DDVB Case Study Generator - n8n Workflow

AI-powered Telegram bot for generating publication-ready case studies in Russian for DDVB branding agency.

Overview

This n8n workflow acts as Ilya Morozov, Senior PR Executive at DDVB, creating professional case studies via Telegram. Case studies are generated in English first, then professionally translated to Russian following strict editorial standards for Russian media outlets. All long-form prompts live in standalone markdown files that the workflow fetches dynamically from GitHub, so prompt updates no longer require editing or re-importing the workflow JSON.

Features

• Telegram Bot Interface: Interact via Telegram messenger
• Russian Input: Send requests in Russian via Telegram (natural for Russian PR team)
• AI-Powered Research: Uses Perplexity API for structured strategic research and company analysis
• English Generation: Internally generates case study in English first with comprehensive reference examples (better AI quality)
• Professional Translation: Translates to Russian with detailed editorial standards and proper formatting
• AI Humanization: Post-translation humanization pass to make text sound natural and human-written
• Russian Media Standards: Follows strict СИТУАЦИЯ-ЗАДАЧА-РЕШЕНИЕ structure with character limits
• Telegram Delivery: Receives completed Russian case study directly in Telegram

Workflow Architecture

Telegram Message (Russian input from user)
    ↓
Validate Russian Input & Parse Entities
    ↓
Research (Perplexity) - Structured strategic analysis
    ↓
Generate Case Study in English (OpenAI GPT-4o)
    ↓
Translate to Russian (OpenAI GPT-4o with detailed editorial standards)
    ↓
Humanize Russian Text (OpenAI GPT-4o to remove AI patterns)
    ↓
Validate Russian Quality - Check standards compliance
    ↓
Send to Telegram - Deliver humanized Russian case study

Workflow Components

1. Telegram Trigger

• Listens for messages from Telegram bot
• Webhook-based activation

2. Parse Telegram Input

• Extracts message text and chat ID
• Parses case study requirements
• Identifies missing information

3. Research Phase (Perplexity API)

• Translates Russian request to English
• Researches company background, industry context, and competitive landscape
• Provides structured research output with:
  • Company background and market position
  • Brand and market context
  • Industry insights and trends
  • Potential project drivers and challenges

4. English Case Study Generation (OpenAI GPT-4o)

• Generates case study in English following comprehensive reference examples
• Follows SITUATION-TASK-SOLUTION structure
• Creates mandatory client and agency quotes (500-700 chars each)
• Integrates DDVB branding naturally throughout
• Produces metadata in English (SEO tags, keywords, pull quotes)

5. Translation to Russian (OpenAI GPT-4o)

• Professional translation with detailed Russian media editorial standards
• Applies proper Russian formatting («кавычки», em-dashes, number spacing)
• Uses correct branding terminology (брендинг, айдентика, нейминг, etc.)
• Maintains СИТУАЦИЯ-ЗАДАЧА-РЕШЕНИЕ structure
• Eliminates канцелярит (bureaucratic language)

6. AI Humanization (OpenAI GPT-4o)

• Post-translation pass to remove AI patterns
• Makes text sound natural and human-written
• Varies sentence structures and lengths
• Removes robotic phrasing while preserving all facts
• Maintains Russian media standards and DDVB branding

7. Quality Validation

• Validates character limits (title ≤90, text 1500-2000)
• Checks СИТУАЦИЯ-ЗАДАЧА-РЕШЕНИЕ structure
• Verifies Russian formatting («кавычки», em-dashes, number spacing)
• Confirms DDVB branding integration

8. Telegram Delivery

• Sends humanized Russian case study to user's Telegram chat
• Includes validation results
• Formatted with Markdown

Russian Media Editorial Standards

All case studies comply with:

