← Home
Partner Basecamp · Cohort 22

London

July 20–21, 2026
+41
Net Promoter Score
55
Day 1 Responses
75%
Matched Pairs
58%
Day 2 Response Rate
Programme Satisfaction
NPS +41 with 53% passives — score is held up by low detractor count, not promoter strength. Passive signal is weak: neither endorsing nor rejecting the programme.
+41NPS · n=32
44% Promoters53% Passives3% Detractors
95% CI: +22 → +60  ·  True NPS lies within this range with 95% confidence (n=32 respondents)
Flag — 53% passive rate is above the programme mean. What in the programme experience left people undecided?
Confidence Arc
Moderate end-of-Day-1 build confidence — cohort is progressing; Day 2 has room to close remaining gaps (mean 3.8/5).
3.8end-of-D1 build / 5
4.1D2 Design / 5
4.3D2 Commercial / 5
4.2D2 Build / 5
Audience
55 participants across 14 organisations — Architect majority with applying in practice the most common AI experience level.
Top Organisations
Capgemini (13) DXC Technology (7) PwC (7) Accenture (4) Bain (4)
Voice of Participant
What participants said — dominant theme from each NPS segment.
Promoters
Interactive delivery and hands-on exercises enabled participants to absorb substantial AI knowledge and practical optimization methods within two days.
Passives
Technical content depth and exercise time were mismatched; some participants wanted more conceptual presentations, while others found basics insufficient post-certification.
Detractors
Content lacked advanced topics and Anthropic engineering expertise, leaving the curriculum at an introductory rather than specialized level.
Survey & Data Quality
Survey & Data Quality
Day 1 responses55
Day 2 responses32 (58% of Day 1)
Matched pairs24 (75% of Day 2)
Orgs resolved55 of 55 respondents matched to named org
Day 2-only respondents8 (25.0%) — suppresses NPS by 3 points; no Day 1 data, excluded from persona & confidence analysis
⚠ Low Day 2 rateOnly 58% of Day 1 respondents completed Day 2. Push the survey link during or immediately after the closing session.
Organisations
All Day 1 respondents · n=55
Capgemini13 (24%)
DXC Technology7 (13%)
PwC7 (13%)
Accenture4 (7%)
Bain4 (7%)
valantic4 (7%)
Deloitte3 (5%)
Slalom3 (5%)
Ascendion2 (4%)
IndiciumAI2 (4%)
The Agile Monkeys2 (4%)
Reply2 (4%)
Infosys1 (2%)
Forgd.AI1 (2%)
Function × Seniority
All Day 1 respondents · n=55
FunctionPractitionerSr PractitionerMgr / Sr MgrDirector+Partner / MDTotal
Engineering7103––20
Architecture12102–15
Business Leadership––4419
Project / Engmt–272–11
Experience Profile
AI experience level · n=55 · % of row
By Persona
SegmentExploringApplyingDeliveringFrontiern
Developer5%75%5%15%20
Transformation Lead25%45%10%20%20
Architect33%40%13%13%15
By Seniority
SegmentExploringApplyingDeliveringFrontiern
Manager or Senior Manager25%38%12%25%24
Senior practitioner (5–9 years)7%71%7%14%14
Director, Senior Director, or Principal25%62%12%—8
Practitioner (0-4 years in role)25%75%——8
Partner, Managing Director, or Executive———100%1
How would you describe your current experience with AI tools prior to Basecamp?
Learning and exploring
11 (20%)
Applying in practice
30 (55%)
Delivering independently
5 (9%)
Operating at the frontier
9 (16%)
Prior to today, how much had you worked with Claude or the Anthropic API?
Not at all
5 (9%)
A little
27 (49%)
Regularly
23 (42%)
Did the depth land for this audience?
