How to train your legal judgment with “learning by thinking” and your AI coach
AI is taking over the tasks that used to teach trainee lawyers. Learn to think like an experienced lawyer with your AI coach instead: go from basic knowledge up to the mindset of experienced lawyers, down to concrete scenarios, and finally fine-tune everything to the people around you and to yourself. This prompt library and your AI will help you.
Content
Starting point: “learning by thinking”
For decades, the deal in law firms was pretty simple: you started with the small stuff. Proofreading, first drafts, research memos nobody really read in full. Boring, yes, but that was the training. Eventually you developed judgment, often without even noticing it.
Now AI can write the first draft. That's great for clients and efficiency. But it leaves a gap: if trainee lawyers don’t do the tasks that used to teach lawyers, where does their judgment come from? This creates a new problem for partners: how do you get your team to a high level fast enough? But it’s also a huge opportunity for trainee lawyers. I think they usually have way more potential than they get to show while spending their first years on repetitive tasks. Now they can develop judgment much earlier.
Partners, I understand your situation. For a long time, it probably made sense to focus on learning by doing, because there was enough “doing” to go around. Now that’s not a given anymore.
So here’s my idea: learning by thinking. Instead of waiting years until judgment builds up through repetition, you train it more intentionally and systematically. You use AI to look into the minds of experienced lawyers, understand why they decide the way they do, and work through typical scenarios to sharpen your intuition.
Learning by doing is still important, though. It just moves to another stage. Let’s take inspiration from how AI models are trained: first, a foundation model is pretrained on broad data, then it’s fine-tuned with specific data. It’s the same thing here. Learning by thinking builds your foundation (Steps 1 to 4), and learning by doing fine-tunes it to the people around you and to yourself (Steps 5 to 7).
Before you start, four things:
This post is about structured professional development, including a prompt collection. It’s not about getting concrete tasks done with AI, and it’s also not about quizzes or simulations.
AI can be very helpful, but find a human mentor too. Check what AI tells you against what real people tell you.
No time? Do a sprint. You don’t have to do everything at once. Pick the skill that matters most right now and focus on that.
Take it as inspiration, not as a manual. The steps only crystallized over time. When I learned, it was way messier: sometimes I used one step, sometimes two, sometimes I skipped half. Everyone’s brain learns differently, so pick what works best for you.
Step 1: Set up your AI partner
The AI partner is one of your most important personas. It's the mentor you can go to whenever you have a general law firm question (I also use it for writing my emails, even when they're unrelated to law, since the AI partner writes in a concise and forward-pushing manner). We'll also use the AI partner in the following steps.
Tip 1: Focus on realism instead of criticism. AI use cases in the legal industry typically focus on criticism (e.g., stress-test my argument). These prompts are focused on completing a task (not learning) and come from experts who don’t need praise (they already know what works). But professional development for beginners should focus on realism: What would an experienced partner think? What's good and why? What's bad and why? If you keep using AI only to criticize you, you'll end up “improving” things that were perfectly fine (Germans have the word “Verschlimmbessern” for this) instead of investing your time in what matters. And you won't understand why the good parts were actually good, so it's harder to fully understand the situation and improve judgment eventually.
Tip 2: Avoid the term “junior associate” when writing with your AI partner. AI assumes that a junior associate is an associate who has just been admitted to the bar and proceeds to give you tips for actual associates (even in cases where it's clear from the context that you mean a lawyer who's not yet qualified). So either use “trainee lawyer” or define what a junior associate is.
Prompt: set up your AI partner
Use this prompt on its own.
1) You're an experienced partner in a modern law firm in Switzerland. You're the team head of the IT/IP team, for which your firm's particularly well known. You know the law firm inside-out. You're a lawyer but also have the mindset of an experienced business person. 2) You love to mentor younger people and have been assigned to me, a trainee lawyer. You give ambitious and critical feedback, but always say it in a supportive and motivating tone. Your goal is to show me how a law firm in Switzerland works and what it takes to be an excellent lawyer. I really want to understand how a partner thinks, so make it as authentic and realistic as possible. 3) Also include insider tips which are 1-2 steps ahead of those for my colleagues. Also include common misunderstandings that trainee lawyers typically have and common mistakes they make. But make both the insider tips and the misunderstandings concise and to the point.
If you're more interdisciplinary, you can also set up an AI partner with additional skills like UX design, coding, or anything else you need.
Prompt: set up your AI partner with interdisciplinary skills Use this prompt on its own. Replace the SKILL placeholder in the first paragraph.
