---

The Challenge

This was not a tools problem. Lowe's India GCC had already rolled out AI access across the team. What was missing was the thinking behind the prompting — how to frame a problem before you touch the keyboard, how to avoid hallucinations, how to bring structured frameworks to AI and reach decisions you could actually defend.

Using AI like a search engine is the floor, not the ceiling. The brief was to raise it: build the habits and mental models that turn AI access into genuine thinking leverage.

The Session

I designed a two-day experiential workshop and delivered it across three batches — roughly 22 professionals per batch, all with 3 to 10 years of experience. Minimal lecture. Participants did 80% of the work; every concept landed through an activity, not a slide.

Day 1 — Think Better with AI: Double Diamond and Lotus Blossom for framing and ideation. Problem vs. symptom — finding the real question before prompting. How to spot and avoid hallucinations. Scenario-based learning on real work problems, not textbook cases.

Day 2 — Apply and Decide: Framing questions to AI for better outcomes. Output vs. outcome vs. insight in real work contexts. Peer discussion — multiple perspectives on the same problem. Action planning — what specifically changes on Monday morning.

The scenario work was the hinge. Teams pressure-tested skills on problems that weren't their own — read the scenario, agree on a recommendation, commit to one card. Learning sticks when it's immediately recognisable, not when it's safely abstract.

What Happened

Day 2 outscored Day 1 in every single batch. That pattern held across all three. Here are the numbers:

Batch	Day 1 NPS	Day 2 NPS	Lift
Batch 1	8.50	9.20	+0.70
Batch 2	8.71	9.38	+0.67
Batch 3	8.00	9.11	+1.11

Not one participant flagged a low score across any session. Zero. 113 respondents, six sessions, three batches.

What participants said was most valuable: Day 1 — "Usage of AI as a thinking partner." Day 2 — "Framing questions to AI for better outcomes." Both are habits, not features. That's the point.

"No low score. It was fantastic." — participant, when asked what could be improved

The Result

Three batches delivered and the engagement is ongoing. On the strength of this work, Lowe's India GCC commissioned a second program — Digital Fluency and Mindset. Same client, new problem. That's the measure I care about: repeat trust, not repeat attendance.

Day 2 outscoring Day 1 consistently is also a signal worth naming. Participants rate the application day higher because they leave with tools they can use immediately, not frameworks to think about later. The design is working the way it's supposed to.

---

Get in touch See the games More case studies

What Happened in the Sessions

Six sessions ran across consecutive months. Each session covered a different cross-functional group — senior individual contributors, team leads, and function heads mixed together rather than kept separate by level. The mixed-level structure was deliberate: the collaborative dynamics between levels were exactly what the programme was designed to surface.

The game used in each session surfaced collaboration and cross-functional decision-making patterns. Within 20-30 minutes of play, consistent patterns began emerging: some participants built early alliances, others played defensively; some communicated proactively, others waited to be asked. By the debrief, participants could name their own patterns specifically — not because a facilitator told them, but because they had just demonstrated those patterns in the room.

The NPS trend across the six sessions is worth noting: the score improved from session one to session six. This happened without changes to the game or the debrief structure. Participants who had heard about earlier sessions came in with higher expectations — and the sessions met them. The social proof within the organisation did part of the preparation work.

What the Numbers Mean

An NPS of 8.82 from 113 respondents across 6 sessions is not a participation satisfaction score. NPS measures likelihood to recommend — it's a proxy for whether participants felt the experience was genuinely valuable, not merely enjoyable. At 113 respondents, the sample is large enough to be meaningful. At 8.82 out of 10, the variance is low: this was not a programme that split the room. It landed consistently across all six groups.

Day 2 outscoring Day 1 in every single batch is the result that Arvindh finds most significant. Day 1 typically benefits from novelty and energy. Day 2 is where participants who have slept on their insights return with more specific questions, deeper reflection, and more pointed application. Scoring higher on day 2 means the programme earned its value from learning quality, not opening-day enthusiasm.

Takeaways for L&D Professionals

• Consistent delivery quality across 6 sessions and 113 respondents is not an accident — it requires a facilitator who understands both the game mechanics and the debrief structure deeply enough to adapt without losing coherence
• Mixed-level participant groups surface richer cross-functional dynamics than homogeneous groups — level-based separation protects comfort at the cost of learning quality
• NPS improvement across a session series suggests positive word-of-mouth within the organisation — early participants are telling colleagues the session is worth attending, which raises the quality of engagement in later sessions
• A 90-minute to half-day serious game session can generate participant insight that multi-day workshops struggle to produce, because the mechanism is different: behaviour under pressure, not content delivered passively