Episode 003 · Recorded September 23, 2026

Start Your Day With the Brief, Not the Forum

AI can do the morning forum read. What to answer, in whose voice, and who should answer it stays with you.

One idea

Every community manager's day has started the same way: open the forum and read what happened overnight. The Daily Brief hands that read to AI, which works like an intern who reads every thread, ranks what needs you, spots the patterns, and drafts a first reply. What stays human is the deciding: which reply goes out, in whose voice, and which person in the company should answer. What changed is access: the brief now runs from a chat plan and a browser, with no IT ticket, API or MCP server required.

Show notes

Overview

Most community managers start the day the same way: coffee, the forum, and an hour of reading to find out what happened overnight. That read is where the job’s judgment begins. It shows who is stuck, which thread is turning, and what the community is quietly asking for again. It is also the part of the job that scales worst. In a community with hundreds of open threads, nobody reads everything, so the morning read becomes a guess about what to skip. AI changes the arithmetic, and the newer point is that you don’t need anyone’s permission to try it. A chat plan and a browser are enough.

Chris Heuer and his co-host Marius Ciortea spend this episode on one working artifact: the Daily Brief, a skill Marius runs on his own community every morning and has rewritten so any community manager can use it. They walk through its six sections, then run it live on two communities. Marius uses a beekeeping forum he belongs to. Chris uses a community he is a member of but doesn’t manage. The episode ends with the extension Marius likes most, which stops answering questions and starts routing them to the right person. The skill is free to download for Claude and ChatGPT.

The morning read, handed off

The old habit is the right one. Marius says so plainly: the first thing a community manager should do each morning is open the forum, read the threads, and get a feel for the night. The brief doesn’t replace that instinct. It does the reading first, so the human starts from a summary instead of a scroll. Its last two sections boil the day down to a plan, and Marius’s rule of thumb is that twenty minutes spent on what the brief puts at the top clears about seventy percent of what the day actually needs.

Start with the intern

The most useful design choice in the brief comes before any prompt is written. Marius didn’t start by asking what AI could do. He asked what he would want from an intern who came in early every day to help him, wrote that list down, and narrowed it to six questions. Framing the work as a human request keeps the prompts natural, and natural requests are what these models handle best.

The brief also needs less plumbing than most AI tools. It doesn’t depend on an API, an MCP server or anyone in IT. Current AI assistants can use a web browser as their eyes. If your community can be read in a browser, the assistant can read it, and if your platform’s admin screens open in a browser, it can work there too. Chris installed it by pasting Marius’s skill into Claude and asking it to set the skill up. When Claude’s own cloud browser was blocked from the community he wanted to read, he co-browsed instead and ran the brief through his local browser, where he was already signed in.

Chris calls the person who works this way a proveloper, borrowing from prosumer cameras: not a developer, but someone thinking like a solution architect and building their own tools. The first run is the hardest step. After that, the brief is something you adjust rather than something you build.

Six questions, every morning

Each section of the brief is its own prompt, and the skill is the collection. Together they follow the pipeline from the previous episode: aggregate, understand, then act.

Top signal is the one thing you need to know today. Is anything on fire, is a question sitting unanswered, is a thread heating up or turning sour?

What’s changed compares today with yesterday: a new post drawing a crowd, a knowledge base article going up, a mood shifting. It improves with use, because it builds on earlier briefs. Run it daily and it can tell you that the thread you answered yesterday has turned positive.

Needs attention ranks open threads, and the ranking is only as good as its definitions. Marius worked out what counts as urgent, medium and low with the model itself. The model can then tell a member whose production system is down from a member asking which sticker to put on a laptop. Both are good conversations; only one needs you before lunch.

Recurring patterns is where the brief starts to understand rather than report. It shows the same question arriving in different words from different people.

Coach’s note comes from Marius’s years in team sports. A coach tells you what to focus on in the time you have. If you have twenty minutes, these are the things to knock out. Marius often reads this section first, then scrolls back up only when he wants to know why something made the list.

Suggested actions turns the plan into drafts. If the brief can see your documentation, it can tell you a question is already answered in the install guide and write the reply that points there. You edit it and post it, or, with deeper access, send it from the tool.

Every one of these questions can be rewritten. The version you can download is a starting point, and the most useful brief is the one tuned to the questions only your community asks.

What the brief saw in a beehive

Marius ran the brief through ChatGPT on the Flow Hive community, a beekeeping forum he recommends to beginners, minutes before recording. Even on a first run, with no history to compare against, the output is useful. The top signal is a question that has gone eleven hours without a reply. The needs-attention list is all medium and low, so nothing is urgent today. One thread involves a member weighing whether their homeowners’ association would ever notice a hive. The recurring pattern is a steady run of questions about honey texture and extraction, which tells whoever runs that forum what their next blog post should be.

The suggested reply shows why a draft beats a blank page. Rather than answering the unanswered question itself, it invites experienced keepers to share what they check before a first harvest and to include their region and season so newer beekeepers can put that experience in context. Nobody prompted the word context. The brief reached for it because the community needed it.

