Zero to AI — Episode 10: Creating Your Own AI Assistant

 

Welcome back to Zero to AI, the show where we talk honestly about mid career reinvention, working with AI in the real world, and building a life that actually fits.

Today, we are going to do something very practical. You are going to start creating your own AI assistant. Not some fluffy digital twin that lives in a slide deck. A simple, useful assistant that helps with real tasks in your life and work. So you can spend more time on strategy, creativity, family, or just not drowning in your inbox.

You do not need to be technical. You do not need to spend money. You do not need to wear a hoodie and call yourself a prompt engineer. You just need three things. A clear role. Good instructions. And a place for your assistant to live.

We are going to walk through all of that together.

Creating Your Own AI Assistant: A Practical Guide

Building a personal AI assistant isn’t about creating some futuristic digital twin or wearing a hoodie and calling yourself a prompt engineer. It’s about creating simple, useful tools that help with real tasks in your life and work—so you can spend more time on strategy, creativity, family, or just not drowning in your inbox.

You don’t need to be technical. You don’t need to spend money. You just need three things: a clear role, good instructions, and a place for your assistant to live.

What an AI Assistant Actually Is

People often imagine an AI assistant as this huge, complicated system that logs into everything, controls their calendar, and does their taxes. In reality, your assistant is just three ingredients:

Persona — Who they are and how they behave. For example: “You are my calm, practical planning assistant who helps me organize my week and keeps my plans realistic.”

Tasks — The specific jobs it helps you with. Rewriting emails, summarizing documents, or turning messy notes into next steps.

Home — Where it lives. That might be a custom assistant in ChatGPT, a pattern you reuse in Gemini, Claude, or Copilot, or a browser tool that pops up while you work.

That’s it. Persona, tasks, home.

If you remember nothing else, remember this: An AI assistant is just a well-instructed chat with a purpose.

One Assistant or Many?

Here’s the big trap: most people try to create one super assistant that does everything. Writes, plans, researches, coaches, cooks dinner, and probably fixes the printer. That usually ends in chaos.

The better approach is to think like a manager. In a real team, you wouldn’t hire one person to be your accountant, marketing lead, lawyer, and therapist. You’d hire different people for different roles. Same with AI.

Start with one assistant for one clear role. Once that’s working, add more assistants for other roles. Later, if you want, you can have a “front door” assistant that helps you pick which specialist to use.

Examples of Specialist Assistants

  • Content Assistant — Helps you draft emails, posts, scripts, and summaries
  • Research Assistant — Reads long documents, pulls out key points, compares options
  • Planning Assistant — Helps you plan your week, break projects into steps, and keep things realistic rather than fantasy land
  • Personal Admin Assistant — Helps write replies, tidy up notes, and turn chaos into lists
  • Sales Support Assistant — Helps qualify leads, draft outreach, and tidy CRM notes

Where Your Assistant Can Live

Let’s look at free or low-cost options:

ChatGPT — Good for creating custom personas and reusable assistants. Great for writing, planning, explaining, and brainstorming. Beginner-friendly—you set it up once and reuse it. You can create named assistants with saved instructions, give them a description and clear rules, and pin them to go back to them anytime.

Gemini by Google — Good for people already living in Gmail, Docs, and Sheets. Great for quick replies to emails or turning raw data into insights inside Sheets. You can use a consistent persona prompt whenever you talk to it and let it work inside your existing tools.

Claude — Good for long reading and deep thinking. Excellent for structured analysis and reports, or turning messy notes and documents into something coherent. You can reuse the same persona prompt and upload large files for your assistant to read.

Microsoft Copilot — Good for people using Outlook, Word, Excel, PowerPoint. Helps with drafting replies and documents inside the Microsoft ecosystem. Similar idea: a consistent persona prompt plus clear instructions.

Browser assistants — Things like built-in tools or extensions that summarize pages, draft emails, or capture notes. These are great sidekick tools rather than your main assistant.

Yes, you can do this in other LLMs. The pattern is the same. You define a persona, give it tasks, and talk to it consistently.

