AI Maturity Self-Assessment: A 30-Minute Diagnostic for Pharma IT Leaders

Why a 30-Minute Diagnostic

Comprehensive AI maturity assessments produce excellent strategic insight but they take weeks to run. The IMD AI Maturity Index evaluates organizations across executive support, technology and infrastructure, operational excellence, workforce and culture, and ethics and risk management — each requiring structured data collection. L.E.K.’s maturity framework, McKinsey’s diagnostic, ZS’s CDIO research, and the various consulting-firm frameworks are similarly comprehensive and similarly time-intensive.

The thirty-minute diagnostic is not a replacement; it is a complement. It produces a directional read fast enough to be useful in three specific situations: when a new IT leader needs to orient on AI maturity in the first week of a role; when an executive team needs a quick maturity read before a board meeting; and when an organization is deciding whether to commission a full maturity assessment and needs an initial read to scope it. In all three situations, the time cost of a comprehensive assessment is prohibitive and a directional read is sufficient.

The discipline of the thirty-minute diagnostic is in choosing the right questions. The right questions are those whose honest answers correlate with maturity across the dimensions that matter most. The wrong questions are those that produce confident answers regardless of maturity. The diagnostic below has been calibrated against the more comprehensive frameworks to focus on questions whose answers are diagnostic of maturity rather than performative.

The Six Dimensions That Matter

The diagnostic covers six dimensions. Each dimension is scored independently, and the scores aggregate into an overall maturity read. The dimensions:

The six dimensions are not independent; they reinforce each other. Strong data foundations enable better use cases; strong governance enables more sustainable scaling; strong talent enables better infrastructure decisions. The diagnostic measures each independently but recognizes that improvement typically happens in correlated clusters.

The Diagnostic Questions for Each Dimension

Each dimension has five questions, scored 1 (low maturity) to 5 (high maturity). The questions are designed to be answered honestly in roughly five minutes per dimension. The IT leader should answer them alone before discussing with the broader team; the honest individual read is more useful than the consensus answer.

Data foundations

1. Can you produce a current inventory of the data assets feeding AI use cases? (1: no, 5: comprehensive inventory with metadata)
2. Is there a documented data dictionary for each AI use case’s inputs? (1: no, 5: comprehensive, kept current)
3. Can you trace the lineage of regulated data through AI workflows? (1: no, 5: full lineage with documented controls)
4. Is there clear ownership and accountability for data quality? (1: no, 5: defined ownership with operational evidence)
5. Are data refresh cadences and drift monitoring operational? (1: no, 5: operational with defined response procedures)

Governance

1. Is there a cross-functional AI governance committee with defined decision rights? (1: no, 5: chartered with clear authority)
2. Is there an AI use case inventory with risk classifications? (1: no, 5: comprehensive and current)
3. Is there a defined validation methodology for AI in regulated workflows? (1: no, 5: documented and operationally embedded)
4. Is AI change control integrated with the existing QMS change control? (1: no, 5: fully integrated)
5. Is there a defined incident response procedure for AI-related incidents? (1: no, 5: operational with documented test of effectiveness)

Talent and capability

1. Does the organization have an internal AI platform team? (1: no, 5: sized appropriately for scale with defined responsibilities)
2. Is there a workforce upskilling program for AI capability? (1: no, 5: operational with measurable outcomes)
3. Can the organization attract and retain AI talent it has hired? (1: high turnover, 5: stable retention)
4. Do business leaders understand AI well enough to make informed decisions? (1: no, 5: broadly true across the leadership team)
5. Are external partnerships used effectively for capability the organization cannot build? (1: no, 5: strategic partnerships with clear value)

Infrastructure

1. Does the organization have compute capacity sufficient for current AI use cases? (1: no, 5: adequate with reasonable headroom)
2. Is there a documented reference architecture for AI in the organization? (1: no, 5: documented and operationally followed)
3. Is there MLOps tooling appropriate to the AI portfolio? (1: no, 5: operational across the portfolio)
4. Are AI capabilities integrated cleanly with existing systems (MES, LIMS, eTMF, etc.)? (1: no, 5: integrated with documented patterns)
5. Is observability adequate to detect AI performance drift in production? (1: no, 5: operational with response procedures)

