TL;DR
Every content session used to start the same way: re-explaining what we post about, what tone we use, what’s already been published, and what’s coming next. I replaced that with a structured Claude setup that already knows all of it. Sessions now take about 20 minutes, posts go straight into the schedule, and I don’t manage a single spreadsheet.
The Problem
Keeping a consistent content schedule as a solo founder is harder than it looks. It’s not the writing that slows things down. It’s the overhead around the writing.
Every time I sat down to plan posts, I’d spend the first chunk of time just rebuilding context. What have we posted recently? What topics are we covering this week? What was that angle we liked but pushed to next month? What does our posting schedule actually look like right now?
And when I tried using AI to help, the problem got worse. Every new conversation started blank. I’d explain our strategy, paste in the content backlog, describe our tone, and by the time we got to actually drafting something, half the session was gone.
That’s not a content problem. That’s a context problem.
What I Built
Think of it as a dedicated AI workspace that never forgets.
I set up a Claude Project specifically for IB Solutions content. Claude Projects are a feature in Claude that let you create a persistent working space with background context always loaded. This one holds our full growth strategy, posting schedule, voice and tone guidelines, and content backlog. Every time I open it, the context is already there — no re-explaining, no catch-up.
But holding context is only half of it. The other half is acting on it.
Connected to this workspace is a content database where every post idea, draft, and scheduled post lives. Each entry has a status — Idea, Draft, Scheduled, Published — and a planned date. Think of it like a shared editorial calendar that both I and Claude can read and update.
The connection between Claude and the database is handled by automation workflows that run in the background. When Claude approves a post, schedules it, or flags an idea as not worth developing, those changes happen directly in the database. I never open a spreadsheet or update a row manually.
There’s one more piece that makes the content itself land better: a My Voice reference library. This is a collection of real posts and messages I’ve written over time. Before Claude drafts anything, it reads those samples. That’s why the drafts sound like me rather than like generic AI output — it’s working from examples, not just instructions.
How a Content Session Actually Works
Here’s exactly what a typical session looks like, from opening the app to closing it.
Step 1: Open the IB Solutions Claude Project
I open Claude on my laptop and navigate to the dedicated IB Solutions project. The background context loads automatically — strategy, schedule, voice guidelines, all of it. Claude already knows where we are.
Step 2: Pull up what’s scheduled
I ask: “Show me this week’s scheduled posts and anything that still needs to be drafted.”
Claude calls the content database and returns the full picture: what’s confirmed, what’s still a draft, what dates are empty. This takes seconds and gives me a clear view of what the session needs to cover.
Step 3: Work through the content together
For each post that needs attention, I discuss the angle with Claude. Sometimes I have a specific idea. Sometimes I ask Claude to propose something based on what we’ve published recently and what fits the week’s theme.
Claude checks our published insights and previous posts first, so it doesn’t suggest topics we’ve already covered. It also follows our posting structure: heavier, story-driven posts on Tuesdays and Thursdays, shorter observations on Wednesdays.
Once we settle on an angle, Claude drafts the post. It pulls from the voice samples automatically, so the tone is consistent without me having to say “write this like I would write it” every time.
Step 4: Review and approve
I read the draft. If it needs adjusting, I tell Claude what to change and it revises. When I’m happy with it, I say so.
Claude then pushes the approved post directly into the content database — setting the status to Scheduled, assigning the correct date, and storing the final copy. I don’t open any other tool. Claude handles the database update via the automation workflows running in the background.
Step 5: Capture any new ideas
If something comes up during the session that’s worth saving for a future post, I tell Claude and it adds the idea to the backlog with a status of Idea. It also notes any strategic decisions we made — a topic we’re leaning into, a framing we’re moving away from — so the next session starts with that context already loaded.
The whole thing runs about 20 minutes.
What This Replaces
Before this setup, content planning involved: a Google Doc for the backlog, a separate calendar for scheduling, copy-pasting drafts between tools, and starting every AI conversation by re-explaining who we are and how we write.
Now it’s one conversation. The AI already knows the context, the database updates itself, and the posts are ready to publish without any manual filing.
Where This Pattern Applies for Marketing Teams
The setup I’m describing is built around LinkedIn content, but the underlying pattern works anywhere a team has a recurring content or communications workflow that currently depends on someone holding all the context in their head.
A marketing director running a content team could use the same approach to hold campaign briefs, channel guidelines, and publishing schedules in one persistent AI workspace. Brief a new campaign once, and every subsequent session picks up from there.
A communications manager handling social across multiple channels could maintain separate voice guidelines and content rules for each, with the AI drafting to the right standard automatically depending on which channel is in scope.
A content strategist managing a blog or newsletter could keep the full editorial calendar and topic backlog in the database, letting the AI surface what’s due, flag gaps, and draft to a consistent brief without a standing planning meeting.
The common thread: recurring work that currently requires rebuilding context every time is a good candidate for this kind of setup.
What This Means
AI tools are genuinely useful for writing. But most people use them as a one-off service — ask a question, get an answer, close the tab. The context resets every time.
The difference here is that the context accumulates. Each session builds on the last. The AI gets better at working with you not because the model improves, but because it has more of the right information available from the start.
That’s not a complicated idea. It just requires building the workflow intentionally rather than treating AI as a tool you pick up and put down.
Want an AI system that keeps your content moving?
We build structured AI setups that fit into how you already work — not generic chatbots, but context-aware systems that know your voice, your schedule, and your strategy.