← Home
Partner Basecamp · Cohort 37

San Francisco

October 27–28, 2026
+37
Net Promoter Score
50
Day 1 Responses
86%
Matched Pairs
98%
Day 2 Response Rate
Programme Satisfaction
NPS +37 with 51% passives — the positive score reflects few detractors more than strong promoter enthusiasm. Passives read as a neutral signal: engaged enough to stay, not yet won over enough to champion it.
+37NPS · n=49
43% Promoters51% Passives6% Detractors
95% CI: +20 → +53  ·  True NPS lies within this range with 95% confidence (n=49 respondents)
Flag — 51% 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
4.0D2 Commercial / 5
4.0D2 Build / 5
Audience
50 participants across 14 organisations — Transformation Lead majority with applying in practice the most common AI experience level.
Top Organisations
IBM (13) Trace3 (9) PwC (7) Sia Partners (5) Deloitte (4)
Voice of Participant
What participants said — dominant theme from each NPS segment.
Promoters
It was very informative and hands on. Loved it!
Passives
Liked the pre work material and the exercises
Detractors
There were a lot of WiFi and API issues internally and unless those are addressed by PwC I don’t think it’s valuable for people to come
Segment Finding
PwC (n=7, NPS -14) detractors cited access/network restrictions on their own side (WiFi, API, system blockers) during Day 1, not content issues — one respondent explicitly says PwC needs to resolve this internally before sending more people. Architects (n=12, NPS +67) rated the hands-on/practical content highly despite the same connectivity friction surfacing once outside PwC too.
Survey & Data Quality
Survey & Data Quality
Day 1 responses50
Day 2 responses49 (98% of Day 1)
Matched pairs42 (86% of Day 2)
Orgs resolved50 of 50 respondents matched to named org
Day 2-only respondents7 (14.3%) — suppresses NPS by 6 points; no Day 1 data, excluded from persona & confidence analysis
Organisations
All Day 1 respondents · n=50
IBM13 (26%)
Trace39 (18%)
PwC7 (14%)
Sia Partners5 (10%)
Deloitte4 (8%)
Protiviti2 (4%)
Making Sense2 (4%)
Salesforce2 (4%)
Bain1 (2%)
Modelit1 (2%)
AWS1 (2%)
Wipro1 (2%)
Cognizant1 (2%)
Robert half1 (2%)
Function × Seniority
All Day 1 respondents · n=50
FunctionPractitionerSr PractitionerMgr / Sr MgrDirector+Partner / MDTotal
Engineering941––14
Architecture2922217
Business Leadership–––314
Project / Engmt6162–15
Experience Profile
AI experience level · n=50 · % of row
By Persona
SegmentExploringApplyingDeliveringFrontiern
Transformation Lead26%63%11%—19
Architect29%41%18%12%17
Developer7%79%7%7%14
By Seniority
SegmentExploringApplyingDeliveringFrontiern
Practitioner (0-4 years in role)18%65%12%6%17
Senior practitioner (5–9 years)36%57%7%—14
Manager or Senior Manager11%78%11%—9
Director, Senior Director, or Principal29%29%29%14%7
Partner, Managing Director, or Executive—67%—33%3
How would you describe your current experience with AI tools prior to Basecamp?
Learning and exploring
11 (22%)
Applying in practice
30 (60%)
Delivering independently
6 (12%)
Operating at the frontier
3 (6%)
Prior to today, how much had you worked with Claude or the Anthropic API?
Not at all
1 (2%)
A little
25 (50%)
Regularly
24 (48%)
Did the depth land for this audience?
Technical depth perception · n=50 respondents · % of row
By Persona
SegmentToo basicAbout rightToo advancedn
Transformation Lead5%84%11%19
Architect6%94%—17
Developer29%71%—14
By AI Experience Level
SegmentToo basicAbout rightToo advancedn
Applying in practice10%87%3%30
Learning and exploring—91%9%11
Delivering independently17%83%—6
Operating at the frontier67%33%—3
Overall depth distribution
Too basic
6 (12%)
About right
42 (84%)
Too advanced
2 (4%)
Did the pace work across the room?
