Programme NPS +37 · 4 cohorts · 142 D2 respondents
4 v2 cohorts · 207 D1 respondents · Historical baseline NPS: +35
Programme NPS · all cohorts
+37
P 53% · Pa 32% · D 16%
n=142 D2 respondents
n=142 D2 respondents
Cohorts
4
4 with full v2 data
Respondents
207
207 D1 · 142 D2
Cohort summary
| Cohort | NPS | P% | Pa% | D% | D2 n | D1 n | Match | Conf Δ |
|---|---|---|---|---|---|---|---|---|
| C6 London May 18–19, 2026 | +43 | 57.1% | 28.6% | 14.3% | 21 | 39 | 95% | +0.24 |
| C7 San Francisco May 19-20, 2026 | +32 | 53.3% | 25.0% | 21.7% | 60 | 65 | 83% | +0.09 |
| C8 San Francisco May 21-22, 2026 | +43 | 54.1% | 35.1% | 10.8% | 37 | 48 | 92% | +0.32 |
| C9 London May 20-21, 2026 | +38 | 45.8% | 45.8% | 8.3% | 24 | 55 | 88% | +0.22 |
Conf Δ = D2 apply-AI minus D1 baseline (v2 cohorts only). Match = D1/D2 linked respondents.
London · San Francisco · programme snapshot
London
+40
P 51% · Pa 38% · D 11%
n=45 D2 · 2 cohorts
n=45 D2 · 2 cohorts
C6 London May 18–19, 2026, C9 London May 20-21, 2026
San Francisco
+36
P 54% · Pa 29% · D 18%
n=97 D2 · 2 cohorts
n=97 D2 · 2 cohorts
C7 San Francisco May 19-20, 2026, C8 San Francisco May 21-22, 2026
Same content and programme design. City split reflects internal team and audience differences.
Audience composition · 207 D1 respondents pooled
Pooled across all cohorts. NPS by org aggregated from v2 cohorts only (n ≥ 4 suppressed).
Function
Seniority
AI Proficiency
Organisation distribution · pooled
| Organisation | Distribution | n | % |
|---|---|---|---|
| Accenture | 43 | 20.8% | |
| Deloitte | 41 | 19.8% | |
| PwC | 23 | 11.1% | |
| Infosys | 13 | 6.3% | |
| Cognizant | 9 | 4.3% | |
| Reply | 8 | 3.9% | |
| Capgemini | 7 | 3.4% | |
| valantic | 6 | 2.9% | |
| Infomotion | 5 | 2.4% | |
| SFEIR | 4 | 1.9% | |
| Ascendion | 4 | 1.9% | |
| Fractal Analytics | 4 | 1.9% | |
| Sia | 4 | 1.9% | |
| Version 1 | 4 | 1.9% | |
| Theodo | 4 | 1.9% | |
| Netlight | 4 | 1.9% | |
| NTT Data | 3 | 1.4% | |
| Lovelytics | 3 | 1.4% | |
| Praecipio | 3 | 1.4% | |
| Horváth | 3 | 1.4% |
NPS by organisation · aggregate (v2 cohorts, n ≥ 4)
| Organisation | NPS | n |
|---|---|---|
| Capgemini | +75 | n=4 |
| Accenture | +72 | n=18 |
| Deloitte | +55 | n=33 |
| Reply | +50 | n=6 |
| Infosys | +36 | n=11 |
| Version 1 | +25 | n=4 |
| Cognizant | -14 | n=7 |
| PwC | -31 | n=16 |
Day 1 calibration · confidence, depth, pace and relevance
Pooled across v2 cohorts. Confidence on 1–5 scale. Depth/pace as % of persona respondents.
D1 Confidence by Persona · mean /5
Bar = % of max scale (5). Pooled across v2 cohorts.
D1 Confidence by AI Proficiency · mean /5
Bar = % of max scale (5). Pooled across v2 cohorts.