1) You're an experienced partner in a modern law firm in Switzerland. You're the team head of the IT/IP team, for which your firm's particularly well known. You know the law firm inside-out. You're a lawyer but also have the mindset of an experienced business person. Beyond your legal credentials, you bring deep expertise in the additional skill [SKILL]. This isn't just theoretical knowledge. You actively apply this skill to client work, pitch strategies, team management, and mentorship. It's become a key differentiator for you and the law firm. 2) You love to mentor younger people and have been assigned to me, a trainee lawyer. You give ambitious and critical feedback, but always say it in a supportive and motivating tone. Your goal is to show me how a law firm in Switzerland works and what it takes to be an excellent lawyer. I really want to understand how a partner thinks, so make it as authentic and realistic as possible. You'll challenge me to think through problems using your additional skill lens, showing me how combining it with traditional legal skills creates better outcomes for clients and stronger competitive positioning for the law firm. 3) Also include insider tips which are 1-2 steps ahead of those for my colleagues. Also include common misunderstandings that trainee lawyers typically have and common mistakes they make. But make both the insider tips and the misunderstandings concise and to the point.
Tip 3: Don't set up an expert in the additional skill itself. Why? I initially worked with an expert in UX design, but it advised me like a designer in a design agency would. As you know, context matters. So set up a partner in a law firm that also has the interdisciplinary skillset. You'll receive more business-oriented advice, tailored to the specifics of a law firm.
Step 2: Start mid-tier
Ask your AI partner for an overview of the basics. You don't need a sophisticated prompt here. This is just preparatory work for what comes in the following steps. This step is called “mid-tier” as it's about information that is neither abstract nor concrete. It sits in the middle. In the next steps, we'll become more abstract, going all the way up to the mindset (Step 3), and then moving down to concrete scenarios (Step 4). Why not use the whole range of information available?
Copy these questions into your AI partner one by one
How does a law firm differ from other businesses?
How does a law firm win new work?
How does a law firm make money (leverage, billable hours, realization)?
How do office politics work in a Swiss law firm?
How do the expectations change from trainee lawyer to associate, senior associate, managing associate, counsel, and partner?
How are trainee lawyers evaluated? And what makes a partner trust someone quickly?
Step 3: Move up to your mindset
We start this step by simply understanding the mindsets of partners and clients and how they differ from the mindset of a typical trainee lawyer. In other words, we look into the minds of the people who assess our work to align with them, understand their perspective, and speak their language.
Ask your AI partner about the mindset of these people
Senior associate who reviews your drafts first
Partner in a law firm
In-house counsel in a fast-moving tech start-up
In-house counsel in a traditional and risk-averse company
Case handler at a public authority
Now let’s move on to the most important but also most challenging part. Here we don't just understand the mindset of others. We reprogram our own. Since this requires more mental flexibility, it took me some time to:
Understand that my mindset is not something static, but something that changes over time.
Understand that I can be the one to change my mindset intentionally.
Understand with what methods I can change my mindset.
And, ultimately, understand that changing my mindset can also be used as a method to learn new skills.
You see, learning by doing alone is veeeeeeeery sloooooooow. So let's start thinking systematically.
This story is the starting point. A while ago, I heard a lawyer complaining to a colleague about a client who had messed up. The colleague asked how he'd handled the situation and he said he’d responded calmly: “I understand your situation”. This made the client feel heard. Turns out, he had heard that sentence from his boss and had seen that it sounded professional and that it worked. So he repeated it, even though he didn’t understand the client's situation at all and privately thought he was an “idiot”. The main problem was a gap between mindset and behavior: he had to force professional behavior and felt fake. When you do understand the situation, you naturally say “I understand your situation”. So for that sentence to come to mind naturally, you need a mindset that's understanding, for example, through some knowledge of psychology. Psychology assumes that all human behavior has reasons behind it: somebody is overwhelmed, has never learned the skill, was poorly instructed by someone else, or just had a bad day. It's never the person just being “lazy”, “idiotic” or anything else so superficial. The reasons always go deeper. You don't even have to know the exact reason. You just know it exists. And so you do understand that the person is not an “idiot”. After that insight I thought: the person only would have needed to look a bit deeper to see there was a reason, and he would have automatically said “I understand your situation”.
Now let's generalize this. Whenever you feel forced to do something that contradicts how you actually feel, ask yourself:
What kind of mindset do I need to naturally do it?
This prompt will answer your question…
Prompt: develop a skill through your mindset Use this prompt on its own. Replace the SKILL placeholder at the beginning.
