The hardest part about AI isn't using it - it's knowing what to use it for.
Here's my tip: don't start with AI prompts or tools - start with where your work feels stuck.
That's where AI can help you today.
- Why AI works best when used to solve specific workflow problems
- 10 friction points where AI genuinely improves your work
- Real-world example of using AI from chaos to clarity
You don't need a 200-prompt template pack or another list of generic AI instructions. What you need is clarity on where your work gets stuck.
AI has become a quiet force behind my productivity - not because I use it for everything, but because I use it specifically for the messy parts of my workflow where I traditionally struggle.
I don't use AI to take over my work. I use it as a thinking partner.
Instead of treating AI as a writing tool or question-answering machine, I've built a collection of thinking partners that help me navigate the messy middle of creative and strategic work - the space between raw ideas and finished products.
The 10 Specific Tasks Where AI Actually Helps Me
I organise by friction points, not tools:
Shaping ideas - turning rambling thoughts into structured outlines
Analysing messy input - making sense of transcripts, notes, and research
Refining drafts - improving flow and clarity while keeping my voice
Creating audience profiles - building target reader/customer personas
Checking accuracy - extracting claims for verification
Writing support content - quick blurbs, emails, summaries, proposals
Keeping me on task - managing priorities and follow-through
Making decisions clearer - framing trade-offs and structuring options
Shortening and simplifying - cutting words without losing meaning
Formatting output - prepping content for different platforms
The most valuable use case? Turning mess into clarity.
My Actual Process: From Chaos to Coherence
Let me walk you through a typical writing workflow:
- I start with chaos - unstructured voice notes or rambling thoughts
- AI helps transform this raw material into a rough outline
- I do another round of expansion and thinking
- AI helps organise this new material using my tone guidelines
- I manually edit and finalise everything
This works because the AI isn't writing for me - it's helping me think more clearly while maintaining my voice.
Custom AI Tools I've Built for Specific Problems
I've created several purpose-built AI tools that address specific friction points:
For writing challenges:
- A developmental editor that identifies weak structure
- A text shortener that preserves key points
- A style analyser that maintains my voice
For strategic thinking:
- A target reader builder for audience clarity
- A scenario planner for complex decisions
- A prompt creator for better AI instructions
For administrative burdens:
- A proposal generator based on past work
- A meeting summary writer
- A GDPR toolkit for compliance
One I use regularly: the Fact-Checker that extracts and lists every claim I make (stats, facts, etc) from my writing so I can verify them - sometimes automatically, sometimes manually (try it with one of your blog posts or emails).
It keeps me accountable without adding hours to my process.
A Real Example From Last Week
I was creating a client strategy document and hit several friction points:
- Initial chaos: I had too many disconnected ideas and couldn't see the structure
- Solution: Recorded a 4-minute voice note of my thoughts
- AI role: Transcribed and organised it into a logical outline
- Weak arguments: Some sections felt unconvincing
- Solution: Used my "developmental editor" AI
- AI role: Highlighted three sections needing stronger evidence
- Accuracy concerns: Wasn't sure all my claims were solid
- Solution: Used my fact-checker AI
- AI role: Extracted 17 claims for verification (found 3 needed additional support)
- Tone mismatch: My casual style didn't fit this slightly more formal client
- Solution: Quick tone adjustment
- AI role: Adjusted formality while keeping my core message
The result? A document completed in half the time, with higher accuracy and better structure - and the client specifically commented on how clear and well-reasoned it was.
The Not-So-Secret: Context Creates Quality
Here's the unsurprising part: generic prompts produce generic garbage. But specific, contextual prompts create remarkable results.
Garbage In, Garbage Out. We've always known that.
When I'm working with AI language models in particular, I try to always provide:
- Voice notes containing my actual thoughts (lots of context)
- Examples of my past work (more context)
- Specific rules about my tone and style (I have this in my profile/settings)
- Project background information (lots more context)
- My specific concerns or questions
The more context I give, the more valuable the AI's contribution becomes.
How to Apply This to Your Work Today
Start by mapping your friction points. Where do you get stuck? What's difficult or so?
Is it starting a project or task? Organising? Editing? Formatting? What specifically makes it difficult? Too many ideas? Unclear?
Then look for AI assistance - just ask it how it can help. Feel free to DM me a question on LinkedIn and I'd be happy to recommend you a specific tool or technique that could help.