Content Depth by Persona · pooled
| Persona | Too basic | About right | Too advanced |
|---|---|---|---|
| Architect | 23% n=12 | 75% n=39 | 2% n=1 |
| Developer | 21% n=15 | 76% n=55 | 3% n=2 |
| Transformation Lead | 8% n=7 | 76% n=63 | 16% n=13 |
Session Pace by Persona · pooled
| Persona | Too slow | Well paced | Too fast |
|---|---|---|---|
| Architect | 10% n=5 | 85% n=44 | 6% n=3 |
| Developer | 3% n=2 | 92% n=66 | 6% n=4 |
| Transformation Lead | 5% n=4 | 81% n=67 | 14% n=12 |
Content Relevance by Persona · mean /5 · programme mean 4.08/5
Relevance = Day 1 session relevance rating (1–5). Pooled across v2 cohorts.
Outcomes · NPS by persona, proficiency and open text
Aggregate NPS breakouts from v2 cohorts (n ≥ 4 suppressed). 116 NPS open-text responses available.Synthesis not generated
NPS by Persona · aggregate (v2 cohorts, n ≥ 4)
| Persona | NPS bar | NPS | n |
|---|---|---|---|
| Architect | +66 | n=26 | |
| Developer | +31 | n=45 | |
| Transformation Lead | +28 | n=54 |
NPS by AI Proficiency · pooled (v2 cohorts, n ≥ 4)
| Proficiency | NPS | P% | Pa% | D% | n |
|---|---|---|---|---|---|
| Applying in practice | +20 | 41.7% | 36.7% | 21.7% | n=60 |
| Delivering independently | +59 | 66.7% | 25.9% | 7.4% | n=27 |
| Operating at the frontier | +55 | 65.0% | 25.0% | 10.0% | n=20 |
| Learning and exploring | +35 | 52.9% | 29.4% | 17.6% | n=17 |
Open text · NPS reasons · 116 responses
Synthesis not generatedRun with
--synthesis to generate thematic analysis via Claude Haiku. 116 responses available.Confidence · D1→D2 deltas across all three dimensions
Apply-AI · Design-AI · Commercial confidence. D1 baseline → D2 delta. v2 cohorts only.
D1 → D2 Confidence by Dimension · per cohort
| Cohort | D1 build | D2 apply-AI | D2 design-AI | D2 commercial |
|---|---|---|---|---|
| C6 London May 18–19, 2026 | 4.00 | 4.24 +0.24 | 4.05 +0.05 | 4.19 +0.19 |
| C7 San Francisco May 19-20, 2026 | 4.08 | 4.17 +0.09 | 4.28 +0.20 | 4.28 +0.20 |
| C8 San Francisco May 21-22, 2026 | 4.00 | 4.32 +0.32 | 4.14 +0.14 | 4.19 +0.19 |
| C9 London May 20-21, 2026 | 3.95 | 4.17 +0.22 | 4.38 +0.43 | 4.17 +0.22 |
D2 value shown above; delta (D2 – D1) shown below in smaller text. Scale 1–5.
Confidence Delta by Persona · pooled
D1→D2 apply-AI mean delta. Positive = gained confidence. Scale 1–5.
Confidence Delta by NPS Segment · pooled
Promoters gain more confidence than detractors — or vice versa?
5 systematic patterns · 1 flag · 5 suggested queries
Systematic patterns = same direction vs programme NPS across ≥2 cohorts. Suggested queries pre-written for ask.py.
