jira data. When Fred says “good morning,” the system pulls the most relevant signals forward: aging work items, backlog health, sprint goals, and where the next team conversation should focus. Instead of starting the week by hunting for data, Fred starts with questions he can bring into the first stand-up: what is stuck, what needs refinement, and what risk is already visible? Making Flow Metrics Visible Inside Jira "If you're on a page, what do you hope you could learn without asking?" Beyond using AI as a thinking partner, Fred used AI to build a Chrome extension that surfaces useful Jira insights directly where the team already works. On the active sprint board, it shows work in progress, aging items, sprint goals, and sprint changes. In the backlog, it exposes backlog health and epic health. On the sprint report page, it adds cycle time per item so the retrospective can move from generic discussion to concrete learning. The point is not to shame the team with metrics. The point is to make the right conversation easier to start: what caused this item to take seven days, what blocked it, and what do we want to learn from that? AI as Extra Eyes and Ears for Team Conversations "It helps make sure that we don't lose sight of these things I can bring back to the team." Fred also uses AI agents to review meeting transcripts and look for patterns that are easy to miss in the flow of daily work. The system can notice hesitation, unresolved topics, or a requirement problem that was mentioned once and never followed up. For retrospectives, Fred’s Atlas setup can review sprint transcripts around day nine of the sprint and suggest themes for the next retro. It can even generate prompts for a visual retrospective board in Copilot or a Miro board from a prompt. This helps Fred avoid relying only on recency bias and creates a better starting point for the conversation the team actually needs. Keep the Human in the Loop, Especially for Outcomes "It's not about building a faster hammer. It's about identifying the outcome you're going for." A major warning in the episode is that AI can make Scrum Masters faster at producing more of the same: more notes, more summaries, more documents, more reports. Fred argues that the useful question is not “can we automate this exact process?” but “what outcome are we trying to achieve?” In a Jira-to-Azure DevOps migration, for example, the goal was not to reproduce the current time-tracking process perfectly. The goal was to provide the information accounting needed. That outcome focus keeps the human accountable for direction, judgment, and value, while AI helps explore better ways to get there. The SPOT Framework for Finding AI Opportunities "If it meets three or four of those, this is a great opportunity for an automation to free me up to do that more human-centric work." Fred uses a simple filter, the SPOT framework, to decide what AI should help with and what should remain human-led. A task is a good candidate when it is simple, predictable, observable, and tedious. Simple means it requires low human judgment and could be explained on an index card. Predictable means it happens repeatedly, either on a schedule or triggered by an event. Observable means the needed data is available to the agent and not locked away in...