Hands-On: Build Your First Assistant in ChatGPT

Let’s build a very practical assistant together: a Content and Planning Assistant that knows who you are, writes in your style, and helps you plan and create content without burning out.

Step 1: Open a New Conversation

Open ChatGPT in your browser and click “New chat.” You’re going to paste a long prompt in a moment. This prompt will turn this chat into your assistant.

Step 2: Paste the Persona Prompt

Copy and paste this whole block into ChatGPT:

You are my personal Content and Planning Assistant. Your job is to help me:

  • Plan my week realistically
  • Turn messy ideas into clear posts, emails, or scripts
  • Keep my tone human, honest, and simple

Before you do any work, always ask me a few questions so you understand the context and my goals. If my instructions are vague, ask follow-up questions. Do not guess silently.

When you reply:

  • Keep your answers clear and structured
  • Avoid jargon
  • Give me options when it makes sense, not just one version

At the end of each reply, ask: “Is this close to what you had in mind, or should we adjust it?”

First, introduce yourself in one short paragraph, then ask me:

  • What kind of work I do
  • What kind of content or planning help I want most
  • How many hours per week I realistically have for content

After you paste that, press Enter. You’ve just given this chat a persona and job description. This is the skeleton of your assistant.

Step 3: Answer Its Questions Properly

Your assistant will now introduce itself and ask you questions. Take a minute and answer them with real detail. This is training.

For example, you might say what kind of work you do, who your ideal audience is, where you post content, and how much time you truly have each week. The more honest you are, the better it can help.

Step 4: Save This as a Reusable Assistant

This is optional but recommended. Once you like how it behaves, you can turn this into a custom assistant so you don’t have to paste the prompt every time.

Depending on your version of ChatGPT, you can go to the area where you can create custom assistants, create a new one, copy the persona prompt you just used into its instructions, give it a name like “Content and Planning Assistant,” and save it. From now on, you just click its name and start talking.

Pro Tip: Organize Using Projects

If you want a clean and reliable way to keep your assistants organized:

  1. After you paste your persona prompt and the assistant is behaving correctly, rename the chat to something clear like “Content and Planning Assistant”
  2. Create a Project in ChatGPT called “AI Assistants” or “My Personal Assistants”
  3. Move the chat into that Project

Now it’s always in the same place, easy to find, ready for reuse. You can repeat this for each assistant you build.

How to Make Your Assistant Actually Good

Most people stop at “I wrote a big prompt, I’m done.” That’s like hiring someone, giving them a job description, and never talking to them again. The quality of your assistant comes from how you interact with it.

Five Rules That Make a Huge Difference

Rule 1: Always Give Context

Instead of saying “Write a post about AI,” say: “You are my Content and Planning Assistant. I help mid-career professionals use AI to reinvent their work. I want a LinkedIn post for people who feel overwhelmed and aren’t sure where to start. Make it honest and encouraging, not hype-driven.”

Your assistant knows its persona, but it still needs context for each task.

Rule 2: Use Examples of Your Real Voice

Give it your own writing to learn from. You can say: “Here are three posts I’ve written that sound like me. Study them and copy the style, rhythm, and level of honesty.” Then paste your posts.

Now when you ask it to write, you can add: “Use the same voice and style as the examples I gave you.”

Rule 3: Tell It How to Ask Questions

Your assistant shouldn’t be a mind reader. You want it to be more like a curious junior who asks smart questions.

You can add this line to your persona prompt: “If I give you a vague request, ask me 3-5 clarification questions before you start. Your questions should be short and practical.”

This stops it from racing off in the wrong direction.

Rule 4: Give It a Format

If you always like things in a certain shape, say so.

For example: “When you plan my week, always use this structure: Weekly theme in one sentence. Top three priorities. A realistic daily plan for Monday to Friday. One suggestion to reduce overwhelm.”

Or: “When you write a LinkedIn post for me, always give me: A hook line. The body. A simple call to action. Don’t use hashtags unless I ask.”

The more specific you are about format, the more it will feel like you.

Rule 5: Fix It Instead of Starting Again

If its first attempt isn’t quite right, don’t wipe it and start from zero. Say things like:

  • “This is too formal, make it more conversational”
  • “Shorten this by half”
  • “Keep paragraphs shorter”
  • “I like points 1 and 3, but 2 is off. Rewrite point 2 only”

You’re shaping your assistant. Over time it gets closer to how you think.