Use case portfolio

1. How many AI use cases are in production (not pilot)? (1: zero, 5: meaningful portfolio with diverse uses)
2. Are the use cases producing measurable value? (1: no, 5: value measured and tracked)
3. Is the portfolio managed against a prioritized roadmap? (1: no, 5: roadmap is current and operational)
4. Are pilots converting to production at a reasonable rate? (1: pilots accumulate, 5: defined conversion gate)
5. Is there a mechanism for retiring underperforming use cases? (1: no, 5: operational with documented decisions)

Executive sponsorship

1. Is there a named executive sponsor for AI strategy? (1: no, 5: clear with regular cadence)
2. Is capital commitment to AI sustained across budget cycles? (1: variable, 5: sustained commitment with multi-year visibility)
3. Is AI integrated into strategic planning? (1: no, 5: deeply integrated)
4. Are AI outcomes reported to the board? (1: no, 5: regular reporting with clear metrics)
5. Does the executive team defend AI investment against competing priorities? (1: no, 5: consistently)

Scoring and Interpretation

Each dimension produces a score between 5 and 25 (five questions, each scored 1-5). The dimension scores aggregate to a total between 30 and 150.

The score ranges map to maturity levels:

• 30-60: Early stage. AI activity may be present but is not yet operating as a coherent program. Investment in foundations is the priority.
• 61-90: Developing. Foundations are in place; the organization is producing AI value but is not yet operating at scale. Investment in scaling and governance is the priority.
• 91-120: Mature. The organization is operating AI as a coherent program with measurable value, sustainable governance, and competitive talent capability. Investment in differentiating capabilities is the priority.
• 121-150: Leading. The organization is among the most AI-mature in the industry. The priority is sustaining the lead and continuing to invest in the disciplines that produced it.

The score ranges are calibrated against the IMD AI Maturity Index and the patterns visible across the major consulting-firm frameworks. Organizations in the leading tier are typically among the seven major pharmas (AstraZeneca, Merck & Co., Eli Lilly, Novartis, Sanofi, Novo Nordisk, GSK) identified as top 100 most AI-mature firms; the developing and mature tiers are where most pharma organizations actually live.

Honest scoring matters. The diagnostic is only useful if the answers are honest. Performative answers (“we have a governance committee, so 5”) produce maturity reads that miss real gaps. Honest answers (“the committee meets but has not made a substantive decision in six months, so 2”) produce reads that drive action.

The Three Maturity Archetypes

Beyond the numeric score, organizations typically fit one of three maturity archetypes. The archetypes are useful because they suggest different priority actions even at similar score levels.

The foundation-strong organization. Strong data foundations, infrastructure, and talent; weaker use case portfolio and executive sponsorship. The foundations are built but the organization has not yet translated them into operational value. The priority is converting foundations into use cases and demonstrating value to sustain executive sponsorship.

The use-case-rich organization. Active use case portfolio with visible value; weaker data foundations, governance, and infrastructure. The use cases are producing value but the underlying disciplines are catching up. The priority is investing in foundations to support sustainable scaling without disrupting the producing portfolio.

The governance-mature organization. Strong governance, executive sponsorship, and risk discipline; weaker talent, infrastructure, and use case portfolio. The organization is well-positioned for sustainable AI investment but is not yet executing at scale. The priority is investing in talent, infrastructure, and use case development to translate governance maturity into operational AI value.

Most organizations fit one of these archetypes more than the others. The archetype is more useful than the numeric score for setting priorities, because it identifies what is over-developed versus under-developed in the maturity profile.

What to Do With the Result

The diagnostic is most useful when it drives specific action. Three uses produce real value.

Identify the lowest-scoring dimension and invest there. The lowest-scoring dimension is typically the binding constraint on overall maturity. Investment in that dimension produces disproportionate maturity gains.

Validate the archetype against operational reality. Discuss the archetype assessment with the cross-functional team. The discussion typically surfaces whether the archetype is real or whether the scoring distorted the picture.

Calibrate the investment ask against the maturity stage. Organizations at the early stage need foundational investment; organizations at the developing stage need scaling investment; organizations at the mature stage need differentiation investment. The investment ask the IT leader brings to the executive team should match the maturity stage.

The diagnostic is also useful for tracking progress over time. Re-running it quarterly produces a maturity trajectory that is more useful than any single point-in-time read. Maturity trajectories that flatten or decline are early warnings; trajectories that climb steadily indicate that the investments are producing results.