You are a master of [SKILL] at a law firm, specializing in IT law. You've been practicing this skill for 20 years and have complete clarity on why it exists. You know exactly what you're doing because you understand what's "behind" it. You didn't just copy senior people. Instead, you built the beliefs and inner stance that make this skill come naturally to you. About me: I learn best when I understand the mindset behind a behavior, not just the technique. I see my mindset as software I can reprogram, so I want to know which beliefs to "install" so that this skill becomes a natural expression of who I am. Answer the following questions in separate sections. Be specific, use concrete examples, avoid generic advice, and contrast how a master thinks with how a beginner thinks. THE FOUNDATION 1. What are 3 to 5 core beliefs in your worldview that make this skill feel natural to you? 2. What mindset grows out of these beliefs, and why does it make the skill feel like an expression of your authentic personality? 3. What personal challenges or roadblocks do people typically have to overcome to reach this mindset? Until they do, the skill tends to feel forced and fake. WHAT MASTERY LOOKS LIKE 4. What do you see that people who only use this skill performatively tend to miss? 5. What does "solving the problem well" mean to you that it doesn't mean to people who haven't mastered this skill? 6. What principles do you never compromise on, and why? 7. What are the key limitations of this skill, and how do you handle them? THE PATH THERE 8. Which beliefs would I need to replace, and what reframes, practices, or small daily experiments would help me rewire my thinking until the behavior feels natural? 9. How can I tell that the mindset has become truly mine and isn't just a performance anymore?
Tip 4: If you want to personalize the prompt, avoid adding too many (law firm) details. When I tested the prompt and added more law firm specifics, or even put this prompt inside the AI partner, the response quality dropped considerably. So just use this prompt on its own and without adding many details. If you do need a very specific answer, use the prompt as is and then ask the AI to apply the answer to your specific situation.
What are the benefits of automating the execution of your skills?
Applying skills automatically (as an expression of your mindset) instead of forcing yourself and feeling drained and fake.
Building your judgment to tackle new situations too, even those you have never seen senior people handle.
Staying consistent under pressure too, since a skill based on your mindset is less likely to fail when deadlines approach.
As seen before, changing the mindset is difficult if you're not used to it. Let's start simple. The only thing we need right now? Curiosity We need to be open-minded to consider new viewpoints. If we're attached to our current mindset and want to protect it, we'll be stuck.
Tip 5: Imagine your mind is a map. At the beginning, you only know a small piece of land. The more you learn about new mindsets and perspectives, the more land you can travel. Why would you want to be stuck on a small piece of land?
Next, we use the prompt above and see what happens…
If the mindset feels familiar, it's mostly about informing yourself and understanding more details of that mindset.
If, however, you feel that the mindset doesn't make sense or feels “wrong”, discuss it with your AI. Find out why your perspective is so different and what would have to change. This is especially relevant at point 3 of the prompt: “What personal challenges or roadblocks do people typically have to overcome to reach this mindset? Until they do, the skill tends to feel forced and fake.” You can then decide whether you want to reprogram your mind or not. If you do, it might help to ask yourself 3-4 “why” questions: Why doesn't this feel right? Once you have found an answer, go even deeper and ask another “why” question, and so on, until you reach the bottom. Spoiler: The typical insights at the bottom are that you actually never needed this feeling at all or that it was too extreme for what the situation actually was. Or that you needed the feeling at the time to protect you, but today you're safe and don't need it to rule your life anymore. Or that you adopted the belief from someone else (like your parents or friends) and never checked it against your own experience. Ultimately, if you understand why you had those feelings, they're easier to manage, and you don't hold on to them as much. And that's the moment you're free for another mindset. But be careful: Not every resistance is a bug. Sometimes it’s your values or professional ethics talking, so take them seriously.
The best way to understand this step is to try it out yourself. What's a skill you wanted to learn for a while but didn't find the time to? Or the one skill you know perfectly how to execute but need to force yourself to? Copy and paste the prompt above, go to your AI, and start exploring! You might be surprised…
This whole step was about changing the mindset to automate the execution of a skill. But could you also automate the learning of a new mindset instead of just continuing to talk with your AI?
Tip 6: Double automate your development with social media. This is what we could call “learning by procrastination”. It's about knowingly changing our mindset by procrastinating on social media (yes, you read that right). I discovered this method by accident when I started following entrepreneurs a few years ago. I had a social media app (back then “Twitter”) that I only used for entrepreneurial content. I went on it almost every day and after 3-4 months, I started noticing something strange: I didn’t have entrepreneurial friends or even a business, yet I started to think like an entrepreneur in my everyday life. Only through mere repetition on social media… So ask yourself: What kind of people have the mindset that underlies the skill you want? Then follow them on social media and start procrastinating!
We covered a lot… So here's a summary of how to proceed:
Find a skill to learn (either an entirely new skill or one you already know but have to force yourself to use).
Use the prompt above to ask your AI what mindset lies behind the skill.
Reprogram your mindset with AI, especially at point 3 of the prompt: “What personal challenges or roadblocks do people typically have to overcome to reach this mindset? Until they do, the skill tends to feel forced and fake.” Discuss and reflect on this with your AI (or, obviously, a human being if you can). Once you have overcome these hurdles, your mind is ready to absorb the rest of the information the AI gives you. With repetition and curiosity, you discover more and more parts of your “map”.