Technical depth perception · n=55 respondents · % of row
By Persona
SegmentToo basicAbout rightToo advancedn
Developer10%85%5%20
Transformation Lead—95%5%20
Architect13%80%7%15
By AI Experience Level
SegmentToo basicAbout rightToo advancedn
Applying in practice3%97%—30
Learning and exploring—73%27%11
Operating at the frontier33%67%—9
Delivering independently—100%—5
Overall depth distribution
Too basic
4 (7%)
About right
48 (87%)
Too advanced
3 (5%)
Did the pace work across the room?
Session pace perception · n=55 respondents · % of row
By Persona
SegmentToo slowWell pacedToo fastn
Developer10%50%40%20
Transformation Lead—70%30%20
Architect7%67%27%15
By AI Experience Level
SegmentToo slowWell pacedToo fastn
Applying in practice—67%33%30
Learning and exploring—55%45%11
Operating at the frontier22%56%22%9
Delivering independently20%60%20%5
Overall pace distribution
Too slow — could have covered more
3 (5%)
Well paced
34 (62%)
Too fast — not enough time to apply
18 (33%)
How confident were participants to build with Claude after Day 1?
End-of-Day-1 confidence · "How confident are you in your ability to build a client solution using Claude?" · n=55 · 1–5
Mean end-of-D1 build confidence by Persona
Transformation Lead3.85/5 · n=20
Developer3.70/5 · n=20
Architect3.67/5 · n=15
Programme mean: 3.9/5 · 15 cohorts
Mean end-of-D1 build confidence by AI Experience Level
Operating at the frontier4.44/5 · n=9
Delivering independently4.20/5 · n=5
Applying in practice3.70/5 · n=30
Learning and exploring3.09/5 · n=11
Overall end-of-D1 build confidence distribution
1
0 (0%)
2
4 (7%)
3
17 (31%)
4
23 (42%)
5
11 (20%)
How relevant was today's content to your current role?
Content relevance rating · n=55 respondents · 1 = not relevant, 5 = highly relevant
Mean relevance by Persona
Transformation Lead4.35/5 · n=20
Architect4.20/5 · n=15
Developer3.95/5 · n=20
Mean relevance by AI Experience Level
Delivering independently4.40/5 · n=5
Applying in practice4.17/5 · n=30
Operating at the frontier4.11/5 · n=9
Learning and exploring4.09/5 · n=11
Overall relevance distribution
1
0 (0%)
2
0 (0%)
3
11 (20%)
4
24 (44%)
5
20 (36%)
How likely are you to recommend attending this programme to a colleague?
n=32 Day 2 respondents · 58% of Day 1
Score distribution · 0 = not at all likely · 10 = extremely likely
Cohort012345678910
Cohort 22110768
■ Detractors 0–6■ Passives 7–8■ Promoters 9–10
NPS Breakdown
44%
53%
3%
NPS +41 (n=32)
Promoters (9–10)Passives (7–8)Detractors (0–6)
Programme mean: +42 · 15 cohorts
Promoters mean: 52%
Passives mean: 38%
Detractors mean: 10%
NPS by Segment
Organisations, personas, and experience levels · ≥4 respondents shown
By Organisation
Promoters (9–10)Passives (7–8)Detractors (0–6)
Capgemini
57%
43%
0%
NPS +57 (n=7)
By Persona
Promoters (9–10)Passives (7–8)Detractors (0–6)
Transformation Lead
50%
50%
0%
NPS +50 (n=8)
Architect
50%
38%
12%
NPS +38 (n=8)
Developer
25%
75%
0%
NPS +25 (n=8)
Programme means · 15 cohorts: Architect +52 · Developer +48 · Transformation Lead +35
By AI Experience Level
Promoters (9–10)Passives (7–8)Detractors (0–6)
Applying in practice
46%
54%
0%
NPS +45 (n=11)
Learning and exploring
33%
67%
0%
NPS +33 (n=6)
Operating at the frontier
25%
50%
25%
NPS 0 (n=4)
Programme means · 15 cohorts: Learning and exploring +45 · Applying in practice +44 · Operating at the frontier +42 · Delivering independently +42
What is the main reason for your score?
n=24 responses · organised by NPS segment
Promoters (score 9–10)· 12 responses
Interactive delivery and hands-on exercises enabled participants to absorb substantial AI knowledge and practical optimization methods within two days.
