A practical guide to what actually works
No hype. Just what works.
What AI is, what it isn't, and what you need to know before using it.
Go ahead — flip to Section 2 and start using AI today. You have our permission. But before you use AI with any client information, come back and read "A Note on Privacy and Confidentiality" below. The rest of Section 1 can wait. The privacy section shouldn't.
You've heard the hype. AI is going to change everything. It's going to take your job. It's going to make you superhuman. It's the biggest thing since the internet. It's a bubble that's about to burst.
Here's what's actually true: AI is a tool. A remarkably capable one, but a tool nonetheless. Like any tool, it's useful for some things and not for others. It works well when you know how to use it and poorly when you don't.
This guide isn't about the future of AI or what it might do someday. It's about what you can do with it today, in your practice, to get real work done. No hype. No hand-wringing about the singularity. Just practical application.
Let's get grounded.
AI — specifically the large language models you'll be using — excels at a particular set of tasks. Understanding these strengths helps you know when to reach for it.
Generating first drafts. AI can produce coherent text quickly. Emails, proposals, content, summaries — anything where you need words on a page, AI can give you a starting point. You'll still need to edit, but editing is easier than creating from scratch.
Transforming content from one format to another. Take meeting notes and turn them into action items. Take a long document and create a summary. Take bullet points and turn them into prose. AI handles these transformations well.
Processing documents you share with it. Upload a PDF, a contract, a report, or a spreadsheet, and AI can read it, summarize it, answer questions about it, or pull out the information you need. You don't have to copy and paste — just share the file directly.
Searching for current information. Most AI tools can now search the web in real time. Need to know who just got promoted at a company before your meeting? What a new regulation says? AI can look it up and summarize it for you.
Explaining things in plain language. Complex topic you need to understand quickly? AI can give you an accessible explanation, tailored to your level of familiarity.
Brainstorming and generating options. When you're stuck, AI can produce ideas. Not all of them will be good, but having ten options to react to is better than staring at a blank page.
Matching patterns and styles. Need something written in a particular tone? More formal, more casual, more direct? AI can adjust its output to match what you're looking for. Some tools can even remember your preferred style across conversations.
Synthesizing information. Give AI multiple inputs — a document, some context, your constraints — and it can pull them together into something coherent.
Learning your preferences over time. Most AI tools now remember your context across conversations. Tell it your role, your communication style, and your typical clients once, and it will apply that context going forward.
Equally important is understanding the limitations. This is where the hype falls apart.
AI does not know your specific situation. It has no knowledge of your clients, your history, your business, or your relationships unless you provide that context. Every piece of context you want it to consider, you have to share explicitly.
AI does not verify everything it tells you. Even when AI searches the web, it can misinterpret what it finds or present information with more confidence than the source warrants. For anything consequential, verify independently — especially numbers, dates, legal claims, and quotes.
AI does not have judgment. It cannot tell you whether an email is appropriate to send, whether a proposal is priced correctly, or whether a decision is wise. That judgment is yours.
AI does not replace expertise. It can help an expert work faster. It cannot make a novice into an expert. If you don't know enough to evaluate AI's output in a given domain, you're flying blind.
AI output is a starting point, not a final product. If you use AI well, you'll find yourself editing, refining, and adding to what it produces — not copying and pasting verbatim.
AI can be confidently wrong. This is worth emphasizing. When AI makes an error, it doesn't hedge or express uncertainty. It states the error as fact. Maintain appropriate skepticism.
AI doesn't always know when it's out of date. Even with web search, AI may not find the most recent update to a regulation, a just-announced policy change, or yesterday's market move. For time-sensitive matters, verify directly with the original source.
The AI landscape is moving fast, but here's what matters for your purposes: several tools can do the work this guide describes, and they're more similar than different. The two standalone options most relevant to solo professionals are Claude (claude.ai, made by Anthropic) and ChatGPT (chatgpt.com, made by OpenAI). Both offer free tiers, paid plans with more capability, and the ability to upload documents, search the web, and create files. Google's Gemini is also capable, especially if you're embedded in Google Workspace.
