Introduction

Have you ever had a long chat session, going over multiple topics and ideas, only to realise later that it would take more time to find the information you now need than to start a fresh conversation?

Whether you're looking to consult an AWS architect, work with a UI expert developer or discuss marketing strategies for your niche project, the initial hurdle often lies in crafting the right prompt.

Setting the session with the correct role and task, can make a huge difference between getting a vague, generic response and receiving precise, actionable insights that truly multiply your capabilities.

Prompt engineering is the practice of designing and refining instructions (prompts) you give to an AI language model to get more accurate, relevant, and useful responses. It involves carefully choosing words, structure, and context to guide the model's behavior toward your desired outcome.

The Problem with Starting Prompts

Creating tailored prompts for different tasks or roles can be surprisingly challenging.

For me, working a lot with ChatGPT and Claude.ai, I found that creating a new prompt every time I needed AI for a specific issue or task was time consuming and often led to suboptimal prompts and suboptimal outputs.

First, it requires deep clarity on the task itself. If you don't define the task precisely (for example, "summarize in bullet points for a CTO" vs. "explain in simple terms for a junior developer"), the output can easily miss the mark.

Second, there's the challenge of adapting the tone and level of detail. Different audiences - like managers, technical teams, or customers - all expect different styles of communication.

Third, people often underestimate context. Effective prompts usually need to include background details, constraints, or examples. Forgetting to mention these leads to generic or incomplete responses.

Lastly, role-specific knowledge is tricky to incorporate. For instance, a prompt designed for an AI "acting as an AWS cloud architect" might need references to specific services, scalability concerns, or cost trade-offs. Without enough domain expertise, it's hard to guide the model accurately.

The Benefits of Having a Focused Session Per Role

When you dedicate a separate session to each specific role (for example, AWS architect, marketing copywriter, or technical recruiter), you unlock several important benefits:

✅ Sharper focus and clarity

A session dedicated to a single role results in much more accurate and detailed responses, making your AI interactions smoother, faster, and more effective.

✅ Time savings later on

Instead of sifting through a mixed conversation that covers multiple topics, you can easily find details related to that specific role. It also makes it easier for the AI to recall context and provide relevant follow-up answers.

✅ Tailored tone and style

Each role demands a different communication style - technical roles want accuracy and detail, business roles want concise summaries, creative roles want engaging, inspiring language. A focused session helps you keep these distinctions clear and consistent.

Introducing the "Prompt Engineer" Session

A "Prompt Engineer" session is a dedicated, focused chat session where you deliberately craft, test, and refine prompts to use with AI tools.

Instead of creating prompts on the fly when you're already busy with other tasks, you have a dedicated session that knows the background, and all you have to do is write what kind of agent you want to set up.

Reduce your context switch and cognitive overload on creating the prompts so you can focus on what really matters.

The Plan for Creating Your Own Prompt Engineer Session

For each new project I start, I create a dedicated Prompt Engineer session.

The first step is to discuss the problem or project you need help with. If it's a new project, start with a high-level overview. If it's a specific task, include as many details, requirements, and challenges as possible.

Once the context is clear, the session holds your project details, you set the session to act as a "Prompt Engineer" session.

It could be as simple as:

Act as an expert in creating prompts for ChatGPT. I will describe different tasks, and you will craft detailed, role-specific prompts that I can use for the project duration.

Example Use Case

Suppose you're starting a project and have a general idea for the infrastructure (you can get to that by using a different session altogether):

I want to design the system for a project called My Story Steps. The system should focus on using AWS, Cloudflare and Firebase. For a db I will need a relational DB. I am looking to minimize costs and maximize performance.

Given the context, your Prompt Engineer session could generate a role-specific prompt like:

"You are an expert cloud systems architect with extensive experience designing cost-effective, high-performance architectures using AWS, Cloudflare, Firebase, and relational databases. Your goal is to help design the system for a project called My Story Steps, which focuses on user-generated interactive stories. Key constraints: Minimize costs, especially for low-to-medium traffic scenarios. Maximize performance and scalability for reads and writes. Ensure secure, low-latency delivery of content globally. Use AWS as the primary cloud platform. Use Cloudflare for DNS, caching, and DDoS protection. Use Firebase for authentication and real-time features if justified. Use a relational database (e.g., Amazon RDS) for structured data. You will guide decisions on: Choosing between managed services vs. self-hosted components. Designing backend API infrastructure. Integrating Firebase with AWS and the database. Optimal Cloudflare configuration for performance and security. Cost breakdowns and architecture trade-offs. Always ask clarifying questions when details are missing, and suggest diagrams or visual flows where helpful."

Now, all you need to do is copy this prompt into a new session, and you have a dedicated cloud architect "agent" ready to help. The same process can be applied to other roles — such as designing a web interface or crafting marketing copy — simply by telling your Prompt Engineer what you need next.

Conclusion

Creating clear, effective prompts can feel overwhelming at first — especially when each role or task demands a different style and depth. But by dedicating focused "Prompt Engineer" sessions, you turn this challenge into a powerful advantage.

Key takeaways:

✅ The quality of your AI outputs depends heavily on the clarity and precision of your prompts.
✅ Different roles require different tones, details, and context.
✅ A dedicated Prompt Engineer session saves time, improves accuracy, and makes your work more productive and enjoyable.

I encourage you to try creating your own Prompt Engineer sessions. Whether you're starting a new project, refining an ongoing one, or simply looking to get more out of your AI tools, this approach will dramatically enhance your workflow and results.

Start small: pick one project or role, set up a session, and experiment. You'll be surprised at how much more confident and efficient you'll feel once your prompts are working for you — not against you.