Partner Basecamp · Cohort 26

How likely are you to recommend attending this programme to a colleague?

n=70 Day 2 respondents · 0% of Day 1

Score distribution · 0 = not at all likely · 10 = extremely likely

Cohort	0	1	2	3	4	5	6	7	8	9	10
Cohort 26					1		2	13	18	14	22

■ Detractors 0–6■ Passives 7–8■ Promoters 9–10

NPS Breakdown

51%

44%

NPS +47 (n=70)

Promoters (9–10)Passives (7–8)Detractors (0–6)

Programme mean: +42 · 16 cohorts

Promoters mean: 51%

Passives mean: 39%

Detractors mean: 10%

NPS by Segment

Organisations, personas, and experience levels · ≥4 respondents shown

By Organisation

No segments with ≥4 responses.

By Persona

No segments with ≥4 responses.

Programme means · 16 cohorts: Architect +51 · Developer +47 · Transformation Lead +36

By AI Experience Level

No segments with ≥4 responses.

Programme means · 16 cohorts: Applying in practice +44 · Learning and exploring +44 · Delivering independently +42 · Operating at the frontier +37

What is the main reason for your score?

n=55 responses · organised by NPS segment

Promoters (score 9–10)· 29 responses

Fantastic appropriate depth for content

9“Loved the sessions but some better walk through of how to do each exercise in Jupyter would be good. Plus more time on agent structure before notebook exercise.”

Passives (score 7–8)· 23 responses

Cool content but could be a bit more social/networky. An evening at a bar (doesn’t even have to be funded) would have been cool

8“Great opportunity to learn and connect. Well organised. However, hackathon is rushed and structure is lacking. I even think it can’t work in teams bigger than 2. Additionally, amount of people is too high in my opinion, if groups were smaller, connection and learning would be better. One more thing, connection and networking part is lacking. Content itself is something that can be distributed, and online, but connection can’t. Evening activities, more emphasis on collaboration would be great. Overall, very valuable experience, presenters are great, content relevant, breaks also amazing.”

Detractors (score 0–6)· 3 responses

Wrong selection of attendance profiles

4“This Partner enablement Course does need fair re-org or restructure. As a senior Architect, expected propositions, enterprise grade ALM, tooling and team design structure data. Did take away tooling info, technicalities and comparable tech understanding (have worked on .net ver of semantic kernel aka copilot agent framework quite a bit) - in that respect the course has rich technical exposure which I enjoyed. However as a partner based architect involved in customer engagements and proposals, I will have to extrapolate with my own materials for this purpose. Wish there were playbooks and enablement materials to work with (sorry, asking to be spoondfed a bit as spoiled by MSFT).”

1	2	3	4	5	Mean
3%	1%	20%	53%	23%	3.91

1	2	3	4	5	Mean
1%	—	20%	60%	19%	3.94

1	2	3	4	5	Mean
1%	3%	11%	59%	26%	4.04

NPS Reason

n=55 responses · grouped by NPS segment

Promotersscore 9–10 · 29 responses

1Fantastic appropriate depth for content10

2Very informative and lots of hands on activities10

3Good environment to explore things you don't have time for in normal work10

4Learnt a lot and enjoyed using Claude10

5For me it was the perfect blend of tech and business context to bring into conversations with clients and potential clients.10

6Fun and meeting peers10

7Enjoyable course with useful information that can impact working practices.10

8It is good refresher for anyone working on genetic design10

9Great practical experience and hands on dealing with clients10

10Structured well for base camp training10

11It was fun and useful. I liked the joined q&a sessions10

12Great content coverage for all levels. We skipped through some content quick but that maybe the day 2 effect10

13Very useful intro to Claude and the application of its capabilities.10

14Fantastic atmosphere, very relevant learning10

15Same answer as yesterday:)10

16The content was engaging consumable and relevant10

17The valuable insights from the training9

18Was fun educational and not at all boring9

19Highly engaging, technically driven, and very relevant to the work I am delivering9

20Very interesting even despite not being technical9

21It was well focused and structured.9

22Loved the sessions but some better walk through of how to do each exercise in Jupyter would be good. Plus more time on agent structure before notebook exercise.9

