Blending design strategy, lean thinking, and AI in enterprise innovation

Executive overview

Enterprise teams default to building fast, but skipping deep customer understanding produces the wrong thing. Combining design strategy (ethnographies, structured interviews) with lean experimentation front-loads learning, so validation cycles run faster overall.

Sergio Gutierrez-Montero applies this at AbbVie as Director of AI Strategy — starting always from the problem, not the technology. AI enters only when it genuinely fits the solution.

Slow preparation enables fast execution: deeper upfront research produces stronger hypotheses and compresses the total time to validated insight.

From design strategy to lean experimentation

• Design strategy and lean thinking are complementary, not competing — combining them yields stronger hypotheses before any test begins.
• Ethnographies and depth interviews inform the problem space; experimentation validates the solution and growth hypothesis.
• In B2B niches, qualitative relationship-based research outperforms high-volume landing-page A/B tests.
• "Just enough research" (Erica Hall) is the target: diminishing returns kick in well before 100 hours of discovery.
• Early career lesson from Steelcase: scrappy A/B tests fail in low-traffic B2B contexts; direct customer relationships work better.

Maintaining a beginner's mindset

• Before field research, map all current assumptions explicitly — this makes them easier to challenge and discard.
• Frame insights as "we thought X; we learned Y" to make the gap visible and concrete.
• Avoid debating methodology with stakeholders; show results instead.
• At McKinsey, "day one answer" culture is the opposite of beginner's mindset — Sergio navigated it by surfacing market signals that updated the hypothesis rather than arguing process.
• Framing assumption-testing in terms of risk ("what's the cost of building the wrong thing?") lands better than methodology arguments.

Managing stakeholders who want to move fast

• Give anxious stakeholders something tangible: a vision prototype, clearly caveated as a hypothesis under validation.
• Tools like V0 and Lovable make clickable prototypes fast enough to satisfy stakeholder urgency while still enabling real testing.
• Parallel-track the hypothesis-driven work and the exploratory work; let results do the persuading.
• Stop trying to "die on the hill" of process — graceful influence beats head-on confrontation.

AI strategy grounded in problem-first thinking

• The ratio in Sergio's current role: mostly lean and design methodology, with AI considered only once the problem is well-framed.
• Many pain points are solved by better processes or workflows, not AI.
• Hypothesis-driven consulting gets you 60–70% of the right answer; the differentiated last 30% requires deeper, more open-ended inquiry.
• "I couldn't care less about AI — I care about the problem space and the right solution."