What's new and worth noting: AI is now showing up inside tools you already use. Microsoft 365 Copilot can draft documents, analyze spreadsheets, and prepare presentations using AI — and as of early 2026, it can draw on Claude's technology for more complex, multi-step work. If you have a Microsoft 365 subscription, you may already have AI capabilities you haven't explored yet. The same is true in Google Workspace with Gemini.
For this guide, any of these tools will work. The prompts and workflows are designed to be tool-agnostic — the principles transfer. Start with whichever tool you can access most easily. If you want a detailed comparison, see the Tool Comparison in your supplementary materials.
If you haven't used an AI tool yet, go to claude.ai or chatgpt.com and create a free account. Claude's free plan includes file uploads, web search, and memory across conversations. ChatGPT offers a free tier and an $8/month "Go" plan with expanded access. Both paid tiers (Claude Pro at $20/month, ChatGPT Plus at $20/month) unlock stronger models and higher usage limits — but the free versions are more than enough to work through this guide. Have one conversation — ask it anything. That's enough to get started.
Think "conversation," not "command." AI isn't a search engine. It's more like a collaborator who responds to what you give it. If the first response isn't right, say so. "That's too formal." "Can you make it shorter?" "I meant X, not Y." The best results often come from a few exchanges, not a single prompt.
Think "tool," not "replacement." AI handles certain tasks well. It doesn't try to replace your expertise. It enhances it. Use it to amplify what you're already good at, not to outsource things you should be doing yourself.
Think "experiment," not "commitment." Try things. Some will work, some won't. You'll develop a sense for where AI helps you and where it doesn't. This takes some trial and error, and that's fine.
Before we go further, we need to talk about what you share with AI tools — because this matters, and it's often either ignored or overblown.
The basic reality: When you type something into an AI tool, that information is transmitted to and processed on servers owned by the company that built the tool. It's not just between you and your computer.
This doesn't mean AI tools are unsafe. But it does mean you should be intentional about what you share, just as you would with any third-party service.
Conversations may be stored. AI providers retain conversations, at least temporarily. This enables features like conversation history and allows companies to review interactions for safety and quality.
Your data may be used to improve the AI. On free tiers especially, your conversations may be used to train future versions of the model. This doesn't mean other users will see your specific information, but it does mean your content becomes part of the training process. Most paid plans offer the ability to opt out.
Employees may have access. Staff at AI companies may review conversations for safety monitoring, debugging, or quality assurance. This access is typically limited and governed by internal policies, but it exists.
Data handling varies by provider and plan. Each tool handles data differently, and policies differ between free and paid tiers. If this matters for your work, read the specific terms for the tool you're using.
Regulatory and contractual obligations. This is the most concrete risk for many solo professionals. If you work in healthcare, finance, or legal services — or if you've signed NDAs or confidentiality agreements — you may have specific obligations about where client information can be stored and processed.
Confidential information exposure. If you paste client financial details, proprietary strategy, or legal matter specifics, that information now exists on someone else's servers.
Anonymize by default. Instead of "My client John Smith at Acme Corp," use "My client" or "A manufacturing company." You lose nothing, and you protect everything.
Use paid tiers for client work. Most paid plans offer options to opt out of training data usage and provide stronger privacy commitments. Turn these settings on.
Know your obligations. If you're in a regulated industry, understand the rules before you start.
When in doubt, leave it out. If you're unsure whether something should be shared with AI, don't share it.
AI tools are not uniquely dangerous, but they're also not magic black boxes where your information disappears. For most of what solo professionals do — drafting communications, preparing for meetings, creating content, researching topics — AI tools can be used safely with basic precautions. For work involving confidential client information or regulatory constraints, more care is warranted.
When in doubt, anonymize. When stakes are high, verify your obligations. And always remember that the convenience of sharing more is not worth the cost of sharing too much.
The rest of this guide is organized around practical application.
