← Home
Partner Basecamp · Cohort 31

San Francisco

September 9–10, 2026
+36
Net Promoter Score
77
Day 1 Responses
89%
Matched Pairs
83%
Day 2 Response Rate
Programme Satisfaction
Positive NPS of +36 — promoters outweigh detractors with moderate passive presence.
+36NPS · n=64
53% Promoters30% Passives17% Detractors
95% CI: +17 → +54  ·  True NPS lies within this range with 95% confidence (n=64 respondents)
Confidence Arc
Moderate end-of-Day-1 build confidence — cohort is progressing; Day 2 has room to close remaining gaps (mean 3.9/5).
3.9end-of-D1 build / 5
4.0D2 Design / 5
4.1D2 Commercial / 5
4.1D2 Build / 5
Audience
77 participants across 17 organisations — Transformation Lead majority with applying in practice the most common AI experience level.
Top Organisations
NEC (17) PwC (14) Deloitte (11) Ascendion (8) AWS (5)
Voice of Participant
What participants said — dominant theme from each NPS segment.
Promoters
“Phenomenal hands-on course for adoption and enablement. Well planned out and engaging.”
“Connections, energy, content are all great.”
“Very hands on and applicable.”
Passives
“I found the sessions interesting, informative and challenging. I'm not a developer or programmer so it was still valuable to seeing Claude in action this way.”
“It really hit home that I hadn't been fully utilizing Claude's capabilities until now.”
Detractors
“Helpful for learning Claude code but more technical that my current work.”
“There is high value to the two days but I think splitting the topics and agenda more between functional and technical.”
Segment Finding
PwC (n=8, NPS -38) detractor verbatims cluster around facilitation format, not content: they wanted more instructor-led walkthroughs and demos versus independent workbook time, and a clearer functional/technical agenda split. Ascendion (n=6, NPS +67) left with concrete applied takeaways (cost optimization, custom agents, prompt engineering).
Survey & Data Quality
Survey & Data Quality
Day 1 responses77
Day 2 responses64 (83% of Day 1)
Matched pairs57 (89% of Day 2)
Orgs resolved77 of 77 respondents matched to named org
Day 2-only respondents7 (10.9%) — suppresses NPS by 4 points; no Day 1 data, excluded from persona & confidence analysis
Organisations
All Day 1 respondents · n=77
NEC17 (22%)
PwC14 (18%)
Deloitte11 (14%)
Ascendion8 (10%)
AWS5 (6%)
DXC Technology5 (6%)
IBM4 (5%)
Accenture2 (3%)
Fractal Analytics2 (3%)
McKinsey2 (3%)
Hellman & Friedman1 (1%)
LTM1 (1%)
phData1 (1%)
Vedara1 (1%)
Andela1 (1%)
ABeam Consulting1 (1%)
Fujitsu Intelligence1 (1%)
Function × Seniority
All Day 1 respondents · n=77
FunctionPractitionerSr PractitionerMgr / Sr MgrDirector+Partner / MDTotal
Engineering10101122
Architecture372416
Business Leadership2135112
Project / Engmt978327
Experience Profile
AI experience level · n=77 · % of row
By Persona
SegmentExploringApplyingDeliveringFrontiern
Transformation Lead23%64%8%5%39
Developer14%55%23%9%22
Architect25%38%12%25%16
By Seniority
SegmentExploringApplyingDeliveringFrontiern
Senior practitioner (5–9 years)8%64%16%12%25
Practitioner (0-4 years in role)29%58%12%24
Manager or Senior Manager14%50%7%29%14
Director, Senior Director, or Principal38%38%15%8%13
Partner, Managing Director, or Executive100%1
How would you describe your current experience with AI tools prior to Basecamp?
Learning and exploring
16 (21%)
Applying in practice
43 (56%)
Delivering independently
10 (13%)
Operating at the frontier
8 (10%)
Prior to today, how much had you worked with Claude or the Anthropic API?
Not at all
9 (12%)
A little
30 (39%)
Regularly
38 (49%)
Did the depth land for this audience?
Technical depth perception · n=77 respondents · % of row
By Persona
SegmentToo basicAbout rightToo advancedn
Transformation Lead10%67%23%39
Developer14%64%23%22
Architect25%69%6%16
By AI Experience Level
SegmentToo basicAbout rightToo advancedn
Applying in practice9%72%19%43
Learning and exploring6%69%25%16
Delivering independently30%50%20%10
Operating at the frontier38%50%12%8
Overall depth distribution
Too basic
11 (14%)
About right
51 (66%)
Too advanced
15 (19%)
Did the pace work across the room?