23Love the design9

24Great content and value9

25Excellent program for people who are new to the world or Claude9

26Good knowledge on anthropic9

27Second day is going into much detail9

28Informative9

29All these concepts only make sense to me when I see how it works in VS code.9

Passivesscore 7–8 · 23 responses

1A lot of new stuff but some are things I knew about previously. A lot of handholding in the workbooks which I feel is a bit too easy8

2Day 1 was really good, lots of hands on work to get to grips with claude. Day 2 wasn't as interesting as there was less hands on and the hackathon wasn't as fun as I'd hoped.8

3Some factual errors in some sessions but overall good content coverage for a 2 day event. Python notebooks are probably not a good format for this.8

4I am in a business role and learned a lot.8

5Learned many things8

6Interesting and meaningfull. Love the hands-on approach. At moments was too basic and the audience seems to wide, so for some people I think it might have not been very good8

7Was really great the two days.8

8Sometimes the time is not enough to finish all work8

9Good content about building AI agents. The evaluation component could be improved, this is an entire discipline in its own right and was skimmed over a bit. Q&A was focussed on the middle tables of the room, people sat at the end tables weren’t really engaged as much.8

10Great opportunity to learn and connect. Well organised. However, hackathon is rushed and structure is lacking. I even think it can’t work in teams bigger than 2. Additionally, amount of people is too high in my opinion, if groups were smaller, connection and learning would be better. One more thing, connection and networking part is lacking. Content itself is something that can be distributed, and online, but connection can’t. Evening activities, more emphasis on collaboration would be great. Overall, very valuable experience, presenters are great, content relevant, breaks also amazing.8

11- Looking at AI solutioning from a more business-oriented POV - Networking with interesting people8

12Cool content but could be a bit more social/networky. An evening at a bar (doesn’t even have to be funded) would have been cool7

13Not sure on audience it may be better over 3 days to give more time for hands on experience7

14The scope of the programme is a bit strange because it mixes different public targets. Sometimes, from a engineer perspective, it feels to basic and other times it feels right7

15Very good Input, but I would Not das that it is for complete beginners7

16Nice teacher, nice food, interesting topics, a bit to basic7

17Good general overview, learned some nice tricks, but expected a bit more depth on the “build ai into your app” vs workflow optimisation7

18Good, engaging, hands-on7

19It can cover more topics if the first day first half things are done prior to the session.7

20Overall I liked the format, I particularly liked the real world examples shared by Mark.7

21It was interesting but I would like more depth in some exercises.7

22The course delivers a lot of information in a short time and there is little time to gain a full understanding of some aspects.7

23No real breakthrough I believe a lot of engineers in the attendance achieved more on claude than some trainers7

Detractorsscore 0–6 · 3 responses

1Wrong selection of attendance profiles6

2I think the content might be too basic for my colleagues, since most of them are already at a more advanced stage in their use of AI. That's why I gave that rating, specifically with the idea of recommending it to my teammates.6

3This Partner enablement Course does need fair re-org or restructure. As a senior Architect, expected propositions, enterprise grade ALM, tooling and team design structure data. Did take away tooling info, technicalities and comparable tech understanding (have worked on .net ver of semantic kernel aka copilot agent framework quite a bit) - in that respect the course has rich technical exposure which I enjoyed. However as a partner based architect involved in customer engagements and proposals, I will have to extrapolate with my own materials for this purpose. Wish there were playbooks and enablement materials to work with (sorry, asking to be spoondfed a bit as spoiled by MSFT).4

Most Valuable

What one takeaway will you share with a colleague or client? · n=50 responses

1Model changes aren’t the answer

2Evals

3Evals are key

4How to save money with caching/bulk requests

5Optimisation

6Prompt engineering and context is more important than the model

7Usage of evals

8Photos from London

9Limitless opportunities

10Hooks are deterministic.

11There are not enough woman in AI

12Modeling for Efficiency is key. For usage for cost for outcomes.

13Good evaluation sweet is the key

14How to use claude

15Evals

16Prompts are critical

17Prompt is more important than the model

18Context control is key.

19Evals, inference optimisation, context management

20Prompts are important than models

21It is not always about the model - the application and settings in General are more important + expensive models are not always the right ones

22Emphasis on Evals.

