5 AI tools I use daily for clients (and why)

The answer in 30 seconds

The AI tools I use daily are: ChatGPT for research and planning, Claude for writing in my style, NanoBanana for visual assets, NotebookLM for in-depth research, and Gamma for presentations. The biggest problem? Most entrepreneurs only get about 20% out of what’s possible, because they don’t create projects, don’t add skills, and don’t give the AI enough instruction.

Why AI tools aren’t delivering as much as you hoped

As an entrepreneur with a digital product or course, you’ve probably played around with ChatGPT before. Maybe you even asked it once to write a caption or draft an email. And then you thought: “Hmm, this is okay… but not wow.”

The problem isn’t the tool. The problem is how you use it.

I work with AI tools for clients every day — from writing ads to creating complete content plans. And what do I keep seeing? Entrepreneurs using AI like it’s Google: you type something in, get an answer, and that’s it.

But AI tools aren’t search engines. They’re virtual assistants you need to train — just like you’d train a new team member. And almost no one does that.

Recent research shows that employees who use AI are about 40% more productive on average. But there’s a difference between people who “use” AI and people who truly leverage it. Frequent users save more than 9 hours per week, and some “superusers” even save 20+ hours per week.

The difference? They know how to use AI tools effectively.

1. ChatGPT: my research and planning assistant

Wat het is: De bekendste AI-tool, ontwikkeld door OpenAI. ChatGPT is een conversational AI die vrijwel elke vraag kan beantwoorden en taken kan uitvoeren.

What I use it for: For clients, I mainly use ChatGPT for:

• Research into audiences, trends, and strategy

• Creating content plans and editorial calendars

• Brainstorming campaign ideas

• Structuring information

A concrete example: A client wanted to launch her course but didn’t know which pain points to address in her ads. Instead of spending hours doing forum research, I asked ChatGPT to compile 20 common frustrations her target audience experiences.

Within 5 minutes, I had a list of concrete quotes and pain points, complete with examples. We then used those in her ad copy.

Result: better click-through rate and lower cost per lead.

Why most people get too little out of it: People type vague prompts like “Give me a content plan” and expect something usable. But without context, ChatGPT doesn’t know who it’s writing for, what the tone of voice is, or what the goal is.

What I do: I create a project with all relevant information: target audience, tone of voice, brand values, previous high-performing content. Then I give specific instructions: “Write 10 Instagram posts in the style of [examples], aimed at [audience], about [topic].”

The difference? Huge.

2. Claude: my writing assistant that knows my style

What it is: Claude is an AI model by Anthropic that excels at nuance and context. Where ChatGPT can sometimes sound a bit “AI-ish,” Claude can match a specific writing style more naturally.

What I use it for: Claude is my go-to for:

• Social media captions

• Newsletters

• Ad copy

• Any text that needs to sound like my voice

The big difference vs ChatGPT: I trained Claude with my Style DNA — a document that lays out exactly how I write, which words I use, how long my sentences are, and what my tone of voice is. Because of that, Claude writes copy that genuinely sounds like I wrote it myself.

A concrete example: For a client, I had to write 15 ads for different audiences. Normally, that would take me at least half a day. With Claude? An hour and a half. Not because I “copy-pasted,” but because Claude delivered a first draft that was already 80% right. I only had to tweak the details.

The mistake everyone makes: People don’t give Claude (or any AI) examples or instructions. They assume the AI will automatically know how they write. It won’t.

What actually works: Create a document with 5–10 examples of your best writing. Upload it to Claude. Ask Claude to analyze your style. From that point on, Claude can write in your voice. That saves hours.

3. NanoBanana: visual assets that stand out

What it is: NanoBanana is an AI image generator you can use to create unique visuals for social media, ads, and websites.

What I use it for: Many clients don’t have the budget or time for a designer, but they do want professional-looking visuals. NanoBanana solves that:

• Ad visuals that match the brand identity

• Social media content with a distinct look

• Illustrations and graphics for landing pages

The advantage over stock photos: Stock photos are generic. Everyone uses them. With NanoBanana, I can create something unique in minutes — and perfectly aligned with what the client wants to communicate.

A concrete example: For a client selling wellness courses, I created a series of ad visuals in a consistent style: soft colors, minimalist composition, natural elements. All visuals done within 30 minutes — and they looked like a designer had spent hours on them.

Why people underestimate it: They think AI-generated images look “fake” or “cheap.” But that’s usually because they don’t give clear instructions. If you describe exactly what you want (style, colors, composition, mood), you can get professional results.

Tip: Use a style reference feature. Upload an image you like and ask the AI to work in that style. Game-changer.

4. NotebookLM: research on another level

What it is: NotebookLM is a Google research tool that analyzes your documents, videos, links, and other sources — and turns them into summaries, infographics, and even audio podcasts.

What I use it for: This is my secret weapon for complex projects:

• Analyzing long documents and reports

• Summarizing webinars and videos

• Creating infographics based on lots of information

• Generating audio summaries so you can learn on the go

How it differs from “just Googling”: NotebookLM works only with the sources you upload. It doesn’t make things up. If you upload 10 documents about Facebook Ads, NotebookLM can summarize them, connect the dots between sources, and even generate a podcast-like audio conversation where two AI voices discuss the topic.