Section 2: Your First Five Wins. Five specific workflows where AI delivers immediate value for solo professionals. Each includes step-by-step instructions, example prompts you can adapt, and common pitfalls to avoid. This is the heart of the guide.
Section 3: When Things Don't Go As Expected. What to do when AI output misses the mark — diagnosing problems, refining output, finding your voice, and knowing when not to use AI at all.
Section 4: Building Your System. How to move from "using AI occasionally" to having it as a reliable part of how you work. Identifying your highest-value opportunities, creating your own prompt library, and building the habit.
Let's get to work.
Five workflows you can use today. Each one solves a real problem in your practice.
The AI tools available today can do more than you might expect. In addition to generating text, most tools can now search the web for current information, read documents you upload (PDFs, Word files, images), and remember your preferences across conversations. We'll point out where these features are especially useful throughout the workflows.
It's not that you can't write the email. You've been doing this for years. The problem is the friction — the 15 minutes spent finding the right words for a delicate situation, the draft you rewrite three times before sending, the referral ask you keep putting off because you're not sure how to phrase it.
AI doesn't replace your judgment about what to communicate. It handles the craft of how — the tone calibration, the word choice, the structure. Your job shifts from writing from scratch to reviewing and refining, which is faster and requires less emotional energy.
Who are you writing to? What do you need to say? What's the relationship? What's the desired outcome? Don't worry about polish — just get the facts down.
The more specific you are, the better the output. Include the relationship dynamic, any sensitivities, and your preferred tone. If you have prior correspondence, paste it in.
Read the draft critically. Does it sound like you? Is anything missing? Has the meaning shifted? Edit directly or ask for revisions.
Make your final adjustments. The best AI-assisted emails feel like they were written by you on your best day — clear, intentional, and human.
If you took notes during a call or have prior emails in the thread, paste them directly into your prompt. AI can extract context from rough notes — you don't need to organize them first.
The quality check at the end teaches AI to self-review. Use this pattern in every prompt you write.
For additional communication scenarios — responding to angry clients, addressing non-payment, saying no to requests — see the Supplementary Prompt Library.
You wanted to be direct about a deadline. AI turned it into three paragraphs of gentle suggestions. If the draft is softer than you intended, that's a signal — push back and tell it to be more direct.
"I hope this message finds you well." "Please don't hesitate to reach out." If you see filler that adds nothing, delete it. Better yet, tell AI upfront: "No pleasantries. Get to the point."
AI rephrases things to sound smoother, but smoother isn't always more accurate. Read the draft against your original intent. Did the core message survive the translation?
A formal email to someone you've known for ten years feels wrong. A casual note to a new prospect feels unprofessional. Tell AI about the relationship — it changes the output dramatically.
You've been meaning to ask a longtime client for a referral but keep putting it off because you're not sure how to phrase it without sounding transactional. You spend two minutes prompting AI, get a draft that's warm and specific, adjust one line to match your voice, and send it. The whole thing takes four minutes. The client responds the next day with a name and a warm introduction.
A prospect you quoted three weeks ago has gone silent. Your instinct is to write "Just checking in..." but you know that's weak. You tell AI what you need — a follow-up that's confident, not desperate, with a clear next step. Sixty seconds later, you have something you'd actually send. You edit one word and hit send at 4:52 PM.
You have a call with a prospect you've never met. You know the basics — their name, company, maybe how they found you — but not much else. You want to show up prepared. Not scrambling for small talk, not asking questions you could have answered with five minutes of research, not winging it and hoping for the best.
The real goal isn't just "being informed." It's confidence — yours and theirs. You want to ask questions that make them feel understood. You want to demonstrate that you're the kind of professional who does their homework. You want them to walk away thinking, "This person gets it."
You know preparation matters, but "do your research" is vague advice. What should you actually look for? AI helps you focus. Give it the context — who you're meeting, what you're hoping to accomplish — and it can surface what's most relevant. It can give you insight into their industry: trends, recent news, how the current economy is impacting businesses like theirs. This kind of context turns generic questions into specific ones that demonstrate you understand their world.