A draft is not a decision

Chris draws the line here. Everything the brief produces is a suggestion, and it will be right often enough that accepting it becomes a habit. He has caught himself doing exactly that while coding with AI, clicking yes again and again until the judgment quietly belongs to the tool. An intern can be wrong, and so can this one. Keep your agency, and use the brief to see what you couldn’t see before, not to stop looking.

The blank page is where the value sits. Starting from a draft is far easier than starting from nothing, and for some people that alone gives back hours a week. Readers are also getting better at spotting AI-written replies, so the edit matters. Rewrite the draft in your own voice, or, as the brief grows, have it call a voice skill of your own. You can take it much further, all the way to short video answers. Marius’s point is simpler: the goal is an easier morning, and following the suggested actions consistently is what makes a community more active.

Anyone can read the room

The brief doesn’t care whose community it reads. If you have MCP access and a connected CRM, it can add who the member is, what company they’re with and how large the account is. Without any of that, it still reads whatever you can see in a browser, including communities that sit behind a login.

That is how Chris ran it on the Skoolers community, where Skool’s own community owners compare notes. He belongs to it but doesn’t manage it. Nothing there looked like a fire, but money questions were piling up without an official answer. That cuts both ways. It is competitive intelligence for anyone curious about another community, and it is a reminder that your members, and your competitors, can now read your community this closely too.

From answers to experts

Marius’s favorite extension changes what the brief is for. Community managers are the link between members and the knowledge that lives in people across the company, and many questions are beyond what any one community manager knows. So he had the AI read a year of employee answers in his community and build a directory of who knows what. Now the suggested actions name the best person to answer each question and offer to ask them. When he says yes, an agent sends that person a Slack message with the original question, and the expert clicks through and answers in the community. Sometimes the right person is the CEO, who would never have found the thread alone.

Answers from the person who actually knows carry more weight than documentation, and the routing still runs through a human. The brief suggests the expert; the community manager decides who gets asked.

The same move works upward. If an executive wants to engage, the brief can watch for the questions they’d want to answer. Time the brief gives back becomes time to show value to product, sales and leadership, as long as the members stay at the center of it.

What’s ahead

Download the Daily Brief, free, with a version for Claude, a version for ChatGPT, and step-by-step instructions to install each and run your first brief. Marius treats it like an open-source project: fork it, make it your own, and send back what you build. The show will feature the changes worth sharing, so the next version of the brief can come from the people running it.

Chapters

Time Chapter
00:00 The morning read
00:58 Welcome to episode three
01:51 A daily brief anyone can use
03:17 No IT ticket: the browser as the AI’s eyes
04:56 Thinking like a builder: installing it in Claude
06:59 Start with the intern, not the AI
08:28 Top signal
09:19 What’s changed
11:06 Memory makes it sharper
11:32 Needs attention
12:04 Defining urgent, medium and low
13:11 Recurring patterns
13:48 Coach’s note
14:59 Suggested actions
16:22 Point it at your documentation
17:36 Our sponsor: Higher Logic
18:23 Running it on a beekeeping community
19:41 What the brief found
22:15 Keep your agency
23:05 A draft reply, ready to edit
24:57 The blank page and your own voice
25:55 Make your mornings easier
27:17 More context, and reading other communities
29:12 Routing questions to the right expert
31:51 When the CEO answers
32:49 Fork it and share it back
33:19 Where to get the skill

Key takeaways

  • The morning read can be handed off. The brief reads the forum first, so the community manager starts the day from a summary instead of a scroll.
  • You don’t need IT to start. If your community opens in a browser, an AI assistant can read it, with no API, MCP server or IT ticket required.
  • Design it like an intern, not like AI. Asking what you’d want an intern to tell you every morning keeps the prompts natural, and natural requests are what these models handle best.
  • A skill is a collection of prompts. Each of the brief’s six sections is its own prompt, so you can rewrite any one of them for the questions only your community asks.
  • Urgency needs definitions. Working out with the model what counts as urgent, medium and low lets it separate a member whose production system is down from one asking about laptop stickers.
  • The brief improves with history. Run it every day and “What’s changed” compares against yesterday, so it can tell you a thread you answered has turned positive.
  • The coach’s note is the shortcut. Twenty minutes on what it puts at the top clears about seventy percent of what the day needs, and you can scroll up to see why.
  • A draft is not a decision. The brief is right often enough that accepting it becomes a habit, so keep your agency and rewrite drafts in your own voice before they go out.
  • Anyone can read any room. Co-browsing reaches communities behind a login, which makes the brief a competitive-intelligence tool and a reminder that others can read your community just as closely.
  • Route questions to people, not just documents. A directory of who in the company knows what lets the brief suggest the right expert, and the community manager still decides who gets asked.

CommunityAssistant.ai Links

Co-host & sponsor

In this episode

About the show

CommunityAssistant.ai is the podcast and open resource library exploring the human work of community management: what AI can take off your plate, what only you can do, and how to grow into the role that’s coming. Hosted by Chris Heuer and Marius Ciortea. Season sponsor: Higher Logic.

Our sponsor. Community Assistant is made possible by Higher Logic, supporting the practitioners who build the communities that matter.

Our independence. Sponsors make this show possible. They don’t make this show. Marius works at Higher Logic; he appears here as a practitioner speaking for himself, not for his employer.

Made with Claude, not by it · CC BY-NC 4.0