Session pace perception · n=77 respondents · % of row
By Persona
SegmentToo slowWell pacedToo fastn
Transformation Lead82%18%39
Developer82%18%22
Architect12%62%25%16
By AI Experience Level
SegmentToo slowWell pacedToo fastn
Applying in practice2%77%21%43
Learning and exploring75%25%16
Delivering independently10%80%10%10
Operating at the frontier88%12%8
Overall pace distribution
Too slow — could have covered more
2 (3%)
Well paced
60 (78%)
Too fast — not enough time to apply
15 (19%)
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=77 · 1–5
Mean end-of-D1 build confidence by Persona
Architect4.31/5 · n=16
Developer4.00/5 · n=22
Transformation Lead3.64/5 · n=39
Programme mean: 3.9/5 · 22 cohorts
Mean end-of-D1 build confidence by AI Experience Level
Delivering independently4.70/5 · n=10
Operating at the frontier4.38/5 · n=8
Applying in practice3.84/5 · n=43
Learning and exploring3.25/5 · n=16
Overall end-of-D1 build confidence distribution
1
0 (0%)
2
7 (9%)
3
20 (26%)
4
25 (32%)
5
25 (32%)
How relevant was today's content to your current role?
Content relevance rating · n=77 respondents · 1 = not relevant, 5 = highly relevant
Mean relevance by Persona
Developer4.14/5 · n=22
Architect4.00/5 · n=16
Transformation Lead3.74/5 · n=39
Mean relevance by AI Experience Level
Delivering independently4.50/5 · n=10
Applying in practice3.93/5 · n=43
Operating at the frontier3.88/5 · n=8
Learning and exploring3.50/5 · n=16
Overall relevance distribution
1
0 (0%)
2
7 (9%)
3
17 (22%)
4
29 (38%)
5
24 (31%)
How likely are you to recommend attending this programme to a colleague?
n=64 Day 2 respondents · 83% of Day 1
Score distribution · 0 = not at all likely · 10 = extremely likely
Cohort012345678910
Cohort 3113161181024
Detractors 0–6 Passives 7–8 Promoters 9–10
NPS Breakdown
53%
30%
17%
NPS +36 (n=64)
Promoters (9–10)Passives (7–8)Detractors (0–6)
Programme mean: +42 · 22 cohorts
Promoters mean: 52%
Passives mean: 39%
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)
Ascendion
67%
33%
0%
NPS +67 (n=6)
NEC
67%
33%
0%
NPS +67 (n=15)
Deloitte
30%
50%
20%
NPS +10 (n=10)
DXC Technology
0%
80%
20%
NPS -20 (n=5)
PwC
25%
12%
62%
NPS -38 (n=8)
By Persona
Promoters (9–10)Passives (7–8)Detractors (0–6)
Architect
62%
31%
8%
NPS +54 (n=13)
Developer
59%
35%
6%
NPS +53 (n=17)
Transformation Lead
37%
33%
30%
NPS +7 (n=27)
Programme means · 22 cohorts: Architect +50 · Developer +48 · Transformation Lead +36
By AI Experience Level
Promoters (9–10)Passives (7–8)Detractors (0–6)
Delivering independently
100%
0%
0%
NPS +100 (n=6)
Applying in practice
48%
30%
21%
NPS +27 (n=33)
Learning and exploring
38%
46%
15%
NPS +23 (n=13)
Operating at the frontier
20%
60%
20%
NPS 0 (n=5)
Programme means · 22 cohorts: Applying in practice +44 · Learning and exploring +44 · Delivering independently +43 · Operating at the frontier +38
What is the main reason for your score?
n=51 responses · organised by NPS segment
Promoters (score 9–10)· 25 responses
Pretty chill and well paced
10“Being able to obtain a place to exchange opinions with strangers globally, and gaining deeper insight into how to use Claude Code." If this is meant to flow as one sentence (e.g., as part of a list of takeaways or acknowledgments), a smoother version might be: "Gaining a venue to exchange ideas with people from around the world I'd never met, and deepening my understanding of how to use Claude Code.”
Passives (score 7–8)· 15 responses
The day two contents were really good, building graders for evals, inference optimisation
7“I am very involved with lots of AI/agentic efforts and have foundational knowledge of Claude in general via both work and personal use. There were some interesting things that were new to me that makes me want to explore more.”
Detractors (score 0–6)· 11 responses
Helpful for learning Claude code but more technical that my current work
6“I really enjoyed the presenters and wish they would’ve spoken more. For someone at my level, it would’ve been very helpful to see them walk through the exercises instead of doing them by myself. My favorite part was the hackathon and I really enjoyed collaborating with my team. I wish more of the sessions were collaborative, but overall felt like I learned a lot.”
End-of-Programme Confidence
n=64 respondents · 1 = not confident, 5 = very confident
How confident are you in your ability to design an evaluation for an AI solution?