Session pace perception · n=50 respondents · % of row
By Persona
SegmentToo slowWell pacedToo fastn
Transformation Lead16%68%16%19
Architect—100%—17
Developer—79%21%14
By AI Experience Level
SegmentToo slowWell pacedToo fastn
Applying in practice7%80%13%30
Learning and exploring—82%18%11
Delivering independently17%83%—6
Operating at the frontier—100%—3
Overall pace distribution
Too slow — could have covered more
3 (6%)
Well paced
41 (82%)
Too fast — not enough time to apply
6 (12%)
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=50 · 1–5
Mean end-of-D1 build confidence by Persona
Architect3.94/5 · n=17
Developer3.93/5 · n=14
Transformation Lead3.79/5 · n=19
Programme mean: 3.9/5 · 26 cohorts
Mean end-of-D1 build confidence by AI Experience Level
Operating at the frontier5.00/5 · n=3
Delivering independently4.33/5 · n=6
Applying in practice3.93/5 · n=30
Learning and exploring3.18/5 · n=11
Overall end-of-D1 build confidence distribution
1
1 (2%)
2
2 (4%)
3
13 (26%)
4
20 (40%)
5
14 (28%)
How relevant was today's content to your current role?
Content relevance rating · n=50 respondents · 1 = not relevant, 5 = highly relevant
Mean relevance by Persona
Architect4.29/5 · n=17
Developer4.07/5 · n=14
Transformation Lead3.84/5 · n=19
Mean relevance by AI Experience Level
Operating at the frontier4.33/5 · n=3
Learning and exploring4.27/5 · n=11
Delivering independently4.00/5 · n=6
Applying in practice3.97/5 · n=30
Overall relevance distribution
1
0 (0%)
2
2 (4%)
3
9 (18%)
4
23 (46%)
5
16 (32%)
How likely are you to recommend attending this programme to a colleague?
n=49 Day 2 respondents · 98% of Day 1
Score distribution · 0 = not at all likely · 10 = extremely likely
Cohort012345678910
Cohort 3712817615
■ Detractors 0–6■ Passives 7–8■ Promoters 9–10
NPS Breakdown
43%
51%
6%
NPS +37 (n=49)
Promoters (9–10)Passives (7–8)Detractors (0–6)
Programme mean: +43 · 26 cohorts
Promoters mean: 53%
Passives mean: 38%
Detractors mean: 9%
NPS by Segment
Organisations, personas, and experience levels · ≥4 respondents shown
By Organisation
Promoters (9–10)Passives (7–8)Detractors (0–6)
IBM
54%
46%
0%
NPS +54 (n=13)
Deloitte
50%
50%
0%
NPS +50 (n=4)
Trace3
44%
56%
0%
NPS +44 (n=9)
PwC
14%
57%
29%
NPS -14 (n=7)
Sia Partners
25%
50%
25%
NPS 0 (n=4)
By Persona
Promoters (9–10)Passives (7–8)Detractors (0–6)
Architect
67%
33%
0%
NPS +67 (n=12)
Developer
29%
64%
7%
NPS +21 (n=14)
Transformation Lead
25%
62%
12%
NPS +12 (n=16)
Programme means · 26 cohorts: Architect +51 · Developer +50 · Transformation Lead +36
By AI Experience Level
Promoters (9–10)Passives (7–8)Detractors (0–6)
Applying in practice
37%
56%
7%
NPS +30 (n=27)
Learning and exploring
38%
50%
12%
NPS +25 (n=8)
Delivering independently
20%
80%
0%
NPS +20 (n=5)
Programme means · 26 cohorts: Delivering independently +46 · Applying in practice +46 · Learning and exploring +45 · Operating at the frontier +34
What is the main reason for your score?
n=43 responses · organised by NPS segment
Promoters (score 9–10)· 19 responses
It was very informative and hands on. Loved it!
9“Was very informative as someone new to Claude code. I think there is room for improvement with the exercises. I think it’d be smart for someone to lead walking through the exercise at the front so those who want to follow along live can - but others can also self pace. I think also having more staff would be helpful to give more hands on help. Lots of people were struggling at times (e.g. with api key) and it felt like it took a while to be resolved.”
Passives (score 7–8)· 21 responses
Liked the pre work material and the exercises
8“Great experience to get more hands-on experience with Claude and best practices. The reiteration of aspects such as evals and inference optimization were great. One aspect that could use improvement is maybe how this ties into dealing with clients. I really enjoyed the event still.”