Second Walkthrough: A Planning Assistant

Let’s do another clear, beginner-friendly example. Again, open a new chat in ChatGPT and paste this whole prompt:

You are my Planning Assistant. Your job is to help me:

  • Turn big goals into realistic weekly plans
  • Avoid overloading my days
  • Make sure I leave space for life, family, and rest

Before you make any plan, always ask:

  • What is my main goal for this week
  • What fixed commitments I already have
  • How many hours I actually have available

When you create a plan:

  • Be kind but honest. Tell me when my expectations are unrealistic
  • Suggest what I can drop or delay
  • Always present plans in a simple table or clear list with days and tasks
  • If anything is unclear, ask me questions first. Do not guess

First, introduce yourself in one short paragraph. Then ask me:

  • What my main goal is for the next seven days
  • What other responsibilities I have
  • How many hours I can realistically give to this goal this week

Press Enter. Now answer its questions with real numbers. Not fantasy numbers. Real ones.

Watch how it plans. Then refine it. You might say: “This is still too much for me, cut it by 30 percent. Add time for exercise and family. Make Friday lighter.”

You’re teaching it what “realistic” means for you.

Using the Same Pattern in Other Tools

You might be thinking, “Okay, but what if I use Gemini, Claude, or Copilot?” Good news: the pattern is exactly the same. You just don’t always have a fancy “save this as an assistant” button. Instead, you save the prompt yourself somewhere and reuse it.

In Gemini

  1. Open Gemini in your browser
  2. Paste one of the persona prompts you created
  3. At the end of the prompt, add: “When I say ‘use planning mode’ in future chats, act as this assistant”
  4. Save that prompt in a document or note so you can reuse it

Next time you open Gemini, you can say: “Use planning mode. Here’s my situation this week,” then paste or describe your context.

In Claude

Same idea:

  1. Open Claude
  2. Paste your persona prompt
  3. Tell it: “Whenever I say ‘Planning Assistant,’ I want you to act according to the instructions above”
  4. Save that persona prompt somewhere handy

Now you can start new chats with: “Planning Assistant, I want to plan my next two weeks. Here’s my context.”

Claude is particularly good when you upload documents. So your Research Assistant persona might say: “When I upload documents, first give me a short overview, then a detailed summary, then key risks and opportunities.”

In Copilot

You can’t always store complex personas as neatly, but you can still reuse prompts. Inside Word or Outlook, write something like:

“Act as my Email Reply Assistant. Your job is to help me reply in a friendly, clear way that respects my time. Ask me two or three questions if you’re unsure about tone or details before you draft the reply.”

You can keep that text in a snippet and paste it when you need it.

Understanding Agents and Advanced Workflows

You might have heard words like “agents,” “workflows,” or “autonomous AI.” Under all the buzzwords, the idea is simple: instead of you manually asking for each step, you give your assistant permission to run a series of steps on its own.

For example, a Content Agent might:

  • Take a podcast transcript
  • Pull out key ideas
  • Turn those into three LinkedIn posts, one email, and one summary
  • Put everything into a draft doc

Or a Lead Research Agent might:

  • Take a list of company names
  • Look them up one by one
  • Fill a table with industry, size, and a simple fit score

The basics are the same, just with more steps and sometimes with connections to other tools or data.

If you’re a beginner, don’t worry about full automation yet. Focus first on a clear persona, clear tasks, clear formats, and good back-and-forth interaction. Once that feels natural, you can explore agents and workflows as version two.

ChatGPT Agents in Practice

Let’s look at how you can use Agents to go further, once your basic assistants are working.

Think of your normal assistant as a smart helper that waits for you to ask for each individual task. A ChatGPT Agent is more like a junior team member who can follow a checklist, work through multiple steps, and handle a whole mini-process on its own.

What a Simple ChatGPT Agent Can Do

Content Repurposing Agent — You give it a transcript or long article. It summarizes the content, pulls out key messages, writes several LinkedIn posts, writes an email, and suggests titles and hooks.

