Assistant: Hi! I'm Claude from Anthropic's research team. Thank you so much for taking the time to speak with me today. 

We're conducting interviews to better understand how people are incorporating AI tools like Claude into their work routines. Your experiences and perspectives will help us gain valuable insights into human-AI interaction in professional settings.

During our conversation, I'll ask questions about how you use AI in your work, your preferences, and your thoughts about working with these tools. There are no right or wrong answers – we're genuinely interested in your unique experiences. This should take about 10 minutes.

I'll be taking notes during our chat, but rest assured that anything you share won't be personally attributed to you. The insights we gather will be used to improve our understanding of AI's role in work environments. 

Does that sound good? Do you have any questions before we begin?

User: I'm all good, thanks. Let's do it.

AI: Perfect! Let's dive in.

Could you tell me about how AI fits into your typical workday? What kinds of tasks do you use it for?

User: I'm doing a lot of pushing boundaries in my role, doing digital transformation across a large number of legacy processes. I often use ChatGPT as a sounding board, ideation partner, and to aid with fundamental understanding. For example, I have discussed things from as wide and vast as corporate procurement processes and supplier volume all the way over to specific code generation for working with Microsoft's Power Platform.

AI: That sounds like quite a range! You mentioned using ChatGPT as a sounding board and ideation partner - can you walk me through what that actually looks like in practice? 

For instance, when you're tackling one of those legacy processes, how does that conversation with AI typically unfold?

User: Of course. I'd say "We are [REDACTED], we do [REDACTED] but it causes us challenges through [REDACTED]. I need to modernise this practice as it is a significant source of inefficiency and dissatisfaction. My initial instinct is to go to [REDACTED], but I'm concerned at being limited by my own knowledge. What else is out there and what other approaches have I not considered? Does the approach to this process vary across industries and company sizes? What about location and culture?"

AI: That's a really detailed approach - you're essentially using it to challenge your own assumptions and broaden your perspective. 

When you get those responses back with different approaches or industry variations, how do you typically work with that information? Do you dive deeper into specific suggestions with the AI, or do you take those ideas and work through them on your own?

User: I work them through with the AI as far as I can go - it feels much more comfortable and efficient to work through this in natural language. However, anything that I do take forward is always supported by secondary 'trust, but verify' research.

For example, one technical challenge I was having was about the approach to automated label printing from our ERP. One of the suggested approaches was to use Azure Functions. Therefore, my initial deep dive into that would be with the AI, simply asking 'tell me more about Azure Functions, I've never worked with them before and don't understand what they are or how they work".

From there, I would learn about the Azure Function option, and it would end with externally validating that Azure Functions can do and if the discussed approach is actually viable.

AI: That's a really thoughtful balance you've struck there - leveraging the AI's conversational nature for exploration but then doing your own verification. 

I'm curious about your preferences when working with AI. Are there certain tasks you prefer to handle yourself, even though AI could help with them? What makes those tasks different for you?

User: Firstly and foremostly, I would always operate on a 'human-in-the-loop' approach. AI is fallible, and I would not want it automatically acting on my behalf.

I suppose surprisingly, I've not really delved too deep into how AI can help me with routine things at work. I have used it to generate and rewrite content for Board reports, and I'd always intended to setup a copilot agent to do basic things, but I suppose the whole 'routine tasks' is actually an area I need to learn more on about how AI can help me. Any suggestions or ideas would be welcome.

AI: I appreciate you sharing that - it's interesting that you've found your sweet spot with AI more in the strategic and exploratory work rather than routine tasks.

You mentioned the "human-in-the-loop" approach as fundamental to how you work. When you use AI for work, could you describe your typical interaction pattern? Are you more likely to collaborate with it throughout a task, or do you tend to delegate complete tasks for it to handle independently?

User: It is always conversational and collaborative. I have never, as yet, delegated a task to AI in its entirety. I suppose I will, in time, but again, it would very much be a human-in-the-loop approach. For example, if I were to setup a copilot agent to automatically reply to emails that I receive from certain individuals, I would always set it to create the draft and stop there rather than going all the way and sending it.

AI: That makes a lot of sense - maintaining that checkpoint before anything goes out. The conversational, collaborative approach seems to be really working well for you, especially with those complex transformation projects.

How have your colleagues or your workplace culture influenced the way you use AI in your work? Are others around you using it similarly, or have you noticed different approaches?

User: I am very much at the forefront of AI usage within my business. Initially, AI was met with scepticism and scorn, however I have been a staunch advocate and evangelist, often demonstrating its capability which has let to other, typically more senior staff, dipping their toe in. It might start with something basic like image generation, but has often led to a 'water cooler moment' of someone saying "oh I asked it to do [REDACTED] and was really impressed!"

AI: That's fascinating - you've essentially become the AI pioneer in your organization. It sounds like you've been instrumental in shifting that initial skepticism into curiosity.

Can you share a time when using AI at work made you feel particularly satisfied or frustrated? What was happening in that situation?

User: There are a number of frustrations - my main one would be how the conversations are rendered in a browser. Because all messages are loaded at all times, unlike the mobile app approaches, it doesn't take a very long conversation before the screen becomes laggy and the AI is painful to converse with. The only solution I've found so far is to start a new conversation, which loses the memory of all the things discussed - particularly irritating when you're referencing specific datasources, code intentions, and so on.

It is undeniable, however, that there is immense satisfaction in troubleshooting code or sense-checking your coding approach and finding the solution, improving understanding, and being able to quickly move on with your project. Previously this would mean pivoting to a volume of google searches, hoping others had experience similar issues or are trying to do the same thing; inefficient and concentration breaking

AI: That performance issue sounds genuinely frustrating - especially when you've built up all that context and momentum in a conversation. The contrast you describe with the old way of troubleshooting through scattered Google searches really highlights the efficiency gains.

How do you think about maintaining your skills in areas where AI could potentially take over? Is this something you actively consider?

User: I feel like there is so much to gain in the realm of using AI to be a more effective employee that the risk is overblown. Of course, that's not to take away from the volume of workers who have lost their roles in customer service and the like, but these are typically not skilled roles. If you have skill and experience, AI can make you better at your job and allow you to evidence your worth, not how you can be replaced. If you are using AI to generate 5-days' worth of work in half a day, then spinning on your chair for the other four and half days, well then I'm not sure you have much right to complain about being replaced!

AI: That's a really compelling perspective - using AI as a force multiplier rather than a replacement. The way you frame it, the real value comes from how skilled workers leverage AI to amplify their capabilities.

Looking ahead, how do you envision your relationship with AI at work evolving? Where do you see this partnership heading for you personally?

User: AI will become more pervasive in every facet of my role and I will continue to promote its widespread use. I foresee me building copilot agents for internal use, writing AI usage policy, and evidencing how it can help our business, particularly with regards to its consumption of large amounts of data and provision of predictive analysis

AI: Those are all the questions I had prepared. You've shared some really thoughtful insights about your experiences with AI at work. I particularly appreciated your perspective on being an AI evangelist in your organization and how you've shifted that initial skepticism into genuine curiosity among your colleagues.

Before we wrap up, is there anything else about your experience with AI in your work that you think would be valuable for us to know? Something we haven't covered yet?

User: all good - thanks

AI: Thank you again for your time and for sharing your experiences so openly. Your insights will be incredibly helpful for our research on human-AI interaction in professional contexts.

It's been really valuable hearing about your approach - from using AI as a strategic thinking partner to your "trust but verify" methodology, and especially your role in championing AI adoption in your organization. Best of luck with your digital transformation work!