Chapter 06
Research dimension: Dynamic Capabilities
How should workflows change for AI?
4 min readPublished Published by Teklens
Workflows should change with every experience, not once at rollout. The research calls this adaptability, Dynamic Capabilities: sensing signals, seizing opportunities with budget, reconfiguring structures and learning continuously. In the DACH analysis it is the lowest dimension at 2.61 of 5. Adapting structures and processes reaches 2.47, the lowest value of all 19 items. At the same time no dimension separates advanced from lagging organisations more sharply.
Contents · Short answer
Three findings
- 01
The lowest score in the study
Q18, adapting workflows, responsibilities and AI solutions, sits at 2.47 with a median of 2. Level 2 reads "project-based thinking: workflows and roles are only adapted under acute pressure".[Probst 2026, Appendix B]
- 02
The sharpest divide between the profiles
Advanced 4.13, intermediate 2.66, early stage 1.84: a gap of 2.29 scale points and the highest F value of all dimensions (83.1). The interquartile ranges of the three profiles do not overlap.[Probst 2026, Figure 8]
- 03
Large organisations are overrepresented in the early stage
20 of the 25 organisations in the early stage have 250 or more employees, although large organisations make up only 59.4% of the sample. The thesis reads this as a hint that resources do not replace adaptability.[Probst 2026, Appendix E]
Evidence
- Q16 Sensing2.63
- Q17 Seizing2.72
- Q18 Reconfiguring (Structures)2.47
- Q19 Reconfiguring (Learning)2.64
Select an item to see its wording, levels and statistics.
Takeaway: Adapting structures and processes (Q18, 2.47) is the lowest score of all 19 items. Three of the four items have a median of 2.
- Unit
- Mean on the maturity scale 1 (least developed) to 5 (most advanced)
- Population
- DACH analysis: 64 cleaned self-assessments from Germany and Switzerland, mostly senior decision-makers
- Denominator
- n = 64 organisations
- Source
- Probst 2026, Appendix B · Appendix B – Mean, standard deviation, median, minimum, maximum and item-rest correlation of all 19 items
Data as a table
| Item | Dimension | M | SD | Median | Min | Max | r (item-rest) |
|---|---|---|---|---|---|---|---|
| Q16 Sensing | Adaptability | 2.63 | 1.19 | 2.00 | 1 | 5 | 0.70 |
| Q17 Seizing | Adaptability | 2.72 | 1.13 | 2.00 | 1 | 5 | 0.70 |
| Q18 Reconfiguring (Structures) | Adaptability | 2.47 | 1.13 | 2.00 | 1 | 5 | 0.69 |
| Q19 Reconfiguring (Learning) | Adaptability | 2.64 | 1.12 | 3.00 | 1 | 5 | 0.74 |
Analysis
Four capabilities, one pattern
The Dynamic Capabilities dimension holds four factors after Teece: sensing (Q16, 2.63), seizing (Q17, 2.72), reconfiguring structures (Q18, 2.47) and reconfiguring as learning (Q19, 2.64). Three of the four items have a median of 2. The distributions are right-skewed: most organisations sit low, a few high values pull the means up.[Probst 2026, Appendix B]
The thesis sees in this dimension the real difference between more and less AI-ready organisations. Structural factors such as resources, talent and leadership backing exist to some degree in early-stage organisations too. What they lack is the capacity to sense, decide and reconfigure in a rapidly changing environment. Adapting structures is not one weakness among many but the most acute gap in the sample.[Probst 2026, Section 5.1]
- Advancedn = 12 · Md 4.00 · IQR 3.75–4.31
- Intermediaten = 27 · Md 2.50 · IQR 2.25–3.13
- Early stagen = 25 · Md 2.00 · IQR 1.50–2.00
Takeaway: The interquartile ranges of the three profiles do not overlap: early stage 1.50–2.00, intermediate 2.25–3.13, advanced 3.75–4.31. The separation holds for the distribution, not just for the means.
- Unit
- Individual organisation scores on the maturity scale 1–5 (mean of four items)
- Population
- DACH analysis: 64 cleaned self-assessments from Germany and Switzerland, mostly senior decision-makers
- Denominator
- n = 25, 27 and 12 individual scores
- Source
- Probst 2026, Figure 8 · Figure 8 – Distribution of Dynamic Capabilities scores by profile (Section 4.4.2)
Data as a table
| Profile | n | M | Median | Q1 | Q3 | Min | Max |
|---|---|---|---|---|---|---|---|
| Advanced | 12 | 4.13 | 4.00 | 3.75 | 4.31 | 3.50 | 5.00 |
| Intermediate | 27 | 2.66 | 2.50 | 2.25 | 3.13 | 1.75 | 3.75 |
| Early stage | 25 | 1.84 | 2.00 | 1.50 | 2.00 | 1.00 | 3.00 |
The distribution shows how stable the separation is. The advanced profile, with a median of 4.00 (interquartile range 3.75 to 4.31), sits entirely above the intermediate profile (median 2.50, 2.25 to 3.13), which in turn sits above the early stage (median 2.00, 1.50 to 2.00). Each profile moves within its own band.[Probst 2026, Figure 8]
Size does not replace agility
In the literature, firm size counts as an enabling factor because resources are available. In the DACH analysis, large organisations are nevertheless overrepresented in the early stage. The thesis does not conclude that size is a hindrance, since large organisations appear in the advanced profile too. It suggests that resource advantages may come at the expense of other preconditions, such as flexibility and adaptability, and points to Pumplun et al. (2019), who describe bureaucratic structures as a counterweight to resource advantages.[Pumplun et al. 2019]
The European survey confirms the pattern for adapting processes: 2.59 of 5, and 59% of the 163 answers sit at level 1 or 2. After rollout, little changes in the majority of organisations.[AI Monitor 2026, Europe]
Teklens interpretation
For your product team
The study describes organisations across sectors. Applying it to software product teams is our interpretation, not an empirical finding about product teams.