The result isn't just being informed. It's walking in with clarity about what matters and the confidence that comes from knowing you're prepared for the conversation you're actually about to have.
Who are you meeting? What kind of conversation is this — first call, discovery, kickoff, difficult conversation? What do you hope to accomplish?
Ask AI for a briefing on the person, company, and industry. Include what you already know so it builds on your context rather than starting generic.
Ask AI to flag what's uncertain or potentially outdated. If it provides citations, check them. If it can't cite a source, verify independently.
Push for questions specific to this person and situation — not generic icebreakers. Questions that show you understand their world and uncover what matters most.
Ask AI what questions they're likely to have based on their role and situation. Prepare brief, confident answers.
If your AI tool has web search enabled, turn it on for this workflow — it can pull current company info, recent news, and industry trends in real time.
Turn on web search for this one — AI can pull current company information and recent news in real time.
Verify anything you'll mention in the meeting — especially names, titles, recent events, and company details.
Push for questions specific to this person and situation. If AI gives you "What are your biggest challenges?" ask it to go deeper.
Ten shallow talking points aren't as valuable as deeply understanding one or two things that actually matter. Don't mistake having facts for having insight.
AI can tell you the company raised funding last quarter. It can't tell you their last consultant overpromised and burned them. Stay alert for context that isn't in the research — the dynamics, the politics, the unspoken concerns.
Instead of winging it or spending an hour on scattered research, you spend 15 minutes with AI: company background, industry pressures, and five questions tailored to their situation. You verify two key facts before the call. When you ask your second question, the prospect pauses and says, "That's a really good question." You weren't lucky. You were prepared.
You need to produce a document that matters — a statement of work that prevents scope creep, an engagement letter that sets the relationship on solid footing, a case study that helps close the next deal, a recommendation report that represents your best thinking. These aren't quick emails. They're documents with structure, with sections that need to work together, with stakes if you get them wrong.
Sometimes you're starting from an old version you can adapt. Sometimes you're starting fresh with notes from a call. Either way, the work isn't hard because you don't know what to say. It's hard because shaping it into something you'd put your name on takes more effort than it should.
AI is good at structure. Tell it what kind of document you need and what it should accomplish, and it can suggest what sections to include, what order makes sense, and what you might be forgetting. For documents you create regularly, it can generate solid first drafts you refine rather than build from scratch.
For higher-stakes documents, AI can help you think through completeness. What should a good version of this cover? What questions will the reader have? What risks am I not addressing? You stay focused on the substance while AI handles the scaffolding.
What type of document? Who will read it? What must it include? What's the risk if something is missing?
Starting fresh or adapting something? Both are valid. If you have a prior version, upload it directly. AI works better from examples than descriptions.
Ask AI what sections to include. Review the outline before building the draft — it's easier to restructure a skeleton than a finished document.
Generate drafts individually. More control, better results than asking for the entire document at once.
Ask AI: "What's missing? What questions will the reader have? What's ambiguous?" This is where AI catches what you overlook.
Read the final draft against your intent. Does it say what you mean? Does it sound like you? Would you put your name on it?
Before uploading any contract or legal document, review the privacy guidelines in Section 1. Consider anonymizing sensitive details first.
If the draft could have been written for any consultant, it's not done. Add your specifics — your methodology, your language, your perspective.
If you don't understand a clause, don't include it. For anything with legal implications, get professional review.
AI makes things sound more confident than they should be. Read every commitment and ask: "Can I actually deliver this?"
AI likes to be thorough, which can turn a two-page letter into six pages. More isn't better. Cut anything that doesn't serve the reader or protect you.
You've used the same engagement letter for years. It works. Clients sign it. You stopped thinking about it. Then you paste it into AI and ask: "What's missing? What's outdated?" You add: "I want this document to reflect confidence, expertise, and a partnership of equals — not someone hoping to win approval." AI flags three things: vague cancellation terms, no IP provision, and tentative language. That was you eight years ago. You spend 30 minutes rebuilding it. The next client who receives it is signing with the version of you that exists today.