23Evals and context engineering

24Claude is more powerful than I previously thought

25It’s not the model

26Model is not always the answer

27Prompt and context engineering is the key to extract the best use of AI

28Building around the model, not changing the model

29Inference, Evals

30Knowledge

31Claude isn’t just for writing code

32Eval Token management

33Prompt caching

34Context engineering, evals

35Claude eases our work

36The takeaway is that the value is in the application you build around the model, that prompt and context engineering are even more relevant today and that evals are essential

37I tried to vibe code some of the exercises and I was surprised with the results but as the trainer mentioned I didn’t learn much. I have to revisit some later

38Informative

39This is the level of detail we need to even begin being able to solution Claude.

40That 'Just use a better model' is rarely the correct answer.

41- Opus/sonnet with no reasoning may be cheaper/more efficient than Haiku with it

42Metrics are very important and how you show and convey them could make a huge difference for clients

43Testing in AI in more relevant than ever

44Models alone are not going to provide you the value/outcome to your customers, you need to have prompts, skills, routing, context management, evals. Evals are the most important thing, take time to understand the problem you are solving, define success criteria, and create tests.

45Be systematic on building the agent

46Anthropic basecamp allowed me to get a very valuable AI boost. While I still need to complete it I encourage such training for all.

47Claude's model and tooling info.

48Claude and other AI tools do not displace the need for engineering good practice, but engineering practices need to accommodate the strengths and quirks of AI-based development.

49The agent is only as good as you can make it. Make sure you understand how it concluded or came to a decision and you can explain it.

50Anthropic's api is expensive

30-Day Intentions

What will you build for yourself or apply at work? · n=52 responses

1A true multi agent system

2Agents to demoncapabloity

3Evals for sure

4a task agent

5First personal agentic app

6Evals

7Simple applications with agents

8Vacation

9Enterprise architecture agents

10Look to apply the concepts into my daily activities and be ai first.

11Almost everything that was covered in the course I will be looking to apply for the next AI project.

12Better prompts, more granular skills

13An agent

14For myself there will be many subproject type environments that I will utilize to become more efficient in day to day work.

15Apply my knowledge for value proposals

16Agentic cybersecuirty platform

17Reworked evals

18Doing repetitive tasks with agents

19Add more skills, use the context effectively.

20I will continue to build the NLQ application I am building

21Inference optimisation,prompt optimisation, evals

22Analyse architecture design against the template and requirements

23Build Claude agent application

24Evals as confidence KPIs for our Agentic Operations platform

25Evala

26New ai agents and agent swarm

27Coordinator plus subagents AIDLC

28Evals and how to think about it when doing it.

29Will be building a multi swarm agent

30Context engineering. Better evals..

31I will apply AI diagnosis methodology, evals and inference optimization.

32Enabling strategic conversations

33Some operations related stuff to reduce process and effort

34Automating documentation and proposals

35Eval

36Enabling developers with Claude code

37Study documentation on Claude website, build more agents, use cases.

38I mainly focus on how the cost of the project can be reduced by using right approach and also on prompt engineering and context engineering

39I want to play with the platform and build more examples

40Build with Claude and use for client use cases

41Claude update documentation learning agent.

42Automation tooling for my day to day work.

43- Evals - Measuring numbers

44Honestly, I'd like to dive deeper into the entire lifecycle of a real AI application and understand its end-to-end workflow—from prompt engineering and context engineering to evaluations (evals) and everything in between.

45Something amazing

46I am going to dive deep into the Evals and context engineering more. Build shared knowledge which my teams can hse

47Probably internal tools to improve my everyday work

48I've several personal and work topics: - improve our coding env with AI to improve delivery ... - make personal agent for trading

49I have a number of evaluation projects to which I would like to apply Claude Code and agent-based solutions, which will help to consolidate my understanding of the information I have learned during the course.

50An agentic use case qualifier

51Agent to help engage teams to plan and track their yearly goals. Keep them on track.

52Skills

Day 1 responses	0
Day 2 responses	70 (0% of Day 1)
Matched pairs	0 (0% of Day 2)
Orgs resolved	0 of 0 respondents matched to named org
Day 2-only respondents	70 (100.0%) — no NPS data; no Day 1 data, excluded from persona & confidence analysis
⚠ Low Day 2 rate	Only 0% of Day 1 respondents completed Day 2. Push the survey link during or immediately after the closing session.