Our recommendation: the four capabilities as four habits of a product team.
Sensing: collect signals in one place
Customer feedback, support cases, model updates and regulation land in different tools today. One shared place where they connect to decisions turns sensing into routine instead of exception.
Seizing: scale with evidence
Q17 asks whether budgets grow where pilots deliver measurable evidence. For a team that means: one Activity, one measurable result, then the decision about more.
Reconfiguring: rebuild after every run
Level 3 on Q18 reads "defined playbooks: clear roles enable plannable adaptations". After every initiative, record which step is dropped, which is simplified and who now approves. That is the core of our AI Product Playbook.
Learning: keep knowledge in the team
Q19 measures how quickly new model classes become productive. A team learns faster when experience from one Activity stays available as context for the next instead of vanishing in chats.
Scope & limits
- The study measures adaptability as a self-assessment, not through observed process changes.
- That adaptability separates the profiles most sharply describes the cluster solution; it is no proof that it causes readiness.
- The size effect rests on cross-tabulations with small cells; the thesis deliberately runs no significance tests.
- Applying this to playbooks and Activities is our interpretation. The study does not measure whether Teklens or any other tool increases adaptability.
Questions & answers
What are dynamic capabilities?
After Teece, an organisation's capacity to sense opportunities and threats, to commit resources to them (seizing) and to reconfigure structures, processes and knowledge. The AI Monitor measures them with four items.
Why does reconfiguring matter more than sensing?
The study does not weight, but Q18 is the lowest score and the thesis calls adapting structures the most acute gap. Sensing without reconfiguring stays observation.
What does level 5 on adapting structures look like?
"Feedback loops and agile product management guarantee permanent reconfiguration of processes, roles and products before structures become outdated." Few organisations in the sample reach that level.
Sources
- Probst 2026, Table 4 · Capturing AI's Potential: How Ready Are Organisations in the DACH Region? (2026). Table 4 and Figure 4 – Means and standard deviations of the six dimensions (Section 4.3). Sample: n = 64 cleaned responses, Germany and Switzerland · Scale: Maturity scale 1–5 · Limits: Self-assessment, exploratory, not representative.Probst, F. (2026). Capturing AI's Potential: How Ready Are Organisations in the DACH Region? An Empirical Analysis Using an Integrated TOE–DCT Framework. Master's thesis, University of St.Gallen, Institute of Information Systems and Digital Business. Supervisor: Prof. Dr. Ingrid Bauer-Hänsel; co-supervisor: Prof. Dr. Benjamin van Giffen. St. Gallen, 21 August 2026. Conducted in cooperation with Teklens.
- Probst 2026, Table 5 · Capturing AI's Potential: How Ready Are Organisations in the DACH Region? (2026). Table 5 and Figure 7 – Dimension scores by profile (Section 4.4.2). Sample: n = 64; profiles of n = 12, 27 and 25 · Scale: Maturity scale 1–5 · Limits: Profiles from K-means on the same six dimensions; differences between profiles are large by construction.Probst, F. (2026). Capturing AI's Potential: How Ready Are Organisations in the DACH Region? An Empirical Analysis Using an Integrated TOE–DCT Framework. Master's thesis, University of St.Gallen, Institute of Information Systems and Digital Business. Supervisor: Prof. Dr. Ingrid Bauer-Hänsel; co-supervisor: Prof. Dr. Benjamin van Giffen. St. Gallen, 21 August 2026. Conducted in cooperation with Teklens.
- Probst 2026, Figure 8 · Capturing AI's Potential: How Ready Are Organisations in the DACH Region? (2026). Figure 8 – Distribution of Dynamic Capabilities scores by profile (Section 4.4.2). Sample: n = 64 individual scores; profiles of n = 12, 27 and 25 · Scale: Maturity scale 1–5 · Limits: Self-assessment, exploratory, not representative.Probst, F. (2026). Capturing AI's Potential: How Ready Are Organisations in the DACH Region? An Empirical Analysis Using an Integrated TOE–DCT Framework. Master's thesis, University of St.Gallen, Institute of Information Systems and Digital Business. Supervisor: Prof. Dr. Ingrid Bauer-Hänsel; co-supervisor: Prof. Dr. Benjamin van Giffen. St. Gallen, 21 August 2026. Conducted in cooperation with Teklens.