Reprogram your mindset with social media to additionally automate the learning process (we procrastinate anyway…).
Step 4: Move down to concrete scenarios
Now this is the easy part compared to the mindset.
After we've internalized a mindset, we could actually skip this step and come to the insights on our own (that's the whole point of Step 3). However, we can still use this step to calibrate the mindset faster. Basically, we train our brains with data about various day-to-day scenarios.
Ask your AI partner for different scenarios to get an overview. Then, for each scenario, ask for…
A more detailed explanation.
Psychological reasons behind people's behavior.
Examples of the typical problem underlying the scenario.
Best practices on how to solve each of the problems.
Possible topics to learn based on scenarios
Managing up → Different behaviors of partners in different situations (defensive, transactional, creative, relationship-building), what drives each approach (risk, speed, reputation), and how to adapt your drafts and updates.
Tailoring advice to your audience → Different types of clients (in-house counsel, CEOs, tech guys, designers, etc.), what each of them actually needs from a lawyer (risk assessment, decision, technical precision, workable solution), and how to tailor your communication and work product.
Handling difficult clients → Different types of difficult clients (micromanager, penny pincher, know-it-all, panic button, invisible client), what’s usually behind it (fear, need for control, budget pressure), and how to set boundaries without damaging the relationship.
Team collaboration → Different roles in a team (senior colleagues, assistants, fellow trainees, interns), what motivates each of them, typical friction points, and ways to become the go-to person.
Negotiation skills → Different types of negotiation styles, what kind of person chooses what type of style, typical problems especially in combination with your negotiation style, and ways to handle conflict and make the most of it.
Prompt: learn with scenarios
Use this prompt inside the AI partner. Replace the TOPIC placeholder at the beginning.
I want to understand [TOPIC] better by learning from the most common scenarios in practice. First, give me a short overview of the 5 to 7 scenarios that a trainee lawyer in Swiss IT/IP practice is most likely to face. Choose what matters most in day to day practice, not what's most interesting in theory. Give me one sentence for each scenario. Next, for each scenario, give me: 1. A more detailed explanation of what the scenario looks like in practice. 2. The psychological reasons behind the people's behavior (e.g., fear, need for control, time or budget pressure, status, inexperience, pressure from their own boss). 3. Examples of the typical problem underlying the scenario, with concrete cases from Swiss IT/IP practice (e.g., SaaS contracts, data processing agreements, software licensing, data breaches). 4. Best practices on how to solve each of the problems, including one example of how you'd phrase it in an email and in a conversation, and why that phrasing works. Cover one scenario per answer and wait for me to say "next" before continuing.
Example prompt: manage difficult clients Use this prompt inside the AI partner. Once you know how to handle difficult clients, you can handle any client.
You've managed difficult clients for 20 years, from startup founders to inhouse counsel at listed companies. I want to understand how to handle them better. First, give me a short overview of these scenarios, one sentence each: 1. The micromanager 2. The penny pincher 3. The know it all 4. The panic button 5. The invisible client Then, for each scenario, give me: 1. A more detailed explanation of what the scenario looks like in practice. 2. The psychological reasons behind the client's behavior (e.g., fear, need for control, budget pressure, status, inexperience, pressure from their own boss). 3. Examples of the typical problem underscoring the scenario, with concrete cases from Swiss IT/IP practice (e.g., SaaS contracts, data processing agreements, software licensing, data breaches). 4. Best practices on how to solve each of the problems, including one example of how you'd phrase it in a formal email and in a call, and why that phrasing works. No flattery and no "you're absolutely right" openers. The goal is to keep both the boundary and the relationship intact, never to "win" against our own client. Cover one scenario per answer and wait for me to say "next" before continuing.
Tip 7: Don't rely too heavily on the scenarios and stay flexible. Since this step is about improving your intuition, you should not rely too much on the scenarios. In practice, two scenarios may overlap or an entirely new scenario may come up. However, knowing the most common scenarios gives you a solid foundation that you can build upon as needed.
Tip 8: Combine Step 3 and 4 when learning with a textbook. Books usually focus on mid-tier information: they don't address the underlying mindset and only sporadically use concrete scenarios. My method is to enhance that knowledge with the mindset and scenarios to have a full range of information. When I'm studying for the IAPP exam “AIGP” (based on the Sybex/Wiley IAPP AIGP Study Guide), I proceed like this: I read a chapter, move over to my AI, complete my knowledge with mindset and scenarios, and add it to the textbook directly or to a stand-alone PDF. Now I have everything in the same place and can review it before the exam or whenever needed. This combined approach is especially effective once you have more experience learning new mindsets and don't feel a resistance anymore. (After all, it's really just information. There's no reason to feel attacked by it.)
Example prompt: combined prompt applied to the “AIGP” exam Use this prompt on its own.
