Q: What explains Accenture's persistent above-programme NPS score across 3 cohorts, and what would change it?

Accenture's above-programme NPS (+73 in C7, +75 in C11, +67 in C9 vs programme avg +42) is consistent but the sample sizes are small enough that persona and proficiency mix — not firm culture — is the most plausible structural explanation.

Accenture appears in three cohorts with org-level NPS data: C7 (NPS +73, n=11), C11 (NPS +75, n=12), and C9 (NPS +67, n=6). In all three, Accenture outperforms the cohort NPS by roughly 30-40 points. However, with n=6 to n=12 per cohort, these scores are statistically fragile — a single promoter or detractor shifting would materially change the result. The verbatim evidence offers a partial clue: the one Accenture promoter quote visible in C7 ('Solid for baselining people across the skill') and C11 ('Interactive and fun') suggests Accenture participants may arrive with calibrated expectations — neither overqualified nor underprepared — which is the single biggest driver of NPS variance across all orgs in this dataset. The recurring detractor pattern programme-wide is advanced participants finding content too basic; if Accenture sends a deliberately mixed or mid-proficiency cohort, they would systematically avoid that failure mode. What would change it: sending a cohort of frontier-level practitioners (as seen in other orgs) would likely drive Accenture's score toward programme average or below, consistent with the broader pattern where 'Operating at the frontier' participants generate most detractor feedback. Key data points: · Accenture NPS: C7 +73 (n=11), C11 +75 (n=12), C9 +67 (n=6) — all above cohort NPS · Programme-wide detractor pattern: frontier-level practitioners consistently cite content as too basic (e.g. 'Few new things... instructors not capable of answering deeper technical questions', 'It's a bit too basic for us') · Cohort-level NPS swings of 30-50 pts are common across orgs when a single firm dominates a cohort (e.g. PwC: -45 in C7 vs +86 in C15), suggesting firm-level scores are highly sensitive to who is sent Caveats: All three Accenture org-level scores are based on n=6 to n=12 — below the 20-respondent threshold for confident segment conclusions. The data does not include Accenture-specific proficiency breakdowns or persona splits, which would be needed to confirm the expectation-calibration hypothesis. Collecting AI proficiency distribution by org would sharpen this answer significantly.

2026-07-28 18:06 UTC

Q: What explains Persistent Systems's persistent above-programme NPS score across 2 cohorts, and what would change it?

Persistent Systems scores +83 (C15) and +80 (C20) against programme NPS of +42, but both readings rest on n=6 and n=5 respectively — too small to identify structural causes with confidence.

Persistent Systems appeared in two cohorts: C15 San Francisco (NPS +83, n=6) and C20 San Francisco (NPS +35, n=5). Both beat their respective cohort NPS (+55 and +35) and the programme average (+42), averaging +81.5 across the two observations. The one Persistent Systems verbatim from C15 is a promoter praising 'the whole cadence of the workshop and the planned topics' (NPS 10, Developer, Applying in practice), while C21 shows a Persistent Systems Architect scoring 9 with 'explanatory and with a good level of hands on.' The single C15 detractor from Persistent Systems wanted 'more advanced learning' on tokens and application integration — suggesting some senior practitioners in the group had unmet depth expectations. The available data cannot distinguish whether the above-average scores reflect a consistently well-prepared cohort profile, effective internal pre-selection of attendees, or cohort-level delivery variation, because no persona or AI proficiency breakdown is available for the Persistent Systems segment specifically. Key data points: · C15: Persistent Systems NPS +83, n=6 (cohort NPS +55, programme NPS +42) · C20: Persistent Systems NPS +80, n=5 (cohort NPS +35, programme NPS +42) · Only 1 Persistent Systems detractor identified across both cohorts — an Architect wanting more advanced token/application depth (C15, NPS 6) Caveats: Combined n=11 across both cohorts is well below the 20-respondent threshold for confident segment-level conclusions — these scores could shift substantially with one or two additional responses. To explain the pattern, you would need persona and AI proficiency breakdowns for the Persistent Systems segment specifically, plus interview data on how they select and prepare attendees for Basecamp.