A concrete example: A client had taken 5 different email marketing courses but couldn’t remember which tip came from where, or how everything fit together. I uploaded all her course materials into NotebookLM, and within minutes she had a clear summary, an infographic with key points, and an audio overview she could listen to while working out.

The feature nobody knows about: NotebookLM can create infographics based on your sources. Imagine uploading a long article or report and getting a visual summary you can share immediately — perfect for social media or presentations.

5. Gamma: presentations in minutes instead of hours

What it is: Gamma is an AI tool that lets you create professional presentations in minutes. You give the AI a topic and structure, and it builds the entire deck for you.

What I use it for: For clients who present often:

• Pitch decks for investors

• Course materials and lesson slides

• Webinar presentations

• Internal trainings

The difference vs PowerPoint: In PowerPoint, you do everything yourself: choose layouts, write text, design slides. Gamma does about 80% of the work. You provide the content and structure, and Gamma creates a visually strong presentation that you only need to refine.

A concrete example: A client needed a 30-slide webinar presentation about Instagram strategy. Normally, that would take her at least 4 hours. With Gamma, she had a complete deck in 45 minutes — with professional layouts, relevant images, and a clear structure.

Why people don’t use it: They don’t know it exists. Or they assume AI design tools “aren’t there yet.” But Gamma is already good enough for 90% of the presentations entrepreneurs make.

The real problem: AI tools are being used like Google

Here’s what I see with almost every entrepreneur who says, “I tried AI, but it didn’t do much for me”:

• They don’t create projects

• They don’t add skills

• They don’t upload examples or Style DNA

• They write vague prompts

• They don’t give the AI context

It’s like hiring a new employee and expecting them to automatically understand your business, your tone of voice, and what you mean by “make something nice.”

That doesn’t work. And it doesn’t work with AI either.

How to use AI tools effectively

Here’s what I do — and what makes the difference:

1) Create projects

In ChatGPT and Claude, you can create projects with all relevant information: who you are, what you do, who your target audience is, and your tone of voice. Every time you ask something within that project, the AI has the full context.

2) Add skills and knowledge

Upload documents the AI needs to know: your brand identity, previous successful content, your Style DNA, your audience analysis. The more context you provide, the better the output.

3) Write specific prompts

Not: “Write an Instagram caption.”But: “Write an Instagram caption of 120–140 words, in my tone of voice (see Style DNA), about [topic], aimed at [audience], with the goal of [conversion/engagement/awareness]. Use a curiosity-driven opening line and end with a question.”

4) Train the AI with examples

Upload 5–10 examples of your best writing. Ask the AI to analyze your style. From that point on, it can write in your voice.

5) Iterate and refine

AI rarely produces the perfect output in one go. Ask for adjustments: “Make it more personal,” “Add more emotion,” “Make it shorter but keep the punch.” Each iteration improves it.

The numbers don’t lie: AI pays off

Research shows that:

• Writing tasks are completed 59% faster with AI

• Complex tasks like programming can be completed up to 126% faster

• Users save an average of 2.2 hours per week (on a 40-hour workweek)

• Frequent users can reclaim 20+ hours per week

But that only happens if you use AI the right way: as a virtual assistant you train — not as a search engine you type into.

Frequently asked questions

Aren’t AI tools expensive?

ChatGPT and Claude have free versions that can already do a lot. For professionals, a paid plan costs about €20–25 per month. If it saves you 5 hours per week, you earn that back in one hour.

Doesn’t AI-written text sound “AI-ish”?

Only if you don’t give the AI instructions. If you train it with your style and use specific prompts, it can write in your voice.

How do I know which AI tool to use for which task?

Experiment. I use ChatGPT for strategy and research, Claude for writing, NanoBanana for visuals, NotebookLM for analysis, and Gamma for presentations. But your ideal combination may be different.

Can I trust AI with confidential information?

ChatGPT, Claude, and other major tools claim they don’t train on your data if you have a paid account. Always check the privacy policy. For extra sensitive information, you could consider locally run AI tools.

Conclusion: the tool isn’t the problem — the approach is

AI tools are incredibly powerful — but only if you use them as intended: virtual assistants you train and instruct.

Stop using AI like it’s Google. Create projects. Add context. Train the AI with examples. Write specific prompts. That’s when you’ll truly see what AI can do for you. And if you do that? You won’t just save 20 minutes a day — you’ll save hours. You won’t get “okay” output — you’ll get content that sounds like you wrote it yourself. You won’t become a little more efficient — you’ll become radically more productive.

That’s the difference between “trying” AI and really using it.

About Britt from B-Connecting

Britt helps entrepreneurs with digital products and courses set up their marketing and systems in a smart way. With years of experience training and using AI tools for a wide range of clients, she knows exactly how to use AI effectively — without it sounding “AI-ish.”

Want to learn how to use AI tools effectively for your business? Send a message and we’ll explore the options together.

Important: This article is based on my personal experience as a marketing automation specialist and general information about AI tools. Results may vary by person and situation. For specific implementation questions, I recommend experimenting with the tools yourself or working with a specialist.