# 🏆 Hall of Fame: Scenario of the Month **URL:** \ **Category:** News\ **Created:** [October 7, 2026, 8:26pm UTC](https://community.make.com/t/hall-of-fame-scenario-of-the-month/111829 "2026-10-07T20:26:11Z")\ **Posts on this page:** 1\ **Showing post:** 2
**Author:** ![vendy](https://dub1.discourse-cdn.com/flex013/user_avatar/community.make.com/vendy/32/12635_2.png) [@vendy](https://community.make.com/u/vendy)\ **Post date:** [October 7, 2026, 8:26pm UTC](https://community.make.com/t/hall-of-fame-scenario-of-the-month/111829/2 "2026-10-07T20:26:11Z")
## **September 2026 edition** 👀 We have found our winners of the first edition of our Scenario of the Month competition! đŸ„‡ @Prem_Patel - Daily Article Producer with an AI editor đŸ„ˆ@Vavin_B - Athena: AI Health Mentor on Telegram, and @Philippe_LARRUE - Invoice & PO Reconciliation Here are their clever solutions: ### Daily Article Producer with an AI editor Most AI content scenarios share one habit: they publish everything they write. It doesn’t matter on day one. By day 90, though, the blog is hundreds of thin articles all competing with each other in search. This month’s winner takes the opposite approach. Its whole job is to say no. It splits the work across three separate OpenAI steps, each with its own prompt, so nothing ever grades its own homework. First, a planner turns a topic into a brief: what’s safe to claim and what to steer clear of. Then a writer sticks to that brief, working in three tracked store links and an exact closing CTA. Finally, an independent reviewer scores the draft on 8 quality dimensions, with a pass mark of 85. The reviewer catches both kinds of failure. Made-up facts get a fail, but so does an article that sounds confident while saying nothing, like one that explains how to choose without ever choosing. If a draft fails once, it gets revised and reviewed again. If it fails twice, it goes to a human with the reviewer’s notes written into the spreadsheet. It never touches the blog. The results are real. In September, the creator ran it on a US used-book Shopify store. Out of the first five articles, one was published, two were saved as drafts (scoring 91 and 94), and two were rejected. One of those was a summer reading guide that didn’t name a single book. Compared with the store’s 125 existing articles, typical length jumped from 526 words to 1,300+, and articles with no internal links dropped from 94% to zero. It’s also built with care. Every module retries twice and then alerts the owner, the kill switch is a spreadsheet cell the owner already opens, and every run logs its QA scores and token usage so the cost of each article is always visible. The result is a daily blog that can’t spam itself – and the smartest thing it does is know when not to publish. > **[Daily Article Producer - Make.com Automation Scenario](https://us2.make.com/public/shared-scenario/AGbQOipCuZ0/daily-article-producer-with-an-independe)** > > Writes one Shopify blog article a day and refuses to publish weak ones. An independent AI editor grades every draft, pass mark 85: fail once and it is revised and re-reviewed, fail twice and it parks for a human, never the blog. * * * ### AI Health Mentor on Telegram Athena is a full AI health coaching assistant that lives in Telegram, and it’s built entirely in Make. One scenario handles everything, with no external backend or database. A data store holds the user records, Google Gemini writes the personalized coaching replies, OpenAI Whisper and text-to-speech power the voice conversations, and Stripe handles subscription billing. Every incoming message is checked against the data store, so Athena knows who’s talking and where to send them. New users get a disclaimer in English, German, or Russian before onboarding. Approved users chat directly with Gemini, which picks up on intent and switches between four coaching modes: - **Warm Mentor:** the empathetic, supportive default - **Clear Plan:** direct facts and protocol steps only - **Deep Science:** detailed biochemical explanations - **Voice-Ready:** ultra-short replies optimized for speech None of that mode switching relies on hardcoded keyword routing in Make. Gemini makes the call. Deep links can pre-fill a user’s profile, including weight, goals, and fasting protocol. Voice works both ways: Athena downloads a voice message, transcribes it with Whisper, answers in the right mode, and turns the reply back into speech. After a trial period, a paywall router offers two Stripe subscription plans, and a promo-code route can upgrade a user to VIP instantly. It’s a real, revenue-generating consumer product, built end to end in one scenario: multilingual onboarding, mode-aware AI coaching, voice in and out, and monetization. Check it out on YouTube: [https://www.youtube.com/watch?v=b26R62aK-RU](https://www.youtube.com/watch?v=b26R62aK-RU) * * * ### Invoice & PO Reconciliation Matching invoices to purchase orders is a job every finance team knows well. This scenario shows what happens when an AI agent does more than compare fields. It makes a real judgment call. The agent tells apart wording differences that don’t matter from discrepancies that do. When it genuinely can’t be sure, it escalates to a human instead of forcing a yes or no answer. Every decision is logged with its reasoning, so the whole process is auditable. That visibility is what makes it practical for finance work: the team can see what the agent decided and why. > **[AI Invoice & PO Reconciliation - Make.com Automation Scenario](https://eu1.make.com/public/shared-scenario/3ZlhioG2tqf/ai-invoice-po-reconciliation)** > > An AI agent reconciles supplier invoices against their purchase orders. It reads both documents, matches line items even when the wording differs, flags genuine quantity or price discrepancies, and escalates anything it can't confidently verify to a... * * * Congratulations to the winner and the top finalists, who will receive their prize shortly. See you in the next edition! --- _[View the full topic](https://community.make.com/t/hall-of-fame-scenario-of-the-month/111829)._