Harvard Data Science Review Podcast Masterminds and Mindware for Agentic AI: Contextualized and Applied
9 snips
Jan 29, 2026 Ulla Kruhse-Lehtonen, co-CEO of DAIN Studios Finland, expert in data, AI strategy and agentic applications. Dirk Hofmann, co-CEO of DAIN Studios Germany, focused on applying AI to transform workflows. They discuss agentic AI as workflow-embedded agents with memory and guardrails. Conversation covers outcome-driven redesign, practical governance, course design for applied agentic systems, and reframing AI as augmentation.
AI Snips
Chapters
Transcript
Episode notes
DAIN's Origins At Nokia
- DAIN Studios began in 2015 as a trio from Nokia focused on translating business strategy into data and AI impact.
- They evolved into an end-to-end consultancy combining strategy and implementation across data roles.
Agentic AI Is Workflow-First
- Agentic AI works at the workflow level, not as isolated task tools embedded into processes and teams.
- Agents need memory, skills, clear inputs/outputs, and guardrails to function as reliable coworkers.
Design Around Outcomes, Not Outputs
- Start by defining the outcome you want, not the outputs the AI can produce.
- Design agentic solutions around measurable business outcomes to avoid wasted effort and disappointment.