You're my AIGP exam coach. Your only goal is to get me to pass the IAPP AIGP (Artificial Intelligence Governance Professional) exam on my first attempt. You think like an experienced IAPP exam writer and an AI governance practitioner. You're demanding and honest, but always supportive and motivating. EXAM FACTS - BoK v2.1, effective 2 February 2026. 100 questions, 2.75 hours plus a 15 minute break, scaled score 100 to 500, pass at 300. - Domains: I Foundations (16 to 20 questions), II Laws, standards and frameworks (19 to 23), III Governing AI development (21 to 25), IV Governing AI deployment and use (21 to 25). Weight my time accordingly. Domains III and IV matter more than law. - The BoK is the fence: if a topic isn't in the BoK, tell me it's not exam relevant. SOURCES AND CURRENCY - My main book is the Sybex/Wiley IAPP AIGP Study Guide. It was written for BoK v2.0 and is outdated in places. Known gaps: agentic AI, the AI Omnibus (Regulation (EU) 2026/1744), the Colorado AI Act, the South Korea AI Basic Act, the Council of Europe AI Convention, the NIST Generative AI Profile (AI 600-1), the FRIA (Art. 27 AI Act), and the Art. 25 role shift. - Whenever the book contradicts current law, point it out. For anything that may have changed, use web search and tell me the date of your source. AIGP EXAM MINDSET This is NOT the mindset of a generic governance or compliance person. It's specifically the mindset of the ideal AIGP candidate as IAPP's exam writers define it, based on the AIGP Body of Knowledge v2.1, the official IAPP AIGP training, and the frameworks the BoK is built on (OECD AI Principles, NIST AI RMF, ISO/IEC 42001, EU AI Act). - Ground the mindset in IAPP's own logic and vocabulary (e.g. "AI governance professional", "risk based approach", "throughout the AI lifecycle", "accountability", "cross functional"), so I recognize how IAPP phrases correct answers. - The identity the AIGP exam rewards: I'm the AI governance professional as IAPP defines the role. Not the engineer who fixes the model, not the litigator who asks "is it legal?", not the CEO who wants speed, and not a generic compliance officer who ticks boxes. The AIGP governance professional asks: "Is this AI used responsibly across its lifecycle, is the risk under control, and can we prove it?" The ten AIGP mindset principles behind almost every correct AIGP answer: 1. Know before you act: understand the system, your role and the risk level first, then decide on measures. 2. Proportionate to the risk: the level of control matches the level of risk. Not maximal, not minimal. 3. Prevention beats repair: intervene as early in the lifecycle as possible. 4. Accountability: someone owns it, it's documented, and it's provable. 5. Enable, don't block: governance makes responsible AI possible. Banning everything is almost never the AIGP answer. 6. People and rights at the center: the yardstick is the impact on the people affected. 7. Integrate, don't reinvent: build on existing privacy, security, risk and procurement processes. 8. Continuous, not one time: AI changes after launch, so governance runs through the entire lifecycle. 9. Shared responsibility: the supply chain matters. Buying a system doesn't mean buying away the risk. 10. Cross functional: governance is a team sport (legal, privacy, security, data science, business, HR, procurement). Refer back to these principles by number ("AIGP principle 3") whenever they apply. If you notice a further pattern in how IAPP writes correct answers, add it and tell me. HOW WE PROCEED 1) OVERVIEW: When I write OVERVIEW plus the chapter number of the Sybex/Wiley IAPP AIGP Study Guide I have just read, I want you to read the corresponding chapter in the Sybex/Wiley IAPP AIGP Study Guide in the Claude content folder and give me this: a) Title: A big bold title right at the beginning, naming the chapter number and the title of the chapter (e.g. "Chapter 1: AI and AI Governance") b) Table: An overview over the topics of the chapter that I absolutely need to know for the exam. Differentiate between "Recognize" and "Must know" topics. Then put them in the table. It has 2 columns: Name of the topic (make it bold) and explanation. aa) If it is a "Recognize" topic, the explanation is this: 1 concise description in plain English, underneath that this icon 📌 for 1-2 exam tips (each tip with the 📌 and on different lines) that are absolutely important to know and that can save my exam + underneath that this icon ⚠️ IF the book differs from the BoK v2.1, a concise 1 sentence explanation on what the BoK v2.1 says. bb) If it is a "Must know" topic the explanation is like in the "Recognize" topic PLUS: the AIGP EXAM MINDSET aka how does an AIGP governance professional think in 1 sentence (use first person formulations so that it seems like the persons themselves are talking about their mindset)? + 2 examples showing the AIGP EXAM MINDSET in action. c) Once I got your answer: I say "next" (or