2026-07-28 18:06 UTC

Q: What explains KPMG's persistent below-programme NPS score across 2 cohorts, and what would change it?

KPMG attendees consistently score ~28 pts below programme NPS (+42), driven by a mismatch between content level and their existing AI proficiency — verbatims point directly to expectation gaps and content calibration failures.

KPMG appears in two cohorts: C12 (NPS +17, n=18) and C21 (NPS +0, n=4). Compared to the programme average of +42, this is a gap of -25 pts in C12 and -42 pts in C21. KPMG verbatims in C12 are the most diagnostic: detractors explicitly cite the content being too low-level for experienced users ('copy and pasting code'), environment setup consuming learning time, and a failure to set audience expectations ('expectation setting should have been done'). A C12 passive goes further, noting that a more experienced attendee could summarise the learning for colleagues without attending, and calls for 'differentiated tracks.' The C21 KPMG score (NPS +0) is based on only n=4 respondents, so must be treated with caution, but directionally consistent. C12's confidence delta is also the only negative in the dataset (-0.11), meaning KPMG-heavy C12 participants left less confident than they arrived — a signal the programme exposed complexity without resolving it for this audience. No org-level proficiency breakdown is available, but KPMG's detractor feedback consistently signals an over-experienced cohort relative to content pitch. Key data points: · C12 KPMG NPS: +17 vs programme avg +42 (gap of -25 pts), n=18 · C21 KPMG NPS: +0 vs programme avg +42 (gap of -42 pts), n=4 — small sample, treat directionally · C12 confidence delta: -0.11 (only negative in dataset) — participants left less confident than they arrived · C12 KPMG detractor verbatims cite: content too basic, copy-paste exercises, environment setup friction, no expectation setting · C12 KPMG passive verbatim explicitly requests 'differentiated tracks' for experienced attendees Caveats: C21 KPMG n=4 is too small for confident conclusions — treat as directionally consistent, not confirmatory. To sharpen the diagnosis, per-org AI proficiency distribution data would confirm whether KPMG attendees skew toward 'Delivering independently' or 'Operating at the frontier' tiers, which would explain the content-level mismatch quantitatively. What would change it: a pre-cohort proficiency screen for KPMG nominees, routing higher-proficiency attendees to an advanced track or deferring until one exists.

2026-07-28 18:06 UTC

Q: What differentiates C16 London June 15-16, 2026's below-average NPS — is this delivery, composition, or programme fit?

C16's NPS +17 (vs programme mean +42) is primarily a programme fit problem driven by Transformation Leads, not delivery failure — but a high-experience cohort composition amplified the gap.

The sharpest signal is persona-level: Transformation Leads scored NPS -12 (n=8) while Developers scored +50 (n=6) — a 62-point gap within the same cohort. This divergence is not unique to C16 but is the steepest observed. The Transformation Lead detractor verbatims are explicit about fit: one Partner-level Deloitte attendee gave a score of 2, stating 'there was no go to market, no what to do to not have failed POCs' and calling the content 'no differentiator to a half day set of exercises I can do remote.' A second detractor asked for business insights and solution design rather than technical exercises. Composition likely amplified this: C16 had a high share of experienced participants (10 'Delivering independently' out of 36 Day 1 attendees) and only 1 'Operating at the frontier' respondent — suggesting a mid-to-advanced cohort encountering content pitched at an introductory-to-intermediate level. Confidence delta (D2-client: 3.96) is the second-lowest across all cohorts, suggesting participants left with limited felt readiness to apply learning in client contexts. The detractor rate of 21.7% matches C7 San Francisco (21.7%), which also had composition and fit issues, but C16's promoter rate (39.1%) is the lowest of any London cohort. Key data points: · Transformation Lead NPS: -12 (n=8) vs Developer NPS: +50 (n=6) — 62-point persona gap within C16 · Detractor rate 21.7% (tied highest across all cohorts); promoter rate 39.1% (lowest London cohort) · D2-client confidence score 3.96 — second lowest across all 16 cohorts, above only C14 (3.06 not shown) and below programme average · Deloitte NPS: +0 (n=6) — Deloitte attendees, who are likely among the more experienced, returned neutral scores · Only 1 of 36 Day 1 attendees rated 'Operating at the frontier', but 10 rated 'Delivering independently' — mid-to-advanced skew without the depth to match Caveats: Persona-level n is small (Transformation Lead n=8, Developer n=6) — directional only, not statistically robust. No facilitator notes or session-level ratings are available in the data, so delivery quality cannot be fully ruled out as a contributing factor; the low D2-client score is consistent with either a delivery or fit explanation.

