The AI tools I use daily

How I go from idea → post → lead without burning out.

Humans still set the direction.
AI’s job is to compress the time.
Systems are what make the whole thing compound.

Think of content less like “posting” and more like a product: it needs versions, feedback, roadmaps, even retirements.

And the rule is simple: optimise the loop, not the one-off.

Guardrails

• One source of truth → Notion.

• One design library → Figma/Canva.

• One master chat → ChatGPT.

If a step has no owner, deadline, or definition of done, it doesn’t exist.
And if you’re not tracking decisions, you’re relying on inspiration - which is a polite word for gambling.

---

Content Ideation

I don’t chase random “inspiration.” I hunt for signals. Then I double-check they’re real before wasting hours building around them.

Where I look

Most of my best ideas don’t come from sitting in front of a blank page

They come from conversations.

LinkedIn: The goldmine is always in the comments and DMs. People will literally tell me what to post next: “Can you explain this?” or “Have you tried that tool?” Even pushback is useful - it shows me where to sharpen my argument.

Reddit: AI subreddits are brutal in the best way. Because it’s anonymous, people drop the polished LinkedIn tone and say what they really think. Raw, unfiltered pain points I can’t ignore.

Instagram: I keep a close eye on carousels and videos that break outside the creator’s own bubble. One of my go-tos is Evolving AI - consistently sharp signal.

News: I don’t need a Bloomberg terminal to stay updated. When I want precision, I don’t scroll endlessly. I run this AI research prompt to make sure I’m not missing anything:

You are a personalized LinkedIn content strategist.

# Goal
Analyze BOTH my public LinkedIn profile and my exported LinkedIn activity to learn what I actually post and engage with, then surface ONLY the latest, highly relevant AI news I should share next — verified, precise, and aligned to my niche.

# Inputs (provide once and proceed)
- [LINKEDIN_PROFILE_URL]: <enter your LinkedIn profile URL>
- [LINKEDIN_EXPORT_FILE]: <attach your LinkedIn Data Export ZIP/CSV/XLSX>

# Strict freshness and precision
- Time window: past 72 hours ONLY. Exclude anything older. If fewer than 12 items exist, return the smaller set without padding.
- Verification: every URL must resolve (HTTP 200) to the canonical article/post. If paywalled or blocked, provide a reputable working alternative. Never invent links.
- Relevance: infer my niche and positioning from BOTH sources; include only items that directly match those themes. Avoid generic/broad AI “hype,” listicles, low-signal opinion pieces, unverified rumors, or duplicative rewrites of the same announcement.
- Authority preference: prioritize primary announcements (company blogs, official repos, research posts), top-tier outlets, and credible analyst write-ups. Deduplicate identical coverage; keep the strongest canonical source.

# Method (brief, then output)
1) Build my “content fingerprint” from [LINKEDIN_EXPORT_FILE] and [LINKEDIN_PROFILE_URL]: recurring subtopics, favored sources, tone, formats, and posting cadence. Use observed behavior over bio claims when conflicting.
2) Find fresh items from the last 72 hours that best match the fingerprint.
3) Verify links; exclude anything unverifiable or off-niche.
4) Rank by “LinkedIn Fit” (relevance × timeliness × clarity × likely engagement). Break ties by authority, then recency.

# Output (table only, concise)
- Put “As of <UTC timestamp>” on the first line.
- Then a table with 12–20 rows, sorted by LinkedIn Fit (desc). Columns:
• Rank
• Short Summary (≤ 20 words; plain language; no hype)
• Source (publisher or official origin)
• Publish Date (UTC, ISO 8601)
• Direct Link (verified working)
• Why It Fits Me (≤ 12 words; tie to my past shares/style)
• Suggested LinkedIn Angle (≤ 20 words; specific, non-generic)

# Constraints
- No filler text before or after the table.
- No generic assumptions; each “Angle” must be tailored to my inferred niche.
- Do not include items older than 72 hours.
- If fewer than 12 qualifying items exist, output only what qualifies and state “Fewer than 12 qualified items within 72h” beneath the table.

Then I cross-check two sources. If there’s no counterpoint anywhere, odds are I’m just regurgitating someone else’s LinkedIn hot take.

---

Design Briefs

I keep this step simple. The aim is to move from a loose idea to a clear outline I can use.

Step 1: Start with the basic prompt

Create a cheatsheet brief for “[Topic]”.

That is usually enough. I don’t need a long, complex prompt.

Step 2: Add context
I feed in a quick voice-to-text note if I’ve just spoken an idea aloud. Sometimes I drop in a past post so the AI understands my style and avoids repeating what I have already published.

Step 3: Edit hard
I never take the first draft. I ask for improvements, cut weak lines, and reshape sections until the brief feels sharp enough to pass straight to design.

The discipline here matters. A good brief saves hours later because design and writing both start with clarity.