9“I think the basecamp provides a great opportunity to learn and share knowledge about AI and Claude and continue building a partners comunity”
Passives (score 7–8)· 11 responses
Technical content depth and exercise time were mismatched; some participants wanted more conceptual presentations, while others found basics insufficient post-certification.
7“I think it’s a good opportunity for being in a positive environment of people willing to learn and adapt to this new paradigm. However, for people that already have some knowledge, like if you already have passed the Claude Certified Architect – Foundations (CCA-F) certification, it can be a bit basic.”
Detractors (score 0–6)· 1 response
Content lacked advanced topics and Anthropic engineering expertise, leaving the curriculum at an introductory rather than specialized level.
6“The covered topics were too basic. There should have been some engineer from Anthropic to cover advanced topics.”
End-of-Programme Confidence
n=32 respondents · 1 = not confident, 5 = very confident
How confident are you in your ability to design an evaluation for an AI solution?
12345Mean
——19%53%28%4.09
Programme mean: 4.1/5 · 15 cohorts
Leaving today, how comfortable are you advising clients on AI when questions arise?
12345Mean
——9%47%44%4.34
Programme mean: 4.2/5 · 15 cohorts
How confident are you in your ability to lead a conversation about Anthropic and Claude with a client?
12345Mean
——16%47%38%4.22
Programme mean: 4.2/5 · 15 cohorts
Confidence Trajectory by Persona
D1→D2 delta · apply-AI confidence · matched pairs only · n=24
Positive delta = confidence grew · negative = dropped
Architect (n=8)
+0.50
Developer (n=8)
+0.38
Transformation Lead (n=8)
+0.25
What one takeaway will you share with a colleague or client?
n=18 responses
Participants discovered Anthropic's technical sophistication exceeds expectations, shifting focus from model selection to implementation strategy.
“Anthropic is further ahead that I initially thought.”
Evaluation frameworks enable production confidence by systematically diagnosing issues as architectural rather than model-related problems.
“That we have ways to evaluate Ai systems and to ensure we can ship to prod with confidence”
Agent architecture and evaluation frameworks matter more than model choice, enabling non-coders to drive continuous enterprise-scale improvement.
“An eval framework is probably the most important substrate to scale across the enterprise and enable the continuous improvement of as it guardrails the uncertain probablistic outputs and catches the dumb stuff - it's the bowling alley floats in the gutters.”
Practical prototyping capability accelerates, but production deployment demands deeper expertise than the training programme provides.
“Faster prototypes but more expertise necessary in prod”
What will you build for yourself or apply at work in the next 30 days?
n=19 responses
Participants want hands-on experience building swarm-style agents and adjusting parameters to solve real productivity challenges.
“Practice building more agents and experimenting with levers.”
Consultants are reconsidering their model selection strategies and exploring Claude's capabilities for client delivery.
“Rethink the model choice in my solutions.”
Participants need more workshop time to build, experiment, and iterate on agent solutions rather than theory.
“Practice building more agents and experimenting with levers.”
Consultants seek practical tools to assess customer environments and integrate AI into their existing architecture practices.
“Scan a customer environment, collect certificate metadata and score them for quantum readiness.”
Participants recognize evaluation frameworks as critical but need deeper guidance on implementing them effectively.
