Tools of the Future: How to Build Your First AI Agent (Beginner's Guide) 23|04|2025 | Destiny
Read the online version here👇
AI Agent 101: Beginner’s Step-by-Step Guide
Today, I want to walk us through something exciting that we've built at DwellPro - an AI-powered chatbot that helps with travel planning, business tools, and wellness resources without needing to code or worry about backend complexity.
Whether you're exploring DeepSeek, using no-code platforms, or just curious about what's possible with modern AI tools, this guide walks you through how to prompt your way into creating something powerful, lightweight, and entirely client-side. Little or no server-side rendering.
The future isn't coming, it's here.
If you've ever wanted to understand how these chatbots work or wished to build your own, this post is for you!
I'll keep it simple, as it may become a series.
AI is still in its early stages. It is not as saturated as many think. The real wave has not even started yet.
A time will come, if not here, when you’ll assign names to your AI employees, each responsible for specific tasks at specific times.
One will patiently wait for the other to complete their part, while the other will review, refine, and provide feedback where needed.
Just like any team in an organisation, there will be a "Chief of Staff AI" who gives the final approval on every given task or project in a typical digital ecosystem.
Your role is to oversee each step of the process and ultimately provide the final endorsement for real-life implementation or deployment.
I'm not writing this for you to read, nor to become a better writer—nope, my friends. I'm writing this to solidify what I've learned and to share it with anyone it may be concerned.
What We Built
We created a friendly, floating chatbot that lives in the corner of our website, ready to assist anyone with:
💼 Business tools (translation, currency conversion) and all in real time.
✈️ Travel planning (destination info, booking tools, weather, and time of day of any country in the entire sphere) via tools like open weather API and Google map API (updated)
🧘 Wellness resources (meditation tools, inspirational quotes) fetched via web scraping.
If you're not familiar with the term API (Application Programming Interface), think of it as a way for different software or systems to talk to each other. Let me explain it in a way that's easier to grasp.
Imagine you're shopping online. You browse through items, add them to your cart, and then place an order. Behind the scenes, it's the API that helps fetch all the necessary information from different parts of the website and puts it together for you to complete your order.
Or think of it like ordering a meal at a restaurant. You tell the waiter what you want, maybe mention you're allergic to gluten or that you want extra chilli.
Even though you don't see the ingredients being prepared in real time, your preferences are passed along and reflected in the final dish you receive. That back-and-forth communication is very similar to how an API works.
In the same way, when we use something like the OpenWeather API, it collects weather data from various sources and delivers it to us in real time. It’s simple, super effective, and efficient.
In simpler terms, an API is what lets different apps, websites, or services talk to each other and get things done behind the scenes.
I hope that helps make the concept a little clearer.
We also built free tools that you can download to supercharge your business—and yes, they're completely client-side with no server-side rendering.
That means you can download the tool and use it without an internet connection.
Crazy, right? We love making life easier. You can have them for FREE if you want to.
They're vital tools you'd normally pay for. DwellPro File Tools (Convert images and generate QR codes with logo support & More
Many people fear losing their jobs because AI is capable of performing nearly everything humans can—often faster, more accurately, and without the need for rest.
The only thing that truly saves humans is creativity itself. AI cannot reason like we do, it lacks emotional intelligence, and it does not possess real skill processing abilities.
It follows patterns; we create them.
Let’s cut through the conventional BS for a second and go a little bit deeper.
AI doesn’t get tired.
It doesn’t ask for time off or shy away from tedious tasks like humans do.
It doesn’t complain, slack, or wait for a certain day to be productive.
If AI could talk like Oliver Twist, it’d be asking:
“Please, mate, can I have some more tasks before you call it a day?”
More prompts. More tasks. More work. No complaints.
The Truth About Chatbots (And Why You Can Build One Too)
Anyone can build a chatbot. Seriously. You just need to understand the foundational steps.
At the core, chatbots are digital customer service reps—always on, hyper-efficient, and insanely reliable.
What we built and trained isn’t your typical, lifeless chatbot staring blankly from the corner of a website. This is a game-changer—if built right with yours.
At Dwell, we decided to stop relying on third-party bots with limitations and instead built our AI-powered assistant that supports users more efficiently.
We update ours frequently with fresh, relevant information so it can genuinely serve our users with valuable calls to action.
Why This Matters
Honestly, we, as humans fear AI because it’s damn near limitless. Others say it's pretty much techy inclined to easily roll along with.