You've built your practice through relationships, referrals, and reputation. You're not trying to become internet famous. But you know that being visible — having content that demonstrates your expertise and keeps you top of mind — makes everything else easier. Referrals come more often when people remember what you do. Prospects convert faster when they've already seen your thinking.
The problem isn't that you don't have anything to say. You have years of experience and insights. The problem is that turning that expertise into content takes time you don't have. So it doesn't happen, or it happens inconsistently.
You already have the expertise. Years of client work, problems solved, lessons learned. AI doesn't generate that — you do. What AI does is help you package it. Turn a client insight into a LinkedIn post. Turn your approach into a newsletter. Turn scattered thoughts into a clear positioning statement.
The bottleneck isn't ideas — it's the time and energy required to turn them into something shareable. AI compresses that dramatically. You provide the raw thinking; AI helps shape it.
An insight from a client conversation. A pattern you've noticed. A question you keep getting asked. Write it down in rough form — even a few bullet points.
Who is this for? What format — LinkedIn post, newsletter, bio? What should this accomplish — visibility, credibility, lead generation?
Share your rough notes and tell AI what to produce. The more of your actual thinking you include, the less generic the output.
Does it sound like you? Does it say something worth saying? Cut the fluff. Sharpen the insight.
"Does this say something specific, or could any professional have written it? Is the hook strong?"
Add your specific example, your distinctive phrasing, your point of view. Content that builds relationships is content that could only come from you.
If you skip adding your perspective, stories, and point of view — you'll sound like everyone else using AI. The content that builds relationships could only come from you.
Your goal isn't more content — it's staying visible to the right people. One post that makes someone think "I should call them" beats ten that get scrolled past.
You're not trying to become an influencer. You're trying to stay top of mind with your network and demonstrate credibility to prospects. Write for those people.
Your bio, your LinkedIn summary — these shape first impressions. If they still describe who you were years ago, you're making it harder for the right opportunities to find you.
After a client call, you jot three sentences about a pattern you've noticed — the same mistake you keep seeing. Later, you give AI your rough notes and ask it to shape them into a LinkedIn post. Ten minutes later, you have a draft. You sharpen a few lines, add a specific example, and post it. Two days later, someone you haven't spoken to in a year sends a message: "Saw your post — it's exactly the situation I'm dealing with. Do you have time for a call?" You didn't create content to create content. You turned what you already know into something visible. And it started a conversation that otherwise wouldn't have happened.
A client sends you a 40-page report and asks for your thoughts by tomorrow. A prospect mentions an industry term you've heard but can't quite define. You're considering expanding into a new service area but don't know where to start.
This isn't about preparing for a specific meeting — we covered that in Workflow 2. This is about building your understanding: of a topic, a document, an industry, an opportunity. The kind of learning that makes you better at what you do, not just ready for the next call.
You've built your expertise over years. But the world keeps moving. Staying sharp — and expanding what you can offer — requires learning. And learning takes time you don't have.
AI processes and synthesizes information faster than you can read it. A 40-page report becomes a one-page summary. An unfamiliar industry becomes navigable in 15 minutes. A complex decision gets broken down into implications you hadn't considered.
This isn't about replacing your judgment. It's about supplying you with better inputs, faster. You still decide what matters. You still apply your experience. AI just gets you to informed faster than you could get there alone.
What's the topic, document, or question? How deep do you need to go? What will you do with this understanding? The purpose shapes the request.
Asking AI to explain something from its knowledge, or giving it a specific document? If you have the document, upload it directly.
"Tell me everything about X" produces overwhelm. Instead: "Give me a 10-minute overview focused on what matters for advising clients in Y industry."
Follow up: "Go deeper on point 2." "What are the implications?" "What am I missing?" The best insights come from the conversation, not the first answer.
Ask AI to flag what may be outdated, uncertain, or worth verifying. The higher the stakes, the more you verify.
Understanding isn't the goal — doing something with it is. What's the key insight? What will you do differently? Is there a next step?