- Probst 2026, Section 5.1 · Capturing AI's Potential: How Ready Are Organisations in the DACH Region? (2026). Section 5.1 – Interpretation of key findings. Sample: n = 64 cleaned responses, Germany and Switzerland · Scale: Maturity scale 1–5 · Limits: Interpretation by the thesis; no causal claim.Probst, F. (2026). Capturing AI's Potential: How Ready Are Organisations in the DACH Region? An Empirical Analysis Using an Integrated TOE–DCT Framework. Master's thesis, University of St.Gallen, Institute of Information Systems and Digital Business. Supervisor: Prof. Dr. Ingrid Bauer-Hänsel; co-supervisor: Prof. Dr. Benjamin van Giffen. St. Gallen, 21 August 2026. Conducted in cooperation with Teklens.
- Probst 2026, Appendix B · Capturing AI's Potential: How Ready Are Organisations in the DACH Region? (2026). Appendix B – Mean, standard deviation, median, minimum, maximum and item-rest correlation of all 19 items. Sample: n = 64 cleaned responses, Germany and Switzerland · Scale: Maturity scale 1–5 · Limits: Self-assessment, exploratory, not representative.Probst, F. (2026). Capturing AI's Potential: How Ready Are Organisations in the DACH Region? An Empirical Analysis Using an Integrated TOE–DCT Framework. Master's thesis, University of St.Gallen, Institute of Information Systems and Digital Business. Supervisor: Prof. Dr. Ingrid Bauer-Hänsel; co-supervisor: Prof. Dr. Benjamin van Giffen. St. Gallen, 21 August 2026. Conducted in cooperation with Teklens.
- Probst 2026, Appendix C · Capturing AI's Potential: How Ready Are Organisations in the DACH Region? (2026). Appendix C – One-way ANOVA and Tukey HSD per dimension. Sample: n = 64 cleaned responses, Germany and Switzerland · Scale: F values, df (2, 61), p · Limits: Describes how sharply the clusters separate on the clustering variables themselves; not a hypothesis test.Probst, F. (2026). Capturing AI's Potential: How Ready Are Organisations in the DACH Region? An Empirical Analysis Using an Integrated TOE–DCT Framework. Master's thesis, University of St.Gallen, Institute of Information Systems and Digital Business. Supervisor: Prof. Dr. Ingrid Bauer-Hänsel; co-supervisor: Prof. Dr. Benjamin van Giffen. St. Gallen, 21 August 2026. Conducted in cooperation with Teklens.
- Probst 2026, Appendix E · Capturing AI's Potential: How Ready Are Organisations in the DACH Region? (2026). Appendix E – Profile composition by industry, firm size and hierarchical position. Sample: n = 64 cleaned responses, Germany and Switzerland · Scale: Frequencies · Limits: Many cells hold fewer than five cases; the thesis deliberately runs no significance tests and reads the patterns as exploratory tendencies.Probst, F. (2026). Capturing AI's Potential: How Ready Are Organisations in the DACH Region? An Empirical Analysis Using an Integrated TOE–DCT Framework. Master's thesis, University of St.Gallen, Institute of Information Systems and Digital Business. Supervisor: Prof. Dr. Ingrid Bauer-Hänsel; co-supervisor: Prof. Dr. Benjamin van Giffen. St. Gallen, 21 August 2026. Conducted in cooperation with Teklens.
- Teece 2007 · Explicating dynamic capabilities (2007). Strategic Management Journal, 28(13). Sample: Theory · Scale: – · Limits: Sensing, seizing and reconfiguring as the core concepts.Teece, D. J. (2007). Explicating dynamic capabilities: The nature and microfoundations of (sustainable) enterprise performance. Strategic Management Journal, 28(13), 1319–1350. https://doi.org/10.1002/smj.640
- Pumplun et al. 2019 · A new organizational chassis for artificial intelligence (2019). ECIS 2019. Sample: Qualitative study · Scale: – · Limits: Used by the thesis to interpret the firm-size pattern.Pumplun, L., Tauchert, C., & Heidt, M. (2019). A new organizational chassis for artificial intelligence – Exploring organizational readiness factors. Proceedings of the 27th European Conference on Information Systems (ECIS), 106. https://aisel.aisnet.org/ecis2019_rp/106
- AI Monitor 2026, Europe · AI Monitor 2026 – European survey, raw export (2026). Recomputed from the raw export of 2026-08-19; documented in the internal editorial brief of 2026-08-18 (addenda). Sample: n = 165 valid responses, 106 with demographic detail; single items n = 163 · Scale: 1–5 per item, each response normalised to 0–100 · Limits: Self-assessment; no country field in the export; industry and size not reconciled. Role comparisons compare different people, not the same organisation.Teklens (2026). AI Monitor 2026 – European survey on organisational AI readiness, raw export of 19 August 2026. Research partners: ETH Zürich and University of St.Gallen. Internal aggregate, published on teklens.ai/ai-monitor.