In 2025, DAIN Studios joined the Harvard Data Science Initiative (HDSI), Harvard Data Science Review (HDSR) and Next Gen Learning (NGL) to design and deliver a new portfolio of executive programs on Generative AI and Agentic AI.
Across the first waves of this portfolio, more than 1,000 participants have learned with a combination of live Harvard faculty, DAIN’s practical frameworks and an AI tutor called Paski. The flagship program is the Agentic AI Intensive, which is now running its fourth cohort. The most recent cohort closed with an NPS of 87, with participants spending roughly half of their learning time in one-to-one tutorials with Paski.
Participants range from senior managers to C-level leaders in large organisations, plus a smaller group of founders and solopreneurs, mainly based in North America. Three patterns keep coming up: leaders need a clear framework for AI decisions, they create value when they redesign workflows instead of chasing tools, and they are starting to change how they learn in order to keep up.
1. Leaders need a framework for AI decisions
Most leaders who join the intensive do not lack interest or ideas. They arrive with long lists of potential use cases, scattered pilot projects and a mix of excitement and unease. What is often missing is a simple way to decide where to invest, what to pause and how to explain those choices to boards and regulators.
In the Agentic AI Intensive, that structure comes from DAIN’s A.G.E.N.T. Framework. It gives leadership teams a backbone for Agentic AI decisions: audit current workflows and roles, gauge value and complexity, engineer agent-first processes, navigate human–agent collaboration and track impact over time. Discussions that begin as broad “AI opportunity” talks quickly move into concrete trade-offs: which workflows are in scope, which risks are acceptable, which data and controls need to be in place before anything goes live.
At the start of the course many participants ask questions such as “How do we use agents in customer service?”. By the end they are discussing one specific workflow, what stays with humans, what is delegated to agents, and what level of autonomy is acceptable.
2. Workflow redesign is where value actually appears
The second lesson is that value and risk from Agentic AI show up in workflows, not in slide decks or tool catalogues.
Many participants arrive with tool questions. Which platform to choose, which vendor has the best features. By the end of the intensive, the questions have shifted. Leaders are asking which specific workflows they want to redesign, who will own them and how roles and controls will change.
In the Agentic AI Intensive, each participant picks one workflow from their own organisation and redesigns it for agent-first operation. Paski, the AI tutor, guides them through this process using DAIN’s strategy and workflow frameworks. Faculty then review and challenge the designs in live sessions, focusing on roles, decision points, data requirements and governance.
By the end, participants leave with a documented AI readiness or strategy outline for their area, a redesigned agentic workflow for one selected process, and a clear view of key risks, dependencies and next steps. For many organisations this becomes the first concrete, end-to-end workflow that bridges from pilots to day-to-day operations. The outcome is a business workflow they are ready to take back into their organisation, instead of only a certificate.
3. Leaders are changing how they learn
The third lesson is about leadership habits. Boards and executive teams expect their organisations to move faster on AI. That expectation now reaches the leaders themselves. They need to change how they learn, alongside what they learn.
The human plus AI model used in the Harvard portfolio points to one answer. Participants engaged strongly with the AI tutor. In the first intensive alone, they spent about nine hours of a single week learning with Paski. For leaders, the value is straightforward: Paski sits in the middle of the learning path as a planned working session after live inputs, and it knows each participant’s role and goals. This makes it easier to stay focused on their own agenda instead of generic examples.
At the same time, the curriculum stays stable. Only a small share, roughly 5–10 percent, is personalised by the AI tutor for each participant. The rest of the content remains shared. This gives leadership teams a common language and set of concepts they can use internally, while still allowing individual leaders to go deeper where their own responsibilities require it.
From cross-industry to healthcare
The early intensives were deliberately cross-industry, with participants from sectors such as healthcare, finance, manufacturing, telecom, technology and the public sector. This diversity broadened discussions around value, risk and operating models. It also led to a clear signal from leaders who wanted more depth within their own domain.
On the back of this demand, HDSI, HDSR, NGL and DAIN Studios are extending the portfolio into domain-focused tracks, starting with healthcare. A dedicated healthcare-focused AI course is already open for registrations, building on the same core curriculum while addressing the specific constraints and opportunities of healthcare organisations:
https://live.hdsrcourses.org/agentic-ai-healthcare-intensive
A laboratory for A.G.E.N.T. and BUILD
For DAIN Studios, these collaborations with Harvard and NGL function as a laboratory for Agentic AI. The cohorts pressure-test the A.G.E.N.T. Framework with a demanding audience that has limited time and clear accountability. The work highlights which parts of the framework resonate immediately and which steps need clearer language or stronger examples.
The programs have also shaped thinking on implementation. Alongside A.G.E.N.T. DAIN is developing the BUILD Framework for bringing agentic workflows into production, partly inspired by how participants wanted to move from design into delivery during course discussions. Insights from these programs feed back into client projects in Europe and the US, and the client work in turn strengthens what is taught in the courses.
More on the A.G.E.N.T. Framework: https://dainstudios.com/services/agent-framework/
What this means for leadership
The main outcome from teaching AI to more than 1,000 leaders with Harvard is a clearer view of what leadership teams actually need: they need a framework for AI decisions that they can defend in front of boards, regulators and employees. They need to see Agentic AI as a series of workflow and governance choices, instead of a pure technology purchase. And they need learning formats that respect their time while helping them make those choices faster and with more confidence.