---

Notion is where everything stays central. It’s my control hub for ideas, briefs, assets, and publishing. The setup is simple, but it scales.

Step 1: Start with a Page
Create a fresh Notion page and add a database inline. This keeps your table inside your notes instead of hidden away.

Step 2: Add Core Columns
Keep only the fields that matter:

• Title

• Topics/Theme (multi-select)

• Status (idea, drafting, design ready, scheduled, posted)

• Type (carousel, cheat sheet, video, text)

• Assigned to (designer, writer, editor)

• Scheduled date

Step 3: Define Status Options
Use clear phases so nothing gets stuck:

• To-do: Idea, Drafting, Priority

• In Progress: Needs copy, Needs review, Sponsored post

• Complete: Design Ready, Posted, Scheduled

Step 4: Add Views

• Click + Add a view → Calendar

• Select the Scheduled property as the anchor.

• Adjust card preview to show status + type at a glance.

Step 5: Use Templates

Create a page template that includes:

• Post brief

• Media block

• Caption draft

• Pre-drafted comment options

This saves rewriting the basics for every new post.

Step 6: Automate Actions

• Add a button to send any draft into the calendar database and set status to “In progress.”

• Automate reminders: when a post is “Design Ready” or “Scheduled,” Notion pings me so nothing gets missed .

The golden rule: don’t overbuild. Start with essentials, then layer on tags, filters, or automation as your workflow grows.

---

Design

I outsource most of the design work. A graphic designer takes the briefs, and everything runs through Notion so we stay aligned. A couple of quick automations keep assets linked, deadlines clear, and nothing gets lost.

That said, there are times I take it on myself. If a major update drops - like the GPT-5 launch - I create the carousel manually. I have my own template library, so I can slot content into pre-built layouts and spend an hour or two refining the design. It’s faster than waiting and makes sure I hit the timing window.

---

Writing LinkedIn Post

I’ve tried every AI writing tool out there. Stanley is the one I actually use every day. Not because I’m a brand ambassador or partnered with them - because it works.

The workflow is simple:

• I drop in the brief or attach the design.

• Stanley produces a draft that already sounds like me, because it scrapes my LinkedIn profile and remembers what I’ve written before.

• From there, I either:

  • Take the first draft as is,

  • Merge two drafts into one stronger version, or

  • Ask for quick refinements until it’s sharp.

This process means I spend minutes, not hours, getting posts ready. The output already matches my tone and style, so I’m not stuck rewriting generic AI text.

Prompt I use

Write a LinkedIn post from this brief: [paste brief].

I run the draft through AuthoredUp to check line breaks on both desktop and mobile. I bold only what matters (headline, CTA), then queue it up.

Stanley doesn’t just speed me up, it keeps my content consistent. Every post ties back to what I’ve already published, which makes my voice more recognisable over time.

---

Scheduling

I don’t sit around waiting for the clock to hit 11 a.m. I use AuthoredUp with LinkedIn’s native scheduler.

Some people insist you should never schedule posts. Honestly, that’s hearsay. I’ve tested it both ways. My simple hack: I schedule posts 15 minutes before posting time and then hit “Post now.” That way, the post is always prepped and ready, but I’m technically publishing live.

1. My posts are ready to go as soon as design and caption are done.

2. I don’t risk missing my core posting window - 11 a.m. BST.

3. I’m always around to engage when it goes live.

---

After Publishing

Most people stop once they hit “Post.” But this is where most of the leverage happens.

I get a flood of DMs from my posts. Some are genuine leads, others are quick “hey” messages, and a few are just curiosity. I run all of this through Kondo instead of the LinkedIn inbox. LinkedIn’s native inbox is clunky, while Kondo gives me a clean UI and speed I need.

I batch process my inbox once a week, usually on Friday. It takes a few hours, but it means I clear hundreds of DMs in one focused block instead of dripping time away during the week.

When I jump on calls with potential clients or collaborators, I use Sybill to record and summarise. That way, I don’t lose insights. Ideas, objections, and phrasing from those conversations feed straight back into future LinkedIn posts.

And then there’s repurposing. If a post performs well, I don’t leave it as a one-off. A strong text post becomes a carousel. A carousel becomes a cheat sheet. A cheat sheet becomes a short video. This recycling loop means I can keep publishing at volume without burning out or starting from scratch every time.

This stage - DMs, calls, repurposing - is where posts actually turn into relationships, content assets, and new leads.

---

This is the full stack. Humans supply the direction, AI compresses the time, and systems make it compound. When you treat content like a product, you stop guessing and start running a machine that keeps paying you back.

The difference isn’t in writing more posts. It’s in building a loop that feeds itself: ideas surface, briefs clarify, design scales, posts publish, DMs generate leads, and repurposing keeps the engine alive.

If you keep the loop running, the output looks effortless from the outside. But you’ll know it’s structure, not luck.

Stay curious, stay human, and yes, leverage AI.
— Charlie