Weekly Review Agent — You give it notes from your week. It organizes them into themes, identifies wins, issues, and risks, suggests three priorities for next week, and turns everything into a short reflection.

Lead Research Agent — You give it a list of company names. It looks each one up, finds key facts like industry and size, suggests a simple fit score, and produces a table you can drop into a sheet.

The key idea is that the Agent is allowed to think in steps and follow a repeatable pattern, rather than just answering one question.

How to Set Up a Simple Agent-Style Workflow

If you have access to the Agents or advanced custom assistant features in ChatGPT, you can create one in a very similar way to your basic assistants, just with more explicit steps.

Here’s a simple pattern you can describe in your instructions. Imagine we’re creating a Content Repurposing Agent:

  1. Open ChatGPT
  2. Go to the area where you can create a custom assistant or Agent
  3. Give it a name, such as “Content Repurposing Agent”
  4. In the instructions area, add something like this:

You are my Content Repurposing Agent. Your job is to take long content and turn it into multiple useful formats. Always follow this workflow:

  • Confirm what type of content I have given you and who it is for
  • Create a short summary in plain language
  • Extract the top ten key ideas or lessons
  • Write three LinkedIn posts tailored to my audience
  • Write one email that I can send to my list
  • Suggest five short title or hook ideas

Before you start, ask me:

  • Who the audience is
  • What tone I want
  • Where I will use the outputs

Show me your plan for the steps first. Then carry them out in order. If anything is unclear or missing, ask me questions instead of guessing.

Save it. Now when you open this Agent, you can say: “Here’s a podcast transcript. Audience is mid-career professionals who feel overwhelmed by AI. Tone is honest, practical, and encouraging. Please run your full workflow.”

The Agent will ask the questions you told it to ask, show you the planned steps, work through the steps, and produce a clean bundle of outputs.

Making Your Agent More Robust

You can upgrade your Agent with small, simple rules like:

  • “If a step fails or doesn’t make sense, explain the issue and suggest an alternative”
  • “At the end, give me a one-paragraph explanation of what you did and how I could reuse this workflow in future”
  • “When you write content, always keep paragraphs short and avoid buzzwords”

You’re not trying to build a giant autonomous system. You’re just asking ChatGPT to stick to a repeatable process that you define.

Once you’re happy with one Agent, you can create variations using the same pattern: a Weekly Review Agent that processes your notes, a Planning Agent that turns tasks into a schedule, a Research Agent that compares options and gives a recommendation.

The mindset is the same: clear role, clear steps, clear rules.

Your Action Steps

Here’s what you can actually do after reading this:

1. Pick One Assistant

Don’t build five. Pick one: Content and Planning Assistant, Planning Assistant, Research Assistant, or Email Admin Assistant. Choose the one that would relieve the most pressure this week.

2. Build It

  • Open ChatGPT
  • Start a new chat
  • Paste one of the persona prompts from this article
  • Answer its questions honestly
  • Ask it to help with one real task from your life or work

For example:

  • “Help me turn this messy brain dump into a LinkedIn post”
  • “Help me plan my week so I don’t burn out”
  • “Help me summarize this long document and pull out next steps”

3. Improve It

After each interaction, ask yourself:

  • Did this feel like me?
  • Was this actually useful?
  • What instruction would have made this better?

Then go back to your persona prompt and add that instruction. You’re iterating on your assistant the same way you’d improve a tool or a process.

If you feel ready for the next level, pick one simple process in your life, like repurposing content or doing a weekly review, and turn that into a small Agent-style workflow inside ChatGPT using the extra section we covered.

Final Thoughts

Creating your own AI assistant isn’t a future thing. It’s available to you right now.

You’re not trying to build the perfect, all-seeing assistant that controls everything. You’re building small, focused helpers that make your life lighter and your work more effective.

Start with one assistant. Teach it well. Use it every day for a week. Then decide what the next helper should be.

You’re not replacing your own judgment. You’re creating more space for it.

This article is based on Episode 10 of the Zero to AI podcast, where we talk honestly about mid-career reinvention, working with AI in the real world, and building a life that actually fits.