12345Mean
6%22%41%31%3.97
Programme mean: 4.1/5 · 22 cohorts
Leaving today, how comfortable are you advising clients on AI when questions arise?
12345Mean
2%22%42%34%4.09
Programme mean: 4.2/5 · 22 cohorts
How confident are you in your ability to lead a conversation about Anthropic and Claude with a client?
12345Mean
5%22%36%38%4.06
Programme mean: 4.2/5 · 22 cohorts
Confidence Trajectory by Persona
D1→D2 delta · apply-AI confidence · matched pairs only · n=57
Positive delta = confidence grew · negative = dropped
Architect (n=13)
+0.23
Developer (n=17)
+0.06
Transformation Lead (n=27)
+0.26
What one takeaway will you share with a colleague or client?
n=40 responses
Claude / Anthropic content
“Overall token economics, moving between the models.”
Evals & testing
“Evals. Llms as a judge.”
AI agents & engineering
“Swam agents”
Hands-on practice
“Hands On!!!!!”
Networking & peer learning
“I will share the prompting tips and tricks with my colleagues.”
What will you build for yourself or apply at work in the next 30 days?
n=44 responses
AI agents & engineering
“Always on agent for a few ideas”
Claude / Anthropic content
“HTML dashboards underpinned by claude”
Evals & testing
“Eval harness”
AI strategy & use cases
“I will definitely apply the compact concept.”
Networking & peer learning
“I'm currently building an AI Adoption Network to encourage non-AI users the ability to start using AI in their day-to-day lives. Focusing on groups that are either less engaged in AI or have never used it.”
NPS Reason
n=51 responses · grouped by NPS segment
Promotersscore 9–10 · 25 responses
1Connections, energy, content are all great10
2Great course learned a lot of new things10
3Awesome exercises and moderation10
4Very helpful sessions and labs10
5To explore Claude and use it in my work10
6Hands On!!!10
7Pretty Interactive.10
8Great content and exercises10
9This was a great way to learn more and get plugged in further to our joint partnership.10
10I liked the interaction for the workshops and hackathon. Also the new framework as well as inference optimization was good10
11Phenomenal hands-on course for adoption and enablement. Well planned out and engaging.10
12Learning and exposure10
13It was fun. A lot of information thrown at you, but in a good way. I felt like I learned a lot and met some cool people :).10
14Being able to obtain a place to exchange opinions with strangers globally, and gaining deeper insight into how to use Claude Code." If this is meant to flow as one sentence (e.g., as part of a list of takeaways or acknowledgments), a smoother version might be: "Gaining a venue to exchange ideas with people from around the world I'd never met, and deepening my understanding of how to use Claude Code.10
15Through the seminar, I could fully understood the features of Claude code and how to use claude code in various situations. In addition, there are many opportunities to communicate internationally.10
16can learn the basics of "current" AI10
17Able to know the logic how to handle llm and prepare the environment to keep the accuracy10
18Pretty chill and well paced9
19Good day today9
20Very hands on and applicable. Access issues can hinder a bit.9
21Perhaps different tracks for different levels of experience.9
22Very informative and immersive. Loved it. Definitely for those more experienced in Claude Code.9
23Practical knowledges. We could apply this even from tomorrow.9
24I noticed how important defining 'eval'.9
25I learned new things about context memory, evals9
Passivesscore 7–8 · 15 responses
1Very interesting session to deep dive o Claude capabilities8
2It was super fun and informative.8
3ケーススタディがとても実際の業務に即した内容であったため。 少なからず日本ではまだまだAIをお客さんにエージェント提案する段階には至ってないため、少し先の未来に直面する課題を取り組めたと感じてます。8
4I loved the hands on approach.8
5It really hit home that I hadn't been fully utilizing Claude's capabilities until now.8
6The day two contents were really good, building graders for evals, inference optimisation7
7Good amount of hands on work. Would have like more time with applied ai folks about what they are seeing and compelling new use cases.7
8Would've been nice to see each exercise through to the end. Not enough time, but getting a set of prompts afterwards with the "answer" would be nice7
9I am very involved with lots of AI/agentic efforts and have foundational knowledge of Claude in general via both work and personal use. There were some interesting things that were new to me that makes me want to explore more.7
10It was very insightful, some activities required more time than allocated for less experienced developers.7
11I found the sessions interesting, informative and challenging. I’m not a developer or programmer so it was still valuable to seeing Claude in action this way.7
12Great experience to see Claude in action while being able to interact with it hands on7
13getting knowledge of how to try ideas easily7
14A lot of it could be done as online training with chat bot support7
15Great experience to see Claude in action while being able to interact with it hands on7
Detractorsscore 0–6 · 11 responses
1Helpful for learning Claude code but more technical that my current work6
2Difficulty with acceas6
3Very technical. I think your general consultant is mainly using AI to assist in analysis and delivery, instead of designing agents.6
4There is high value to the two days but I think splitting the topics and agenda more between functional and technical.6
5I really enjoyed the presenters and wish they would’ve spoken more. For someone at my level, it would’ve been very helpful to see them walk through the exercises instead of doing them by myself. My favorite part was the hackathon and I really enjoyed collaborating with my team. I wish more of the sessions were collaborative, but overall felt like I learned a lot.6
6No “new” information was shared, almost all of this content could be done remotely with a very similar impact. Very good content but i didn’t really feel that i got much benefit from being in person6
7I believe cohorts should be divided by skill and proficiency5
8Felt like a lot of the workbooks were run autonomously without me understanding what the learning was supposed to be4
9It’s not interactive enough. The coding exercises are very monotonous. It should be more interactive and instructor driven.4
10I would have liked more instructions and demoes since some of this was pretty new.4
11This was very dev focused. A lot of things were not explained well and I fell into trying to figure out coding and using VSCode3
Most Valuable
What one takeaway will you share with a colleague or client? · n=40 responses
1Overall token economics, moving between the models.