2026-07-28 18:07 UTC

Q: 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?

C9's 46% passive rate is driven by experienced participants finding content too basic — the two detractor verbatims both cite insufficient depth for advanced users, and passives gave neutral rather than enthusiastic responses.

C9 had only 2 detractor verbatims, both explicitly naming depth as the issue: one stated 'few new things... instructors did not seem capable of answering deeper technical questions' (NPS 4, Developer, Horváth, frontier-level) and the other said the content was 'more for less experienced recipients' (NPS 5). The passive verbatims do not echo this as strongly — responses like 'great content' (NPS 8), 'it helped a lot to know what we can do with Claude' (NPS 7), and 'gave a sense that outcomes are achievable' (NPS 8) are positive but lack the enthusiasm needed to push to 9-10. C9's AI proficiency mix is notably advanced: 13 participants were 'Delivering independently' and 13 were 'Operating at the frontier' out of 55 day-1 attendees — a higher concentration of advanced practitioners than many cohorts. The Transformation Lead persona scored lowest at NPS +29 vs Architect at +75 and Developer at +40, suggesting non-technical passives may have wanted more applied or strategic framing rather than deeper technical content. Converting passives to promoters likely requires two levers: for advanced technical participants, deeper or frontier-level content tracks; for Transformation Leads, clearer client-application framing. Key data points: · Passive rate: 45.8% (11 of 24 respondents), matching detractor-low rate of 8.3% · Frontier + Delivering independently participants: 26 of 55 day-1 (47%) — skewing experienced · Transformation Lead NPS +29 vs Developer +40 vs Architect +75 — largest persona gap in C9 · Both detractor verbatims explicitly cite content being too basic for experienced users · Passive verbatims are positive but brief — none name a specific unmet need Caveats: C9 final NPS n=24 is small; persona-level splits (Architect n=4, Transformation Lead n=7) are below the 20-respondent threshold for confident conclusions. Only 3 passive verbatims are available for C9, limiting qualitative depth — additional open-text responses or post-event interviews would sharpen the conversion diagnosis.

2026-07-28 18:07 UTC

Q: What is driving the high passive rate in C13 London June 2-3, 2026 — what do the verbatims say and what would convert them to promoters?

C13's 60% passive rate is driven by three passives who felt Anthropic staff were absent, content didn't stretch them, or the programme only enabled non-technical users — not by detractors, of whom there were zero.