Detractors (score 0–6)· 3 responses
There were a lot of WiFi and API issues internally and unless those are addressed by PwC I don’t think it’s valuable for people to come
6“This will be too technical for my colleagues. I found it useful but maybe there needs to be levels to this, with a beginner, advanced etc”
End-of-Programme Confidence
n=49 respondents · 1 = not confident, 5 = very confident
How confident are you in your ability to design an evaluation for an AI solution?
12345Mean
—2%10%65%22%4.08
Programme mean: 4.1/5 · 26 cohorts
Leaving today, how comfortable are you advising clients on AI when questions arise?
12345Mean
—8%10%57%24%3.98
Programme mean: 4.2/5 · 26 cohorts
How confident are you in your ability to lead a conversation about Anthropic and Claude with a client?
12345Mean
—4%16%57%22%3.98
Programme mean: 4.2/5 · 26 cohorts
Confidence Trajectory by Persona
D1→D2 delta · apply-AI confidence · matched pairs only · n=42
Positive delta = confidence grew · negative = dropped
Architect (n=12)
+0.50
Developer (n=14)
-0.07
Transformation Lead (n=16)
-0.06
What one takeaway will you share with a colleague or client?
n=34 responses
Claude / Anthropic content
“Its always never the model”
Evals & testing
“The evaluation harness for agentic systems”
AI agents & engineering
“The evaluation harness for agentic systems”
Hands-on practice
“How hands on it was and very kind instructors”
Networking & peer learning
“i will be sharing the inference information and some context engineering commands because i think those are under utilized”
What will you build for yourself or apply at work in the next 30 days?
n=37 responses
AI agents & engineering
“agentic solutions”
Evals & testing
“Evals and different prompt rescue techniques”
Claude / Anthropic content
“Evals and different prompt rescue techniques”
Tool setup / readiness
“Build an productivity tool for consultants”
Hands-on practice
“See the APIs we have available and build an agent that connects to be able to run daily agents for my team”
NPS Reason
n=43 responses · grouped by NPS segment
Promotersscore 9–10 · 19 responses
1It was very informative and hands on. Loved it!10
2Very well run and structured, clear guidance and well presented.10
3It was so helpful to get hands on and instructive code examples with incredible instructors10
4Very informative and practical in nature10
5learned a lot of valuable information.10
6Great quality10
7Engaging and in person - cool Location and food - cool people10
8Well planned and thought out content and schedule. It gave good understanding and overview of agentic development, diagnosis and optimization that will be helpful.10
9I enjoyed the workshops and meeting others to learn about Claude and it’s true capabilities.10
10The way sessions were organized10
11Learned a great deal10
12It gives a good start to experience and onboard you using Claude.10
13Great hands on labs10
14Learned alot. Very immersive and hands on9
15Great content and aligned with what we are doing9
16Great content. Need more depth.9
17Was very informative as someone new to Claude code. I think there is room for improvement with the exercises. I think it’d be smart for someone to lead walking through the exercise at the front so those who want to follow along live can - but others can also self pace. I think also having more staff would be helpful to give more hands on help. Lots of people were struggling at times (e.g. with api key) and it felt like it took a while to be resolved.9
18Heavy hands on, good depth9
19Learned a ton…really built confidence9
Passivesscore 7–8 · 21 responses
1Liked the pre work material and the exercises8
2practical exercises8
3I appreciated the pace and how hands on it was. Seeing some of the slides before hand would’ve helped with retention for me personally but the repos are going to be a good resource to refer back to8
4Depends on their expertise level. If mid-high, maybe no need to come. I’d beginner. Yes.8
5Was good content on how to use Claude but not sure if it prepared me to go in and train partners.8
6Interesting content8
7Hands on activities8
8It was a great training. I enjoyed some of the content and liked the notebook style of learning8
9The issues on access that were a blocker for the first half of the day due to our system restrictions.8
10Hands on experience, thoughtful activities, very helpful and enthusiastic instructors8
11I enjoyed learning about how to fix solutions already in place. The evaluation exercise was also super helpful!8
12I’d like a little more theory applied with demos and showcases, I feel that I could do this from home8
13Great experience to get more hands-on experience with Claude and best practices. The reiteration of aspects such as evals and inference optimization were great. One aspect that could use improvement is maybe how this ties into dealing with clients. I really enjoyed the event still.8
14Content and brand8
15API struggles on day 17
16I think for a lot of new users would get good information. The beginning was a little basic, but today was really good and helpful in my opinion7
17Learnt a lot about AI based code generation and agent building7
18It was well organized and dynamic, good sessions to learn and practice7
19I found the hands on aspects.7