AI doesn’t just replace tasks—it scales them. Faster. Cheaper. Better.
And while AI can do, it can’t create on its own—that’s where humans still win (for now). Creativity is our superpower.
Humans get bored. We need rest.
Our veins run with blood, not code.
AI-bloodless. Tireless. Loyal.
It never sleeps nor slumbers. It never stops or complains.
And if your prompt is right, it works 24/7—fetching APIs (like we chatted earlier), offering solutions, and helping your business run while you're chilling in Bali or on cloud 9 (Deep sleep).
How It All Started...
We were utilising Tidio for live chat support. It worked—until I hit the message cap (100 per month or so).
After that, it was like radio silence on what was happening on my site. That’s when I decided: Enough is enough. I’m building my own AI chatbot.
While our chat agent is still in its preliminary stage, the future is promising.
Also, I don’t like paying for things I can build myself if I’m resourceful enough.And resourcefulness? That’s the heart of human creativity. I love it.
How to Build Your Chatbot
Building your chatbot is simple when you have a project you're working on. Without a project, it becomes frustrating, because your mind lacks a clear goal to focus on or improve—nothing is pushing you to find smarter ways to bring your ideas to life.
I single-handedly prompted my way to an easier life using DeepSeek and other reliable tools at my fingertips—some paid, others on free trials, or free for life and so on.
I built everything using a clean, SEO-optimised embeddable widget with no backend. I chose not to stress over backend development, focusing instead on the MVP until one of my projects grows enough to require a more robust database setup.
Step 1: Understand the basics.
Let’s break it down:
AI (Artificial Intelligence) – Machines mimicking human-like decisions.
ML (Machine Learning) – AI that learns from data.
LLMs (Large Language Models {Brain 🧠}) – Think ChatGPT, Claude, Gemini.
AI Agent (A digital employee that can understand instructions and take necessary actions to actualise a given task).
Limitations? Oh, it has a few:
It’s not truly intelligent—it just predicts patterns or pattern recognition.
It can hallucinate (aka, make up stuff).
It reflects the biases in its training data.
🔗 Resources:
https://ai.google
IBM’s AI Fundamentals
Step 2: Learn How It Works (High-Level Overview)
Data → Training → Model → Output → Feedback Loop
Types of AI:
- Generative AI – Creates text, images, and code via promptings (instructions).
- Predictive AI – Forecasts trends, makes recommendations.
- Computer Vision - This is the field of AI that enables machines to "see" and interpret visual information from the world, like images or video. It helps systems detect objects, recognise faces, track motion, and more.
- Memory in an AI context refers to the system’s ability to retain and recall information over time. This could be 1. Long-term memory, such for storing user preferences or past experiences for future reference. 2. Short-term memory, like remembering what was said a few seconds ago in a conversation.
How they work together:
For example, in a self-driving car or the example we cited earlier:
- Computer vision helps the car recognise stop signs and pedestrians.
- Memory helps the car remember the layout of a familiar neighbourhood or a past detour route.
Therefore, while computer vision and memory are not the same, they can complement each other to create smarter, more adaptive systems.
💡 Try no-code AI tools:
Step 3: Master Everyday Tools
Writing: ChatGPT, Claude, Gemini
Images: MidJourney, DALL·E
Coding: GitHub Copilot, CodeLlama
Voice: ElevenLabs, Google Notebook LM
🔑 Prompting Tip: Use the "5w Rule"
Who, What, When, Where, why (+How)
Example:
“Act as a marketing expert. Give me 5 Instagram post ideas for a vegan bakery in Melbourne, targeting millennials.”
The better your prompting, the better the results and the time saved.
Step 4: Recognise the Ethical Risks
AI is a bastard, and it is important you know this afore time.
Before you dive headfirst, know the risks:
Bias – AI can unintentionally discriminate.
Privacy – It can collect or infer personal data.
Deepfakes – Misinformation is easier than ever.
Step 5: Stay Updated (Without Overload)
🔍 Follow these 3 categories:
News: The Algorithm – MIT Tech Review
Research: arXiv.org
Tools: Futurepedia – AI Tools Directory
That's it for this mini starter kit.
To get better at any of these tools, build a real-world solution.
Thank you for reading this letter.
Please share this with your friends and family. It might help ease some of the tension and fear surrounding modern tools.
I was honestly terrified at first, but I went into it anyway, and it has been worth every step.
I'll see you in the next one.
Until then
– Destiny
{Melbourne, Victoria, Australia}
Unsubscribe · Preferences
