A content plan for a month in one evening: prompts and a live example

This kind of plan used to eat a whole day: write out topics, check they don't repeat, arrange them by logic, come up with headlines. Now it's one evening and an AI. I run four client projects and a personal channel of 20,000 followers, and I build the content plans for all of them on the same scheme. That's the one I'll show – with every prompt you can copy and paste right now.

Section 01Why bother with a plan if you can just go "on inspiration"

There's a pretty myth: a real expert writes from the heart when the mood strikes, and that's exactly what lands. Sometimes it does strike, and a strong post really comes out. But a channel runs on rhythm, and bursts of inspiration don't replace it. Your audience gets used to you when you show up regularly. Disappear for two weeks – they forget you, reach drops, and only half of them see the next post.

Working "on inspiration" breaks for three reasons. First: inspiration comes irregularly, but posts are needed on a schedule. Second: when you come up with a topic in the moment, you grab the first thing that pops into your head, and that's usually the most obvious one. Third: without a plan you don't see the whole picture, and the content comes out jagged – three posts in a row about the same thing, then an abrupt selling post that looks like a shout in the silence.

A plan solves all of this at once. You sit down once, calm, look at the month from above, and lay out topics so they lead the reader by the hand. Value, story, more value, engagement, a soft sell. The reader doesn't get tired and doesn't feel sold to head-on. And in the morning you don't agonize over the choice – you open the plan, take the day's topic, and put a post together in 15 minutes.

Section 02Five inputs to gather before the AI

The classic beginner mistake is to open a chat and type "come up with a content plan for a psychologist." The AI will hand you twenty topics on the level of "5 signs of burnout" and "what personal boundaries are." Technically those are topics. They won't sell, because every psychologist on earth writes them and your audience has read them a hundred times.

For the plan to come out as yours, not average-of-the-ward, the AI needs context. Gather five things ahead of time – it takes 15 minutes and decides the quality of the whole plan.

Section 03The prompt builder: setting the task for 30 topics

Now the main thing – the prompt that turns five inputs into a working plan. It has five blocks: context, voice, task, structure, response format. Each block plugs one leak where quality usually drains out. Here's the ready template; paste your data into the square brackets.

You are my content strategist with experience in info-marketing.
Help me build a 30-day content plan for a Telegram channel.

1. CONTEXT
I am [niche: a nutritionist for moms on parental leave].
Audience: [women 30-45, back from leave, want to lose weight
without harsh diets; afraid of slipping up, don't believe
in results, ashamed of their bodies].
I sell: [a nutrition program for $200 and a review for $40].

2. VOICE
I write like this (hold this tone across every topic):
[paste 2-3 of your own posts in full].

3. TASK
Come up with 30 topics for 30 days. Each topic is one idea
that hooks my audience and gently leads to the product.
Topics must be specific and non-generic: my own observations,
breakdowns of real situations, stories, numbers. No general advice.

4. STRUCTURE
Split the topics across content pillars:
40% - value (how-tos, breakdowns, instructions),
30% - engagement (stories, takes, questions),
20% - values (views, inspiration, behind the scenes),
10% - selling (offer, cases, reviews).
Alternate the pillars; don't place selling posts back to back.

5. RESPONSE FORMAT
A table: Day | Topic | Pillar | Hook (first line of the post) |
Format (post / reel / carousel / poll).
Write hooks that grab, in plain living language, no corporate-speak.

Let me walk through why each block sits where it does.

Context sets the boundaries. Without it the AI writes for an abstract person and the topics come out about everything at once. With an audience profile it strikes specific pains.

Voice is what sets your plan apart from anyone else's. Two sample posts change the result more than any adjectives like "write in a friendly, expert tone." The AI copies what it sees, not what you describe in words.

Task with the "specific and non-generic" qualifier cuts the filler. Without saying it, half the topics will be along the lines of "the importance of a sleep schedule." A direct request for personal observations and numbers turns the AI toward your experience.

Structure with pillar percentages is the reason the plan will sell, not just entertain. More on the pillars themselves below.

Response format as a table saves you an hour. You see the whole month at once: where the value is, where the sell is, which post fills which day. And a hook in the first line is already half of the future post.