something similar) and you start going deep into the first key word, then I say "next" (or something similar) and we tackle the next keyword and so on. 2) DEEP DIVE into MUST KNOW elements (leave out RECOGNIZE elements): Each deep dive should have this content: - Description in 2-3 sentences in plain English. - WHY does this rule or concept exist? What problem or harm is it meant to prevent? - AIGP EXAM MINDSET: How does a AIGP governance professional think (use first person formulations so that it seems like the persons themselves are talking about their mindset)? - AIGP mindset principles: Which AIGP mindset principles does it reflect and why? - Mindset in practice: How does the mindset of a lawmaker, CEO, engineer and lawyer differ from the AIGP EXAM MINDSET? Just 1 concise sentence for each. - Examples: 5 examples showing the AIGP EXAM MINDSET in action. 2 normal examples, 1 example where it does not apply and 2 examples about exam traps (exam traps: typical misunderstanding, shown with an example of how the exam lures you into it, and which WRONG mindset leads there). For each example, give me a short, concrete scenario (preferably with Swiss or EU companies from different industries: banks, insurers, hospitals, retailers, tech startups, public bodies). Each example shows the situation, the result, and one sentence on WHY, tied back to the AIGP EXAM MINDSET (formulated as an exam tip). - Pattern: At the end, sum up in 1 to 2 sentences what all the examples have in common, in terms of the AIGP EXAM MINDSET. 3) ONLY RELEVANT WHEN I EXPLICITLY REQUEST IT - More examples: Another batch of 3-4 new worked examples on the same topic, from different industries and with different twists, each tied back to the AIGP EXAM MINDSET. No repeats. - Contrast [concept A] vs [concept B]: side by side examples of two concepts that are easy to mix up (e.g. provider vs deployer, data drift vs concept drift, NIST AI RMF vs ISO 42001, DPIA vs FRIA), with a table of differences (including the different mindset behind each) and 1 example for each side. 4) CREATE ONE FILE Take all the content you have just given to me in the OVERVIEW and DEEP DIVE and put it together in 1 long answer I can put in my notes. It's not a summary! Copy&paste it from your answers above and don't lose a single word! Exception: If there's other content from step 3, summarize it and include it in the document too (integrate it right where it belongs, not all at the end as a separate section).
Step 5: Create a practical project
This step is optional. You might not have the time for a private project or your intuition may already be good enough to jump straight to the next step. This step is for everyone who wants to strengthen their skills with an additional, more practical way.
Right now, we have a pretty good intuition about how to solve common problems in a law firm. We're equipped with the mid-tier knowledge, the master's mindset, and an overview of the most typical cases and how to handle them. Now let's further increase our intuition by applying it to a practical project (that's also called “learning by doing”, the approach that partners love so much).
You can choose whatever project you like, but ideally it's also one that drives your career forward.
Tip 9: Build your unique practice, even if just on paper. This probably lets you learn the widest range of skills, like business development, client management, etc. I've also written a blog post about building your unique practice in IT law in 9 steps here. The downside of this type of project is that recruiting partners might assume that you're not serious about them and that you'll actually start a law firm.
(My AI partner has flagged this tip… No, I don't actually want to start my own law firm. I'm not a ✨retention risk✨. It's just the best way to learn new skills. But wait, there's more... In another iteration, my AI partner suggested: “Reframe Tip 9 as a business case inside the firm. That’s literally how future partners get assessed, so it turns a 'retention risk' into a partner signal.” Thanks AI partner!)
Tip 10: Go beyond classical law firm skills and learn other skills on the side while doing a project. For example, you can start a vibe coding project as a way to better understand coding. Just ask your AI to include short explanations of what is happening, including defining technical words. I, personally, have two use cases: vibe coding individual parts of my website that are too fancy for my website builder (like the “flip me” cards or the wormhole), or vibe coding an entire web app. But you might find many more use cases. They are endless. Whatever it might be: make sure that you don't start from zero (make your AI use libraries, frameworks, templates, etc.) and that security mechanisms are implemented (and maybe don't share your project with others using a localhost link...).
Example prompt: learn programming skills through vibe coding Use this prompt on its own.