20It was nice but some of these concepts might be out of date in a couple months. Although I already I’ve heard about a lot of these concepts, this is my first time actually using them and taking a deep dive in them which was helpful7
21It's a fantastic set of demos and work, with great interaction with the Anthropic team. Yet the program can be solved by Claude itself and is focused almost entirely on building applications with use cases that feel a little toy like in the AI age.7
Detractorsscore 0–6 · 3 responses
1It’s good and engaging, would be higher but it does not go as deep technically as I would like, include more case studies6
2This will be too technical for my colleagues. I found it useful but maybe there needs to be levels to this, with a beginner, advanced etc6
3There were a lot of WiFi and API issues internally and unless those are addressed by PwC I don’t think it’s valuable for people to come1
Most Valuable
What one takeaway will you share with a colleague or client? · n=34 responses
1inference optimization techniques
2How hands on it was and very kind instructors
3I can give them a hands on experience
4Its always never the model
5The evaluation harness for agentic systems
6You can make Claude cost less
7Importance of eval
8Token usage across different models
9The eval lesson was so helpful to address client concerns about the ability to properly test an agent
10Evals and optimizations
11Claude models are very powerful now. If we are facing a issue it is probably the system around the model call and not the model
12Different techniques for Evals
13That we were very good!!!!!! I think Anthropic staff liked us :)))))
14Exciting times
15The depth went into around tokenomics
16Evals and context engineering I would like to dive deeper into
17Eval framework and agent swarm
18Concept of optimization
19i will be sharing the inference information and some context engineering commands because i think those are under utilized
20Course usefulness
21Multi agent development, diagnosing issues with AI, inference optimization
22The problem is most likely the user, not the model.
23The ease of building in Claude and all the necessary steps to keep the guardrails on
24Some is the ways to reduce token usage and the cost savings
25Building an agent with Claude isn’t too hard
26I think the ways we can guide the AI journey are super important. The discussions we had on day 1 are helpful and the ability to quickly spin up a POC is really the expectation now.
27Focus on Evals and next comes context engineering
28Claidw capabilities
29Good session thank you
30How important is to work on harness and not only changing the model
31quality evals are very important
32Implementing practices to use Claude efficiently and not just go gung ho.
33Note: wifi needs improved. Too slow.
34AI is in fact self-improving when directed at a goal, but the evals are where it's at.
30-Day Intentions
What will you build for yourself or apply at work? · n=37 responses
1Build an productivity tool for consultants
2agentic solutions
3Working on dynamic routings for an application at work
4Will need to find a use case to apply the learnings.
5Evals and different prompt rescue techniques
6Improvements in an internal home agentic system I have
7Learning more about features of Claude and the SDK
8Eval
9A regulatory reading agents to help understand ever changing protocols in the healthcare space
10More productivity tools!
11Claims solution for life insurance
12A personal assistant agent to keep track of news in my field
13Building POCs
14Pretty much everything
15Agent runtime governance engine
16A process agent that can truly address exceptions.
17Too much
18Risk Assessment narrative agent
19Llm as judge and other eval frameworks I expect to use and implement in my upcoming contracts
20See the APIs we have available and build an agent that connects to be able to run daily agents for my team
21the inference levers today could be really helpful and will probably get incorporated
22Develop data analytics automation using AI
23Use evals to improve performance
24Multi agent development.
25Implement Claude in all of our workflows, allows for faster deliverables. Also implementing more evals snd checks in prompts, to make sure we are maximizing our tokens.
26A multi agent SOW reviewer
27I will be building in some evaluation components to my current tools
28Would like to see how we can leverage Claude to support a third-line function’s activities, helping them to simplify processes and improve efficiencies.
29AI agent to synthesize transcripts
30A solution for my practice the team has been looking to build.
31A personalized agent for consumer - shopping assistants
32I will be working on personal agent tool for a research project idea I have been thinking about
33I am planning to build an agent which should help in job and recruitment process. This could be my internal project which should add confidence by practice these concepts again.
34Stronger evaluations and prompt optimizations
35going to refine evals and inference optimization
36I already started building a swarm of agents for a side project. Within work, I look forward to auditing my Claude, and implementing these ideas into active engagements whether significant or not.
37Evals at work and exponentially evolving the human with hard, in-the-trenches up and down the complexity and value chain. Obviously this is transformative