C13 had 15 respondents with 0% detractors and 40% promoters, meaning all dissatisfaction sits in the passive band (9 passives). The three captured passive verbatims reveal distinct but related friction points: one passive (NPS 8, b.telligent Developer, Delivering independently) explicitly noted 'Anthropic wasn't there' — suggesting the absence of Anthropic staff reduced perceived credibility or access. A second passive (NPS 7, Deloitte Transformation Lead, Applying in practice) said the programme left them 'with more questions than when I came' — positive framing but unresolved depth. A third passive (NPS 7, Infosys Transformation Lead, Learning and exploring) described the value as 'enable someone non-technical to use Claude', implying the content felt pitched below their expectation or role. Promoter verbatims by contrast praised instructor quality, evals content, and perspective shifts — suggesting the programme works well for mid-level practitioners new to Claude but loses more experienced or senior attendees. The cohort was also heavily skewed toward Engineering (14 of ~19 Day 1 attendees), with only 1 Architect and 1 Business leadership participant, and 1 'Operating at the frontier' respondent — so the passive signal likely reflects experienced Engineers finding insufficient stretch rather than a broad dissatisfaction. Key data points: · 60% passive rate (9/15), 0% detractors — all score drag comes from 7s and 8s, not 6s or below · Passive verbatim 1: 'Anthropic wasn't there' (NPS 8, b.telligent, Delivering independently) — staff presence cited as gap · Passive verbatim 2: 'Enable someone non-technical to use Claude' (NPS 7, Infosys, Learning and exploring) — content perceived as pitched below seniority · Engineering-heavy cohort: 14 of ~19 Day 1 attendees in Engineering function, with only 1 at 'Operating at the frontier' proficiency · D2 confidence scores were healthy (apply 4.27, client 4.33) — passives left more confident, but not wowed enough to promote Caveats: Only 3 passive verbatims are available for 9 passives — 6 passives gave no open text, so the themes above are directional, not representative. n=15 is below the 20-respondent threshold for confident segment-level conclusions; treat findings as hypothesis-generating rather than definitive.

2026-07-28 18:07 UTC

Q: What is driving the high passive rate in C20 San Francisco July 14–15, 2026 — what do the verbatims say and what would convert them to promoters?

C20's 44% passive rate is driven by a content-relevance mismatch: Transformation Leads (NPS +17, n=18) found the programme too technical for their consulting context, while PwC attendees (NPS -11, n=9) were the primary detractor bloc.

C20 produced only 3 verbatim responses across passives and detractors, so direct passive verbatims are thin — but the pattern is clear. Passives gave generic positive signals ('Hands on and engaging', 'Very insightful content', 'Well organized') with no specific criticism, suggesting satisfied-but-not-wowed rather than a resolvable complaint. Detractor verbatims are more diagnostic: one PwC Transformation Lead rated it 6 citing 'not as directly relevant for my work in finance consulting'; a Deloitte Architect (score 6) wanted advanced content ('ontology and use of ontology with Claude to create agents'); and a PwC Transformation Lead (score 6, 'Learning and exploring' proficiency) said it was 'very technical' while acknowledging the networking value. The Transformation Lead persona — 18 of 42 matched respondents — scored NPS +17, far below Developers (+70, n=10) and Architects (+43, n=14), pointing to a persona-level fit problem not a delivery problem. PwC's NPS of -11 (n=9) versus Deloitte's +44 (n=9) in the same cohort reinforces that org-level use-case fit varied sharply. Confidence scores were the lowest in the dataset for apply-AI (4.02), suggesting the programme moved passives forward technically but not enough to feel ready to act. Key data points: · Transformation Leads: NPS +17 (n=18) vs Developers +70 (n=10) — 53-point gap within same cohort · PwC NPS -11 (n=9) vs Deloitte NPS +44 (n=9) — same cohort, opposite outcomes · D2 apply-AI confidence 4.02 — lowest of any cohort in the dataset, suggesting content landed below threshold for readiness · 3 of 3 detractor verbatims cite relevance or depth mismatch, not delivery quality · Passive verbatims are positive but vague — no actionable friction surfaced from passives themselves Caveats: Only 3 C20 detractor verbatims and 3 passive verbatims are available; this is insufficient to draw statistically confident conclusions about what specifically would convert passives. A short follow-up pulse to C20 passives — specifically PwC and Transformation Lead segments — asking one open question about what would have moved their score to 9/10 would sharpen this considerably.

2026-07-28 18:08 UTC