Section 04A live example: the dialog with Claude from start to plan

Let me show it end to end. I'll take a real niche – a nutritionist for moms on parental leave – and walk through the whole dialog. You'll see what the AI answers, where it stumbles, and how to fix it with a single line.

I send the prompt builder with the inputs filled in. Claude thinks for a couple of seconds and returns a table. The first rows look like this:

Already not bad. The hooks are alive, the topics hit recognizable situations. But I see a problem: by day three there are two value posts in a row about food and hacks, and the personal-voice stories run out. I fix it in one line.

The AI rebuilds the table. Now the rhythm is even: value, value, story, engagement, value, values, soft sell. I read on and notice the selling posts sit too early – on days 6 and 9, when the audience isn't warmed up yet. One more correction.

Three short replies – and the plan went from "fine" to working. The whole dialog took about ten minutes. That's the essence of working with an AI: it produces the rough volume in seconds, you fix the meaning and rhythm with short commands. To phrase requests more precisely, see the piece on the five-part prompt structure.

Section 05Splitting topics across content pillars

Content pillars are types of posts by their job. Without them the plan skews: either wall-to-wall value that gets boring, or wall-to-wall stories after which no one can tell what you even sell. Here's the proportion tested on my channels.

Here's why it works this way. Value is the foundation: it proves you know your stuff, and people save it. Engagement lifts reach, because comments and reactions tell the algorithm the post is alive. Values make you a person worth trusting – people buy from people. Selling harvests what the first three built.

Ten percent for selling scares beginners: it feels like you should sell more often or no one will buy. In reality it's the opposite. When nine of ten posts give value and warmth, the tenth selling post reads calmly, as a natural continuation. When every third post pushes something, people get tired of the channel and unsubscribe.

The AI holds this proportion on its own if you set it in the prompt. But check with your eyes: ask it to print a summary of "how many posts of each pillar came out." Sometimes it gets carried away and gives 50% value. One line, "bring it back to the 40/30/20/10 proportion," and everything falls into place.

Section 06A 30-day calendar: laying it out by day

You have the table of topics. Now turn it into a calendar you can actually live by. Three things matter here: frequency, selling days, and slack for the spontaneous.

Frequency. For Telegram a healthy rhythm is 4-5 posts a week. Every day isn't needed – it wears out both you and the reader. Thirty topics for the month land nicely on "every other day plus weekdays." Ask the AI: "lay out the 30 topics across weekdays, keep weekends lighter – Saturday a lifestyle post or a poll, Sunday a break or a repost of the best."

Selling days. Place selling posts when the audience is more active and in the mood to spend. In my experience that's Tuesday to Thursday, first half of the day. Monday people are digging out of their tasks; by Friday evening their minds are already on the weekend. Tell the AI to shift all the selling posts to Tuesday and Wednesday.

Slack. Don't nail down all 30 slots. Leave two or three days empty for the live stuff: a strong review comes in, something happens, news breaks in your topic. A spontaneous post on something hot often outperforms a planned one. A plan is a frame, not a straitjacket.

Lay out the 30 topics across a month's calendar:
- 4-5 posts a week, on weekdays;
- Saturday - a light post (lifestyle, poll),
  Sunday - a break;
- shift all selling posts to Tuesday and Wednesday,
  the first half of the day;
- leave 3 days empty for spontaneous content,
  mark them as "reserve."
Print the final table with dates starting from [start date].

Section 07From topic to finished post: the second prompt

The plan is a map. But you still have to walk it: turn each topic into text. That's the second prompt, which makes a finished post out of a table row in 15 seconds. And again the key is voice, so keep the post samples in the same chat or in a Project.

Take the day 7 topic from the plan:
"[topic + hook from the table]".

Write a Telegram post in my voice (samples above).
Structure:
- a grabbing hook in the first line (you can take it from the plan);
- short paragraphs of 2-3 sentences, the text breathes;
- one specific example, story, or number;
- end with a soft question or a light call to action.

No markdown, no long dashes, address the reader directly.
Length 1,200-1,800 characters.

The AI hands you a finished post. From there, two paths. If it hit – you copy it, tweak it a bit, publish. If it missed – you fix it in one line: "too long, cut it in half," "the first line is weak, give me five hook options," "add a specific story from practice." In two or three iterations the post is ready.