“evaluations”
Partner Basecamp Survey Report · Cohort 22 · Generated by analyze.py v2.3
NPS Reason
n=24 responses · grouped by NPS segment
Promotersscore 9–10 · 12 responses
1Amazing training full of insights and hands on work10
2I learned a lot in 2 days. Good investment of time.10
3Some great insights into how to optimise use of AI and the methods around how to build it into a system.10
4I especially enjoyed the section on evals, as well as all the valuable content that was shared10
5Great content10
6It’s a fantastic 2 day workshop , no slides , hands on for any level , leadership to intern calude primer / workshop , so why not ?10
7Huge amount of useful info presented by enthusiastic team10
8Very good9
9Great presenters. Very interactive9
10Learned a lot. Very interesting.9
11Useful and valuable content in a compact format9
12I think the basecamp provides a great opportunity to learn and share knowledge about AI and Claude and continue building a partners comunity9
Passivesscore 7–8 · 11 responses
1Overall a great experience.8
2The way it supports design and architecture decisions.8
3Good insights how we can improve optimization when clients are not happy with their current solution8
4Poor lab instructions8
5The format benefits still from a relatively similar capability audience - generally liked the interactivity of it.8
6Small hints with big impacts8
7content7
8Very technical content but little time to really run the exercises7
9Hands-on exercises are good, but it would helpful to have more presentations (like day 2) to present concepts and solutions rather than experiencing with it.7
10it needed more depth7
11I think it’s a good opportunity for being in a positive environment of people willing to learn and adapt to this new paradigm. However, for people that already have some knowledge, like if you already have passed the Claude Certified Architect – Foundations (CCA-F) certification, it can be a bit basic.7
Detractorsscore 0–6 · 1 response
1The covered topics were too basic. There should have been some engineer from Anthropic to cover advanced topics.6
Most Valuable
What one takeaway will you share with a colleague or client? · n=18 responses
1inference optimisation
2Speed of development
3Anthropic is further ahead that I initially thought.
4evaluations
5It is possible to evaluate model performance. Also, the cause of issues is almost never the model
6Don’t agonise over model choice
7You don't need to have an extended technical knowledge , Claude cna teach and guyde you
8It's not about the model but how you architect the agents.
9Use AI via APIs, not just chat.
10Claudes analysis and awareness capacity
11That we have ways to evaluate Ai systems and to ensure we can ship to prod with confidence
12LLMs can hallucinate, but there are ways to keep those hallucinations under control
13Evals are great!
141. Agentic value lies in the fast flowing repetitive work that’s slightly unpredictable but in a defined domain. 2. An eval framework is probably the most important substrate to scale across the enterprise and enable the continuous improvement of as it guardrails the uncertain probablistic outputs and catches the dumb stuff - it’s the bowling alley floats in the gutters.
15Faster prototypes but more expertise necessary in prod
16Exchange of Ideas and direction
17Use Claude , improve your overall quality of your day to day work life . Use Claude inference optimisation in your architectures and make a slide to share the savings with your clients .
18The capability of the models and how we dont need coders to make progress in agent development. We just need people who understand the principles of agent development and customer requirements
30-Day Intentions
What will you build for yourself or apply at work? · n=19 responses
1Create a side hustle
2More agentic, swarm style solutions.
3Rework the code
4Rethink the model choice in my solutions.
5evaluations
6Optimize my token consumption
7Accelerators for personal tasks
8Locally running agents to help me with productivity and time Management
9RFP processing framework, new project scaffolding framework, ai harness for greenfield / brownfield projects
10Practice building more agents and experimenting with levers.
11As a fun project probably to convert a GPX cycle route into recommendations of stops. Probably improve a project I have relating to OCR. At work, taking the skills I have developed for analysing infrastructure and Databricks, and converting that into Agents or MCP.
12Wip
13I”ll be automatizing workflows at work for sure!
14I’ll play around with the Agent SDK and explore how I can use it to optimize my workflows
15Let’s see if we can apply something from the inference optimization
16Batch projects
17Revise workshop content
18Evolve my day to day work , architecture and program management skills , tools and llm as judge ( with myself in the loop ) … plus an AI adoption platform for my clients …
19Scan a customer environment, collect certificate metadata and score them for quantum readiness.