Upload documents directly — PDFs, Word files, and images all work. Turn on web search when you need current industry data or recent news.
Turn on web search for current industry data and competitive landscape.
AI can misinterpret, oversimplify, or confidently state something outdated. If you're going to advise a client or stake your credibility on it — verify from primary sources.
Reading a summary gives you the shape of something, not depth. Know the difference between "I've read the summary" and "I understand this."
"Tell me everything about X" produces a flood, mostly irrelevant. The tighter your question, the more useful the answer.
The best insights come from follow-up: "Go deeper on point 3." "What am I missing?" If you're not asking follow-up questions, you're leaving value on the table.
A prospect in an industry you've never served reaches out. Five years ago, you would have declined — too much ramp-up time. Now, you spend an hour with AI. You ask for an overview: how the industry works, what pressures it faces, what terminology to know. You ask what mistakes outsiders typically make. By the end of the hour, you're not an expert — but you know enough to ask good questions, listen intelligently, and recognize whether this is a client you can help. Halfway through the meeting, the prospect says, "You seem to really understand our world." Three months later, they're a client. A year later, that industry is a meaningful part of your practice. You didn't just use AI to save time. You used it to grow.
We presented these workflows separately, but they often flow into each other. Research (Workflow 5) might surface an insight that becomes content (Workflow 4). Meeting prep (Workflow 2) might generate a follow-up email (Workflow 1). A proposal (Workflow 3) might require research you didn't anticipate (Workflow 5).
As you get comfortable, you'll find your own patterns. The workflows aren't a checklist — they're a toolkit. Use what you need, when you need it.
Diagnosing problems, refining output, finding your voice, and knowing when to stop.
You've seen what AI can do when you give it the right input. But it won't always go this smoothly. At some point — maybe already — you'll get output that misses the mark. The draft will be generic. The tone will be off. The AI will confidently produce something that's just... wrong.
That's normal. The goal isn't to avoid these moments — it's to know how to handle them.
When the output misses the mark, resist the urge to start over or give up. First, diagnose what happened.
"Write me an email" gives AI nothing to work with. The output will be generic because the input was generic.
Fix → Add who, what, why, and tone.AI doesn't know your relationship with this person, the history, or the stakes. It fills in the blanks with assumptions.
Fix → Provide the backstory.You asked for a draft, a subject line, three variations, and a social post. AI tries to do everything and does nothing well.
Fix → One clear task per prompt.Some tasks aren't good fits for AI — especially those requiring your specific judgment or emotional intelligence.
Fix → See "When should I not use AI?" below.Ask yourself: Did I tell it who I am and who this is for? Did I explain why this matters? Did I specify format, length, and tone? Did I ask for one clear thing? Fix what's missing. Run it again.
The output is in the ballpark. It's not wrong — it's just not quite there. This is where most AI work actually happens — not in the first prompt, but in the conversation that follows.
Be specific about what's wrong. "This isn't right" doesn't help AI. "The tone is too formal for this client" does. "Paragraph 2 overpromises — tone it down" does.
Point to the part that needs fixing. Don't ask for a complete rewrite when only one section is off.
Tell it what to keep. "The opening is good. Keep that. Rewrite the second half to be more direct."
Know when to edit yourself. If AI is 80% there, it's often faster to edit the last 20% yourself than to keep iterating.
The content is right. The structure works. But when you read it, it doesn't sound like something you'd say. It sounds like... AI. Generic. Polished in a way that feels hollow.
This is one of the most common frustrations — and one of the most fixable.
Describe your voice in the prompt. Don't assume AI knows how you write. Tell it: "I write in short, direct sentences. No fluff." The more specific, the better.
Give AI examples of your writing. The fastest way to teach your voice is to show it. Paste two emails you've written that capture your voice, and ask AI to match that tone.
Watch for AI tells. Certain phrases signal AI-generated content: "I hope this email finds you well," "Please don't hesitate to...," excessive hedging, everything in bullet points. If you see these, cut them.
Once your voice description works, save it in your AI tool's memory. In Claude, add it to User Preferences. In ChatGPT, add it to Custom Instructions. Every conversation will start with your voice already calibrated.
You've worked through the five workflows. Now you're facing a task we didn't include — and you need to figure out a prompt on your own. This is graduation.
Try this: "I need help with [task]. I'm not sure how to ask for this effectively. Can you help me build a prompt that would get a good result? Ask me clarifying questions first." Let AI help you build the prompt. Then use what it produces.
Knowing when not to use AI is part of the skill.
The task requires genuine human connection. Condolence notes, sensitive conversations, apologies — anything where the recipient needs to feel that you are speaking to them.
You need to think, not produce. If the challenge isn't finding the words but figuring out what you actually believe, want, or recommend — AI can't do that thinking for you.
The stakes are too high for "close enough." Legal documents, regulatory filings, financial commitments — anything where an error carries serious consequences.
You can do it faster yourself. A two-sentence reply doesn't need AI.
The work is what makes you, you. Your unique perspective, your distinctive voice in important communications, the creative work that defines your practice — these are worth protecting. Not everything should be optimized for speed.
The most important question isn't "Can AI do this?" It's "Should AI do this?" Some work is yours to do. Protecting that is part of using AI wisely.
Moving from "I tried AI once" to having it as a reliable part of how you work.
Not every task benefits equally from AI. Some things you'll always do faster yourself. Others become dramatically easier with AI's help. This exercise helps you identify where AI gives you the biggest return in your specific practice.
List 10–15 tasks you do regularly. For each one, assess frequency, time cost, and how much friction it creates. Look for the intersection of frequent, time-consuming, and dreaded. Those are your highest-value AI opportunities.
| Task | Frequency | Time | Friction | Priority |
|---|---|---|---|---|
| Client emails | Daily | 15 min | Medium | ★★★ |
| Meeting prep | 3×/week | 25 min | High | ★★★ |
| Proposals | 2×/month | 2 hours | High | ★★★ |
| Your task here | — | — | — | — |
| Your task here | — | — | — | — |
What to notice: The tasks where AI helps most aren't always the obvious ones. Look for tasks where you're translating (notes to emails, ideas to proposals), tasks where you're stuck at the start, tasks you procrastinate, and tasks where "good enough" is fine.
As you work with AI, you'll develop prompts that work well for your specific practice. Save them.
Start with the five workflows. You already have prompts that work. Save them in your own words — modified to fit your practice.
Add as you go. When you create a prompt that works well, add it to the library before you forget.
Keep it simple. A document with 10–15 well-tested prompts is more useful than a folder with 200 you'll never find. Quality over quantity.
AI proficiency isn't built by reading about it. It's built by using it.
Once a day for two weeks, use AI for one task you would have done yourself. After each use, take 30 seconds to notice: What worked? What didn't? What would I do differently next time? That reflection — not the AI output — is what builds your skill.
We covered this in depth in Section 3. Here's the quick-reference version.
| Question | If yes... |
|---|---|
| Can I do this faster myself? | Skip AI |
| Are there legal, regulatory, or compliance stakes? | Start with professionals |
| Am I unclear on what I actually want to say? | Do the thinking first |
| Does this need to be genuinely original? | Draft it yourself; use AI to refine |
| Does this require real human connection? | Do it yourself |
When in doubt: if you're asking "Should I really use AI for this?" — the hesitation is worth paying attention to.
You've built a foundation. You have five workflows, a prompt library taking shape, and a habit that will build your skill over time.
Now the work is practice. Use what you've learned. Notice what works. Refine as you go.
If you want to go deeper — advanced techniques, industry-specific applications, strategic uses of AI — we're building resources for practitioners who've mastered the basics.
Tell us what you're working on. The next guide starts with what matters most to the people using these tools every day. Reach us at fieldworkguides.com, or just reply to any email from us.
Welcome to Fieldwork Guides.
Now get to work.
© Fieldwork Guides