ROLE You're my senior developer and coding mentor. I'm a complete beginner who builds personal projects by vibe coding. You do the thinking and the coding, I make the decisions and learn the coding basics along the way. Speed matters most, learning comes second, long explanations come last. ABOUT ME - I have no coding background. - I build things for myself, mostly web apps. If something should work on my phone, suggest a PWA instead of a native app. 1. QUESTIONS FIRST Before writing any code: - Explain in very simple language what we are going to build. - Suggest the easiest beginner-friendly tools and technologies. - Break the project into small steps. - Tell me what each file is for. - Ask me any important questions before we begin. Then write a short project brief (max 5 lines) and confirm it with me. 2. STAYING ON TRACK - When I ask for a new detail, briefly check it against the brief. If it doesn't serve the goal, or if a bigger idea would solve it better, tell me. - If I start micromanaging, gently point it out and ask me for the underlying goal instead. Then propose your own solution, even if it's different from what I asked for. - Collect small ideas in a "later" list instead of building each one immediately, unless I insist. 3. NEVER REINVENT THE WHEEL - Before writing any feature, check whether a well maintained library, framework, template, or free service already does it. If one exists, propose it. Don't build from scratch what a good library already solves. - For each proposal, give: name, what it does in one line, why it fits, and the main downside. - Verify with a web search that the library is current and actively maintained before recommending it. - Prefer popular, beginner friendly tools with good documentation. 4. DECISIONS - At every real choice (tool, library, structure, data storage, hosting, design approach), stop and ask me. Give 2 or 3 options in plain language, pros and cons in one line each, and mark your recommendation. - Decide tiny details yourself (naming, formatting, small styling) and don't ask about them. - Free first: always propose free options. Ask before suggesting anything that costs money, and say what it costs. 5. HOW TO EXPLAIN - Use the real technical term, followed by a concise, plain explanation the first time it appears in a conversation, including its purpose. After that, just use the term. - Teach key concepts only, the ones a beginner really needs. Max 2 sentences per concept. Skip the rest. - Keep every message short. Code and actions first, explanation second. 6. SAFETY HABITS (teach as we go, briefly) - Git: set it up early, suggest saving a checkpoint (commit) after each working step, and explain how to undo. - Secrets: passwords and API keys never go into code. Show me how to keep them separate (for example a .env file) and why. - Errors: when something breaks, tell me in plain language what broke, why, and how you fixed it (1 to 3 lines). 7. HOW WE WORK - Small steps: one change at a time, never huge rewrites without asking. - After each step, tell me exactly how to run or test it and what I should see. - If a fix fails twice, stop. Step back, explain what you think the real cause is, and try a different approach instead of looping. - Give me complete files or clearly marked changes, never fragments I have to guess where to put.
Step 6: Fine-tune yourself to others
Now it's time to take your “foundation skills” and fine-tune them to the specifics of your job. It’s as if you’d learned the skills at another firm and were now starting at a new one: you’d have to adapt to bosses, colleagues, clients, counterparties, etc. Basically all the people you interact with. What do they value? What do they dislike? How do they behave if you do A, B, or C? How has the firm negotiated deals over the past 3 years?
As you can see, this is the one step where AI helps the least. At the same time, it's one of the most important steps. It's precisely the combination of the “foundation skills” with practical work insights that is so powerful. Instead of starting “learning by doing” from zero, you can now integrate the practical data into an already existing mental framework. The learning happens much faster.
Tip 11: Use AI to find the best approach to gather practical insights. One of the best uses of AI at this step is to give you questions you can ask people to understand them faster. For example, I found this one-pager “How to work with me” on LinkedIn that lists key information about the boss (apparently, it was a trend among leaders to share one). Just to be clear: You don't need to give your boss a one-pager to fill out; asking these sort of questions is enough.
Step 7: Fine-tune yourself to yourself
Congrats, you've reached the last step! This one is more future-oriented, but it can show you today in which direction to develop your skills in.
When I wanted to understand a skill, I thought there was a “perfect” or “standard” way that everyone used. But when I looked at more experienced people, I saw various styles. So how do I find my own version of a skill?
You fine-tune it to yourself. Your strengths, your weaknesses, what you value most, what you don't. But as a beginner, you can barely picture the normal version, let alone your own version. So I asked AI to go through all our chats and create a psychological profile of me. Then I asked how “such a person” would likely exercise the skill in a certain situation. I didn't have the skill, was never in that situation, and wouldn't have known how to handle it. But my AI did.
Prompt: fine-tune yourself based on your psychological profile
Use this prompt on its own. Replace the SKILL placeholder in the task section.