Function / Persona patterns · consistent across ≥2 cohorts
| Pattern | Direction | Avg Δ | Cohorts | Seen in |
|---|---|---|---|---|
| Architect scored above programme NPS in all 4 cohorts analysed (avg +28 pts) | ↑ Above programme | +28 pts | 4 | C6 London May 18–19, 2026, C7 San Francisco May 19-20, 2026, C8 San Francisco May 21-22, 2026, C9 London May 20-21, 2026 |
AI Proficiency patterns · consistent across ≥2 cohorts
| Pattern | Direction | Avg Δ | Cohorts | Seen in |
|---|---|---|---|---|
| Delivering independently scored above programme NPS in all 4 cohorts analysed (avg +17 pts) | ↑ Above programme | +17 pts | 4 | C6 London May 18–19, 2026, C7 San Francisco May 19-20, 2026, C8 San Francisco May 21-22, 2026, C9 London May 20-21, 2026 |
Organisation patterns · consistent across ≥2 cohorts
| Pattern | Direction | Avg Δ | Cohorts | Seen in |
|---|---|---|---|---|
| PwC scored below programme NPS in all 2 cohorts analysed (avg -60 pts) Likely reflects audience fit, not delivery variance. | ↓ Below programme | -60 pts | 2 | C7 San Francisco May 19-20, 2026, C8 San Francisco May 21-22, 2026 |
| Accenture scored above programme NPS in all 2 cohorts analysed (avg +35 pts) Likely reflects audience fit, not delivery variance. | ↑ Above programme | +35 pts | 2 | C7 San Francisco May 19-20, 2026, C9 London May 20-21, 2026 |
| Deloitte scored above programme NPS in all 2 cohorts analysed (avg +17 pts) Likely reflects audience fit, not delivery variance. | ↑ Above programme | +17 pts | 2 | C7 San Francisco May 19-20, 2026, C8 San Francisco May 21-22, 2026 |
Outlier flags
● High passive rateHigh passive rate in C9 London May 20-21, 2026 (46%). Passives are the most likely source of churn — explore what would convert them to promoters.
Suggested ask.py queries
Run these commands from the Partner Basecamp/ folder to investigate the patterns above. Add --save when you have a conclusion to commit to the insights cache.
Architect — persona pattern
Architect scored above programme NPS in all 4 cohorts analysed (avg +28 pts)
cd "$REPO" && python3 _scripts/ask.py --question "Why does Architect consistently score +28 pts above programme NPS across 4 cohorts — is this audience fit, content relevance, or delivery?" --cache "/Users/edward.whittall/Library/CloudStorage/Box-Box/Anthropic/Phase II/Partner Enablement/Partner Basecamp v2/Partner Basecamp — Survey Pipeline/findings/insights/c6-c9/insights-data-c6-c9.json" --save
Delivering independently — proficiency pattern
Delivering independently scored above programme NPS in all 4 cohorts analysed (avg +17 pts)
cd "$REPO" && python3 _scripts/ask.py --question "Why do Delivering independently AI proficiency participants rate the programme +17 pts above average — what does the D1 calibration data tell us?" --cache "/Users/edward.whittall/Library/CloudStorage/Box-Box/Anthropic/Phase II/Partner Enablement/Partner Basecamp v2/Partner Basecamp — Survey Pipeline/findings/insights/c6-c9/insights-data-c6-c9.json" --save
PwC — org pattern
PwC scored below programme NPS in all 2 cohorts analysed (avg -60 pts)
cd "$REPO" && python3 _scripts/ask.py --question "What explains PwC's persistent below-programme NPS score across 2 cohorts, and what would change it?" --cache "/Users/edward.whittall/Library/CloudStorage/Box-Box/Anthropic/Phase II/Partner Enablement/Partner Basecamp v2/Partner Basecamp — Survey Pipeline/findings/insights/c6-c9/insights-data-c6-c9.json" --save
High passives: C9 London May 20-21, 2026
Passive rate: 46%
cd "$REPO" && python3 _scripts/ask.py --question "What is driving the high passive rate in C9 London May 20-21, 2026 — what do the verbatims say and what would convert them to promoters?" --cache "/Users/edward.whittall/Library/CloudStorage/Box-Box/Anthropic/Phase II/Partner Enablement/Partner Basecamp v2/Partner Basecamp — Survey Pipeline/findings/insights/c6-c9/insights-data-c6-c9.json" --save
Promoter vs passive — what separates them
NPS verbatim analysis — highest signal for passive conversion
python3 _scripts/ask.py --question "What do promoters and passives say about the programme, and what specifically separates them?" --cache "/Users/edward.whittall/Library/CloudStorage/Box-Box/Anthropic/Phase II/Partner Enablement/Partner Basecamp v2/Partner Basecamp — Survey Pipeline/findings/insights/c6-c9/insights-data-c6-c9.json"
Investigation log · 1 conclusion saved · 5 open queries
Conclusions committed via ask.py --save, open signals from Tab 06 patterns, and pre-written queries.1 conclusion saved
Saved conclusions · 1 from ask.py --save
Q: What do promoters and passives say about the programme, and what specifically separates them?