2Swam agents
3Get started using Claude asap and learn as you go
4Best practises
5Evals. Llms as a judge.
6Build!
7Evaluation frameworks
8How to make agents
9Evals are a non negotiable. You can’t go from poc to production without measuring through evals and optimising inference.
10it’s great
11Hands On!!!!!
12Better prompt engineering.
13MD tables as input rather than json
14Evals are important
15Experiment every day
16I created an HTML of the course files, that I'll share with the team at our AI Adoption Labs.
17I will share the prompting tips and tricks with my colleagues.
18Claude code does magics
19Have an open mind about what Claude can do
20Bring in more skills, mcp severs, and start on an eval harness. The only thing is getting concrete feedback on what to eval 😅
21Claude is only as limited as your imagination.
22How to tell Claude story to client better.
23How to improve prompts, and how to optimize them
24How to connect the knowkedge of claude code features to architect
25Evals, prompt caching details
26“Token efficiency is the next new trend when we make a new AI solution.”
27Impact and importance of model usage
28Its not just about tokens. You can do some really amazing magic with this stuff tho!
29Ensure you have a foundational understanding of the relevant applications you would use in the session.
30Versatility, of agent applications and where they can be embedded
31Look at your AI pilots and decide upfront the right Claude model to both get the right outcome while maximizing on the right economics
32nice food in program
33come prepared with all resources ready ahead of time
34eval
35We need to build more agents
36Higher models doesn’t necessarily mean right output
37calc cost.
38Will definitely recommand to switch Claude models and how to save the cost
39Look at your AI pilots and decide upfront the right Claude model to both get the right outcome while maximizing on the right economics
40when we face AI problems, it is important to make clear what problems is, and consider how to solve it, and make criteria for evaluation
30-Day Intentions
What will you build for yourself or apply at work? · n=44 responses
1Yes
2Automation pipelines
3HTML dashboards underpinned by claude
4Always on agent for a few ideas
5I will build more agents and use what I’ve learned here to become a better developer.
6Travel agent I hope
7Agentic workflow collecting data sources into a single pane of glass dashboard
8Email monitoring agent
9More efficient agents
10Eval harness
11Build regression tests
12Use Claude for proposal documents
13More Agents and swarms that may ease my work
14I will definitely apply the compact concept.
15I will build a demo repository
16An agent
17An awesome learning repo for our team!
18Maybe proceed with hackathon idea
19I'm currently building an AI Adoption Network to encourage non-AI users the ability to start using AI in their day-to-day lives. Focusing on groups that are either less engaged in AI or have never used it.
20I hope to build more agents.
21Many agents for my routine work
22Agent evals for sure!
23Build out my personal website, and I have a bunch of duplicated files across my homelab that I want to consolidate. Claude has been helping me write a tool to hash the files and compare them.
24Automation, agents.
25Few custom agents.
26Thinking about ways to apply this within projects, and improving existing skills
27The education set for my temammates
28..
29Skill development and eval testing
30I want build a teammof supporting agents for my day to day
31I’m looking at talking the Claude Associate certificate.
32Not sure
33Nothing specifically but o know much more to have an intelligent conversation
34study for claude code architect certified
35I would like to build my own agent
36I learned the basics of AI, and I'll re-take demo/handson again after that, I'll create demo for colleague
37We will build agents to handle RFPs
38Evals.
39loop of vulnerability
40Config generator for f5 bigip with its API
41Definitely will apply compact and hash , and how to save the token to reduce the overall cost. Eval before making model changes
42Consider agents capable of efficiently executing incident response.
43Nothing specifically but o know much more to have an intelligent conversation
44build a system for collecting internal news and make dashboard to see what is hot topic in my company