An important note on editing. Even a good AI post is worth passing your eyes over and dropping in something only yours: a client's name, a detail from yesterday, your favorite phrase. It takes a minute and turns "AI text" into "your post, built with AI." The reader feels the difference even without noticing it.

Section 08One topic – five formats

This is where the plan starts to pay off big. Each topic is not one post. It's a post, a reel, a Threads thread, a carousel, and stories. One idea, five touchpoints across different platforms. Thirty topics turn into a hundred pieces of content, all from one evening of planning.

A third prompt unfolds the finished post into other formats:

Take this post and repurpose it, keeping the meaning and my tone:

1. REEL (a 40-second script): a hook in the first 3 seconds,
   3 short points, an ending with a call to action. Written for
   voiceover, the way people talk, not the way they write.

2. THREADS (a 4-part thread): the first part a grabbing
   statement, then one idea unpacked per part.

3. CAROUSEL (6 slides): one idea per slide, slide 1 the
   cover with the hook, the last one a CTA. Give the text of each slide.

The logic is simple. The post lives in Telegram. The reel catches new cold audience on Instagram and Meta. The thread works on Threads, where organic reach is cheap right now. The carousel gets saved and forwarded – it has a long tail. One idea, served in different ways, catches a person wherever they spend their time.

You don't have to drag every topic into all formats. Take a rule: unfold the strongest topics of the month (wow-effect value and gripping stories) into a reel and a carousel, and leave the rest as posts. That way you amplify the hits and don't drown in production. On building carousels on autopilot there's a separate piece on Instagram carousels, and on running Threads with AI – the one on Threads automation.

Section 095 mistakes that leave a plan in a drawer

Section 10If the plan came out weak: quick fixes

Sometimes after the first prompt the plan disappoints: boring topics, limp hooks, all like everyone else's. This is fixable – no need to rewrite the inputs from scratch. Here are five commands that pull a plan out of the swamp.

• Topics are generic. "These topics are too general; everyone in my niche writes them. Give me topics from my own experience, contrarian takes, breakdowns of specific situations, non-obvious observations."
• Hooks don't grab. "Rewrite all the hooks. Each must stop the scroll: a question tied to a pain, an unexpected statement, a number, the start of a story. No 'want to learn.'"
• Everything's about one thing. "The topics are too narrow, it'll get monotonous. Widen it: add topics about money, relationships with the people around them, inner fears, not just [the main topic]."
• Doesn't lead to the product. "Tie the topics to my product. Let the value naturally point to the full solution being in [product]. Without being pushy."
• Doesn't sound like me. "You write too smooth and formal. Here are three more of my posts – pick up the choppy rhythm, the humor, and the plain words. Rewrite the hooks in that tone."

Section 11A content plan for different niches

The scheme works for anyone who sells services, courses, or consulting. Only the inputs and the emphasis change. Here's what a starting flagship topic and the main pillar look like across different niches.

The principle is the same for everyone. Find the audience's main pain, come in through it with a story or value, show the result, lead to the product. The AI tunes topics to the niche if you gave it a precise audience profile. The more specific the inputs, the fewer edits on the way out.

Section 12The whole path in a minute

Here's everything in a short route to keep on hand.

1. Gather the five inputs before the AI: niche (narrow), audience profile (pains and dreams), product, two or three of your own posts, formats. 15 minutes.
2. Send the prompt builder of five blocks: context, voice, task, structure, response format. Get a table of 30 topics.
3. Refine the plan in dialog: mix the pillars, add personal, push the sells to the end, sharpen the hooks. Three or four edits.
4. Check the 40/30/20/10 proportion of pillars and lay the topics out on a calendar: 4-5 posts a week, sells on Tuesday-Wednesday, three days in reserve.
5. Run the topics through the second prompt in a batch – a week of posts in an hour, all queued.
6. Unfold the strong topics with the third prompt into a reel, a thread, and a carousel. Thirty ideas – a hundred pieces of content.

One evening instead of daily agony with an empty editor. The AI takes on the volume and the grunt work; you make the decisions and add the live parts. The channel fills up evenly, the audience gets used to the rhythm, and sales come in soft and steady. And best of all – your head is free for clients and life, not stuck on the eternal "what should I write about tomorrow."