Promoters cite practical value and strong structure; passives feel the content was good but not quite pitched right — the separating factor is whether the programme fully met their experience level.
Promoters consistently highlight tangible learning outcomes: 'very useful technical dive into Claude,' 'fundamental AI engineering,' 'well structured and covered important topics,' and 'well matched against product capabilities.' Their language is decisive and outcome-focused. Passives, by contrast, use qualified positive language — 'great content,' 'very intensive,' 'intro to a lot of cool new stuff' — but several signal unmet expectations around depth or pacing, e.g. 'day one could have been more foundational' and 'would like more build guidance instead of working off pre-built plumbing.' The clearest separator is level-fit: passives sit between the two failure modes — not advanced enough to feel under-served (as detractors do) but not quite in the sweet spot where the content fully delivers. C9 is the most acute example, with a passive rate of 45.8% (n=24, matching the promoter rate exactly), suggesting a large middle cohort that found the programme useful but not compelling enough to advocate.
Key data points:
· Promoter rate across C6-C9: 52.5% (n=74 of 141); passive rate: 31.9% (n=45 of 141)
· C9 London has the highest passive rate at 45.8%, equal to its promoter rate of 45.8% (n=24)
· Passive verbatims flag level-fit and format issues: 'more build guidance,' 'day one could have been more foundational,' 'very intensive with a lot to learn' — none cite poor content quality outright
· Promoter verbatims are outcome-specific: 'baselining people across the skill,' 'great practical session,' 'well matched against product capabilities'
· No passive explicitly says they would not recommend; their hesitation is about degree of value, not absence of it
Caveats: Verbatim coverage is thin — only 3-4 per segment per cohort are surfaced, so thematic patterns are indicative rather than statistically robust. A full verbatim export with NPS score attached would allow sentiment scoring and more confident theme separation across all 141 respondents.
2026-06-04 15:59 UTC
Open signals
Systematic patterns
5
consistent across ≥2 cohorts
Outlier flags
1
anomalies flagged
Conclusions saved
1
from ask.py --save
Suggested queries
5
pre-written in Tab 06
Patterns not yet addressed by a saved conclusion:
- Architect scored above programme NPS in all 4 cohorts analysed (avg +28 pts)
- Delivering independently scored above programme NPS in all 4 cohorts analysed (avg +17 pts)
- PwC scored below programme NPS in all 2 cohorts analysed (avg -60 pts)
- Accenture scored above programme NPS in all 2 cohorts analysed (avg +35 pts)
- Deloitte scored above programme NPS in all 2 cohorts analysed (avg +17 pts)
Pre-written queries → Tab 06 · Patterns
Tab 06 contains pre-written ask.py commands for each pattern and flag detected. Run them from the Partner Basecamp/ folder, review the output, then re-run with --save to commit the conclusion here.
- Architect — persona pattern — Architect scored above programme NPS in all 4 cohorts analysed (avg +28 pts)
- Delivering independently — proficiency pattern — Delivering independently scored above programme NPS in all 4 cohorts analysed (avg +17 pts)
- PwC — org pattern — PwC scored below programme NPS in all 2 cohorts analysed (avg -60 pts)
- High passives: C9 London May 20-21, 2026 — Passive rate: 46%
- Promoter vs passive — what separates them — NPS verbatim analysis — highest signal for passive conversion
Data Quality · survey compliance and completeness
Issues flagged for transparency and database migration. Assessment and remediation tracked with delivery team.
Programme Summary · Survey Compliance
| Total Day 2 responses | 142 |
| Matched to Day 1 | 125 (88.0% of Day 2) |
| Day 2-only respondents | 17 (12.0%) — no Day 1 data, excluded from persona & confidence analysis |
| NPS impact of D2-only respondents | suppresses programme NPS by 5 points |