Using the Diagnostic With Leadership and the Board

The diagnostic is particularly useful in conversations with the broader leadership team and the board. Three patterns work well.

Sharing the dimension scores rather than the total. The total score is useful for tracking; the dimension scores are useful for action. Leadership discussions oriented around dimension scores produce more substantive conversations than discussions oriented around the total.

Comparing the IT leader’s read against the cross-functional team’s read. Running the diagnostic with multiple stakeholders and comparing the results often surfaces gaps in perception that are themselves diagnostic. When the IT leader scores governance at 4 and the QA leader scores it at 2, the gap is worth understanding.

Using the trajectory rather than the snapshot. Boards and executive teams respond better to maturity trajectories (“we have moved from developing to mature over the last twelve months, with the most progress in data foundations and use case portfolio”) than to point-in-time snapshots. The trajectory framing supports continued investment by showing returns.

The discipline of using the diagnostic well is in remembering that it is a directional tool, not a precision instrument. The honest directional read is more valuable than a precise but performative read. As IMD’s analysis of AI trends in pharma emphasizes, the difference between organizations that capture AI value and organizations that do not is more about execution and organizational readiness than about technology choice. The diagnostic measures the readiness; the value comes from acting on the read.

When to escalate to a comprehensive maturity assessment

The thirty-minute diagnostic is sufficient for most ongoing maturity management. There are three situations where escalating to a comprehensive maturity assessment is worthwhile.

First, when a significant strategic decision turns on the maturity read. If the executive team is considering a major capital commitment or a transformative AI program, the diagnostic’s directional read should be validated with a comprehensive assessment before the decision is made.

Second, when the diagnostic produces results that the IT leader does not believe. If the diagnostic shows maturity that does not match the operational experience, a comprehensive assessment helps surface what the diagnostic missed.

Third, when external stakeholders (acquirers, partners, regulators) need the assessment for their own purposes. External stakeholders typically want comprehensive assessments rather than internal diagnostics; preparing for those engagements requires more rigor than the thirty-minute version provides.

In all other cases, the diagnostic is sufficient. Quarterly re-running, combined with disciplined action on the lowest-scoring dimensions, produces sustained maturity improvement without the overhead of comprehensive assessments.

The honest evaluation of executive sponsorship

One dimension worth dwelling on: executive sponsorship is consistently overestimated in maturity assessments. IT leaders rarely score executive sponsorship at 1 or 2 because doing so reflects poorly on the executive team and is uncomfortable to surface. The honest assessment is harder than the performative one, but it is also more useful. If executive sponsorship is genuinely weak, the priority action is rebuilding the sponsorship case rather than continuing to invest in capabilities that the sponsorship cannot sustain.

The signs of weak sponsorship: AI budget is variable across years rather than sustained, AI is dropped from strategic planning under pressure, AI investments are defended weakly when competing priorities emerge, and the executive sponsor cannot articulate the AI value proposition without prompting. These signs are diagnostic of low sponsorship maturity regardless of the surface artifacts (named sponsor, board reporting). The diagnostic should score the substance, not the surface.

The discipline of running the diagnostic on a defined cadence

The diagnostic produces the most value when it is run on a defined cadence rather than ad hoc. The recommended cadence is quarterly, with the trajectory across quarters used to evaluate whether the maturity investment is producing returns. Organizations that run the diagnostic once and never repeat it miss the diagnostic’s most valuable use, which is the trajectory rather than the snapshot.

The quarterly cadence has several practical benefits. It produces a regular forcing function for honest assessment. It surfaces emerging maturity gaps before they become critical. It provides the basis for sustained executive conversations about AI maturity rather than annual surge conversations tied to planning cycles. And it builds organizational fluency in maturity assessment that makes the more comprehensive assessments, when they are needed, easier to commission and interpret.

The discipline of running the diagnostic quarterly is also valuable in itself, regardless of the specific scores produced. The act of structured assessment, repeated on a cadence, builds the organizational muscle for evaluating AI maturity rigorously. Organizations that lack this muscle frequently default to performative assessments that miss real gaps; organizations that have built it through repeated practice produce honest assessments that drive action. The thirty-minute diagnostic is not just a tool for measurement; it is a tool for building the organizational discipline that sustainable AI maturity requires over the long term.

References & Sources