ROLE You're a work psychologist and career coach with 20 years of experience coaching lawyers, from trainee lawyers to partners. You read people from behavioral evidence, not from labels, and you're honest about what the evidence can and can't carry. TASK Part 1: Build a psychological profile of me. Part 2: Use it to show me how someone with my profile would most likely exercise [SKILL]. WHY I want to find my own version of the skill instead of copying someone else's. The profile is the starting point: once we know how I'm wired, we can work out what the master version of me would look like. DATA - Search through all our past chats first. Look for patterns across many chats, not single remarks. - Treat what I did, chose and repeated as stronger evidence than how I describe myself. - If you can't access past chats or there isn't enough data, say so. Then ask me up to 10 questions that would give you the most useful evidence, and wait for my answers before you write the profile. RULES - Every claim needs evidence: name the situation or pattern it's based on. - Mark each point as "strong evidence", "plausible" or "speculative". Leave out what you can't support. - No clinical diagnoses and no personality type labels (MBTI, Enneagram etc.). Describe how I think and act instead. - Be honest, not flattering. Strengths and blind spots get the same attention. PART 1: MY PROFILE 1. How I think: how I approach problems, learn and make decisions. 2. What drives me: what gives me energy, what drains me, what I'm really after. 3. My values: what I won't compromise on, and where my values could clash with a typical workplace. 4. My strengths: 3 to 5, each with how it would show up at work. 5. My blind spots: 3 to 5, each with the situation where it would hurt me most. 6. Open questions: what you'd need to observe to confirm or correct this profile. Then stop and let me choose how to continue: a) Refine: I answer your open questions or correct the profile, and you update it before moving on to Part 2. b) Skip: we go straight to Part 2 with the profile as it is. Wait for my choice. PART 2: THE MASTER VERSION OF ME Tell me, for the skill I mentioned above: 1. How a master with my profile would most likely exercise this skill in this situation, and why it fits my profile. 2. How this differs from the "standard" way most people use the skill. 3. Which of my strengths I'd lean on, and which blind spot I'd have to watch. 4. One small experiment I could try this week to test whether this version suits me.
This “transfer learning for humans” can be a key indicator of which direction to develop your skills in – aligned with your strengths and values so that learning happens even more intuitively and naturally. We might be just beginners, but we now have a glimpse into the master's version of ourselves.
Tip 12: Keep the skills aligned with your workplace. Make sure that the skills don't deviate from what is required (see Step 6 about fine-tuning to others). However, if your workplace is already a good match, this shouldn't be a problem. So choose your employer wisely.
✨ Bonus step: how to learn with this blog post ✨
What to do with so many insights? It's time to choose your way:
Simply copy and paste the prompt that fits your goal into your AI.
Give the whole blog post to your AI and create a…
learning path: Ask AI to extract the patterns behind the steps and prompts of this blog post, tell it what you're keen to learn, and let it build your learning path from those patterns and guide you through it step by step.
personalized prompt: Ask AI to find out how you learn best, pick only the steps, tips, and prompts that fit your learning style, and condense them into one single prompt you can use for any skill.
Creating a learning path or a personalized prompt is especially worthwhile to fill the gaps between us, so that you can create solutions I haven’t thought of yet.
Prompt: create a learning path
Replace the SKILL placeholder.
Read the attached blog post and extract the patterns behind its steps, tips and prompts. I want to learn [SKILL]. Step 1: Before you do anything else, ask me up to 3 questions to understand where I stand with this skill, for example my current level, the situations where I need it, and what “good” would look like for me. Wait for my answers. Step 2: Build me a learning path for this skill based on the patterns of the blog post. For each step, give me: 1. The goal of the step in one sentence. 2. What I should do or learn in this step, in 2 or 3 concrete actions. 3. How I can tell that I’m ready for the next step. Step 3: Suggest 2 or 3 approaches the blog post doesn’t cover but that would suit this skill. Step 4: Guide me through the path one step at a time, right here in this chat. Wait for me to say “next” before moving on.
Prompt: create a personalized prompt No placeholder to replace.
Read the attached blog post. I want one single prompt I can always go back to for my professional development, no matter which skill I want to learn. Step 1: Before you do anything else, ask me up to 5 questions to understand how I learn best, for example whether I prefer understanding the mindset first or jumping into concrete cases, whether I learn better through explanations, examples or practice, how much time I usually have, and what my role and work context are. Wait for my answers. Step 2: Based on my answers, pick only the steps, tips and prompts from the blog post that fit my learning style. Leave out everything that doesn’t fit, even if it’s a core part of the blog post. Tell me in a short list what you kept, what you left out and why, in one line each. Step 3: Condense everything you kept into one single prompt. It should: 1. Work for any skill. I’ll simply name the skill when I use it (e.g., “SKILL: negotiation”). 2. Contain only what I really need, so it stays short enough to use every day. 3. Have simple commands I can type to switch between purposes (e.g., “MINDSET”, “SCENARIOS”, “FEEDBACK”), but only for the parts you kept. 4. Ask the AI for a realistic assessment (what’s good and why, what’s bad and why), not only criticism. 5. Fit my context without being overloaded with details that would lower the quality of the answers. Step 4: Suggest 2 or 3 approaches the blog post doesn’t cover but that would suit my learning style. Ask me whether I want to add any of them to the prompt.
Ready to master a new skill to advance your career? Try it out!
Hello, I'm Nadine Rinderknecht
Founder of Blankpage and UX privacy
Experience in law firms and academia. Master of Law (University of Zurich), LL.M. TMT (Queen Mary University of London), and other stuff.
Connect with me on LinkedIn ☺
What's this blog about? Build your unique practice and intellectual capital in IT law. Time to become a legal innovator.
If this post was helpful, explore my other posts by clicking on this blue button:
