In partnership with
Your Tax Data, Finally in One Place
Are you tired of hunting down data, fixing errors, and manually updating disconnected spreadsheets?
Tax reporting isn’t a simple as it used to be. You need real-time, flexible reporting so you can confidently make decisions backed by accurate, centralized data.
Learn how bringing all your tax information into one central system automates repetitive tasks, improves scenario planning, and frees your team to focus on strategy instead of data entry.
Whether you operate in one country or dozens, Longview Tax scales with you—reducing risk, speeding up your close process, and helping you optimize tax policies across all jurisdictions.
The reason your AI outputs feel generic, surface-level, and basically just slightly better than a Google search? It's not the model. It's you. Or more specifically, it's how you're talking to it.
I've worked with hundreds of business owners who've adopted AI tools over the past couple of years, and the pattern is almost always the same. They try Claude, ask it to "write a marketing email" or "summarize this document," get something okay but not great, shrug, and go back to doing it themselves. Then they tell anyone who'll listen that AI is overhyped.
It's not overhyped. You're under-prompting.
This edition is about fixing that. Not theory. Not fluff. A concrete framework you can implement today that will fundamentally change the quality of everything you get out of your AI tools.
The Three-Layer Problem
Most prompts fail for one of three reasons, and usually all three at once.
First, no context. You walk up to the AI like it's a vending machine. You punch in what you want and expect magic to come out. But AI models don't know anything about your business, your audience, your tone, or your goals unless you tell them. Explicitly. Every single time.
Second, no constraints. You ask for "a blog post" without specifying length, reading level, structure, what to include, what to avoid, or what action you want the reader to take. Unconstrained prompts produce unconstrained results. Which is to say: bloated, generic, and forgettable.
Third, no role. You're not telling the AI who to be. The difference between asking Claude to "write a sales email" versus telling it "you are a direct-response copywriter who has spent 15 years writing for B2B SaaS companies, and your job is to write a 200-word cold email that leads with a sharp problem statement" is the difference between a C-minus and an A.
The CARE Framework
Here's what I want you to start using immediately. I call it the CARE framework. It stands for Context, Action, Role, and Examples.
Context is the background. Who are you? Who is the audience? What's the situation? What does success look like? The more specific, the better. Don't say "small business owner." Say "a 42-year-old woman who runs a $2M/year e-commerce business selling premium pet supplies, and she's skeptical of AI because her last vendor overpromised."
Action is what you actually want the AI to do. Be precise. "Write" is not precise. "Write a 5-bullet summary in plain English" is precise. "Generate three variations of a headline" is precise. "Rewrite this paragraph to sound warmer and less corporate" is precise.
Role is the persona you're assigning to the model. This single change will improve your outputs more than almost anything else. Think about who the world's best person is for this task. A Harvard-trained negotiator. A seasoned supply chain operator. A conversion rate optimization specialist who's worked on 300 landing pages. Give the AI that role explicitly.
Examples are optional but powerful. When you show the AI what good looks like -- even a single sentence or two of your own voice -- outputs get dramatically better. This is especially true for tone-sensitive work like emails, social posts, or anything customer-facing.
Real-World Application: A Before and After
Here's what a before-and-after looks like in practice.
Weak prompt: "Write a LinkedIn post about AI for business owners."
CARE prompt: "You are a sharp, no-nonsense business strategist who writes LinkedIn posts that get shared by operators and entrepreneurs. Your audience is business owners making $500K to $5M annually who are skeptical of tech hype. Write a LinkedIn post (under 200 words) that makes the case for why AI is not about replacing employees but about eliminating the 3 hours of daily admin work that keeps owners out of strategic thinking. Use a direct, conversational tone. No buzzwords. No corporate language. End with a question that sparks real comments."
The second prompt takes you 90 seconds to write. The output is ten times better. That's not an exaggeration.
Implementing This In Your Business This Week
Here's your actual action plan.
Start by building what I call a Master Context Block. This is a 200 to 300 word description of your business that you paste at the top of every new AI conversation. It covers who you are, what your business does, who your customers are, what your tone is, and what you never want the AI to do or say. Build it once. Use it forever.
Next, create a Prompt Library. Every time you write a prompt that produces a great output, save it. You're essentially building proprietary IP -- a collection of instructions tuned specifically for your business. After 30 days, you'll have a toolkit that lets you get great results in under a minute.
Then, start assigning roles to every non-trivial prompt. Keep a shortlist of 5 to 10 roles relevant to your business. Customer success expert. Financial analyst. Brand copywriter. Technical explainer. Pull from that list based on the task.
Finally, review outputs with the question: "Did I give the AI everything it needed to produce this?" Not "Is this bad?" but "What context was I holding in my head that I never gave the model?" That mindset shift alone will improve your prompting dramatically.
How Long Does This Actually Take?
Building your Master Context Block: about 20 minutes, once.
Adding role and constraints to a prompt: an extra 45 to 90 seconds per prompt.
Building your Prompt Library: passive, just save the good ones as you go.
Weekly time savings from better outputs that don't need heavy editing: conservative estimate is 3 to 5 hours.
That's the math. Spend 20 minutes once and 90 seconds per task, recover 3 to 5 hours per week. Every week. Indefinitely.
The Bigger Picture
AI is not magic. It's leverage. But like any tool, the output quality is directly proportional to how well you use it. A power drill in the hands of someone who doesn't know what they're doing leaves a mess. In the hands of a craftsman, it builds something solid.
You are the craftsman. The AI is the drill. The CARE framework is how you stop stripping screws.
Start with your Master Context Block today. Seriously, block 20 minutes on your calendar after you finish reading this. The version of you who does that and the version of you who doesn't will have a very different Q2.
Ready to go deeper? I built an entire system around this, including prompt templates for 22 common business tasks, a role library you can pull from, and a workflow for continuously improving your outputs without starting from scratch each time. It's all inside the AI Workflow Blueprint.
Reply to this email with the word BLUEPRINT and I'll send you details. It's $47 and it'll pay for itself the first week.
Jordan Hale | The AI Newsroom