• Title: ≤90 characters, active voice, no preposition/number at start
• Subtitle: Expands title (1-2 sentences, no duplication)
• Main Text: 1,500-2,000 characters
• Structure: СИТУАЦИЯ-ЗАДАЧА-РЕШЕНИЕ (Situation-Task-Solution)
• Quotes: Both client AND agency (500-700 chars each)
• Total Length: ≤3,000 characters
• Formatting: Russian quotation marks «», em-dashes (—), numbers "10 000"
• Language: Clear, active Russian eliminating канцелярит (bureaucratic language)

Installation

Prerequisites

1. n8n Cloud account or self-hosted n8n instance
2. Telegram Bot Token (from @BotFather)
3. API credentials:
   • Perplexity API key
   • OpenAI API key (GPT-4o access)
4. Public Git hosting for prompts (GitHub raw URLs are used in this repo and can be swapped for your own)

Setup

1. Create Telegram Bot:

   • Message @BotFather on Telegram
   • Use /newbot command
   • Follow instructions to get bot token

2. Import Workflow:

   # In n8n, go to Workflows → Import from File
   # Select: workflow/ddvb-case-study-generator.json

3. Configure Prompt Source (optional but recommended):

   • In the Assemble Prompts code node, update GITHUB_BASE_URL to your repository if you fork the prompts
   • Switch PROMPT_VERSION to load prompts from different branches/environments (e.g., main, staging)
   • Ensure your n8n instance can reach the raw GitHub URLs (or replace them with S3/GitLab/etc.)

4. Configure Credentials:

   • Add Telegram Bot credentials (bot token)
   • Add Perplexity API credentials
   • Add OpenAI API credentials (ensure GPT-4o model access)

5. Activate Workflow:

   • Enable the workflow in n8n
   • Telegram bot will start listening for messages

Usage

Starting a Conversation

Send a message to your Telegram bot in Russian:

Пример:

Создай кейс DDVB для Sostav.ru о ребрендинге крафтовой пивоварни "Хмель & Солод".
Продажи выросли на 45% после запуска новой айдентики от DDVB.

Подробный пример:

Нужен кейс для Forbes Russia о ребрендинге финтех-стартапа "FinTechPro".
После ребрендинга от DDVB компания привлекла инвестиции на $5 млн.
Проект включал разработку новой айдентики, нейминга и позиционирования.

What to Include

Provide as much detail as possible in Russian:

• Client name and industry
• Target publication (Sostav.ru, Forbes Russia, RBC, VC.ru, etc.)
• Project type (branding, rebranding, packaging, naming, identity)
• Business results (metrics, growth percentages, timeframes)
• Client quotes (if available)
• DDVB team members involved
• Deliverables created

Expected Output

You'll receive a Telegram message containing:

1. Complete Russian Case Study:

   • Professional title and subtitle
   • СИТУАЦИЯ-ЗАДАЧА-РЕШЕНИЕ narrative structure
   • Client and agency quotes
   • Team composition section

2. Validation Status:

   • Quality check results
   • Standards compliance confirmation

File Structure

case-study-generator/
├── README.md                            # This file
├── QUICKSTART.md                        # Fast setup checklist for n8n + Telegram
├── AGENTS.md                            # Helper agent instructions for collaborators
├── CLAUDE.md                            # Claude-specific collaboration notes
├── docs/
│   └── customization.md                 # Deep dive on extending the workflow
├── prompts/                             # All long-form prompts consumed at runtime
│   ├── editorial-standards.md
│   ├── openai-generation.md
│   ├── perplexity-research.md
│   ├── reference-examples.md
│   ├── russian-humanization.md
│   ├── russian-translation.md
│   ├── system-prompt.md
│   └── prompt-config.json
└── workflow/
    └── ddvb-case-study-generator.json   # Main n8n workflow export

Note: All prompts are fetched at runtime from the raw URLs configured in the four “Fetch … Prompt” nodes and merged inside Assemble Prompts. Editing any markdown file and pushing to Git automatically updates the live workflow—no JSON re-imports required.

How It Works

Russian Input with English-First Generation

1. User sends request in Russian via Telegram (natural for Russian PR team)
2. Perplexity translates & researches - structured strategic analysis in English
   • Company background and market position
   • Brand context and target audience
   • Industry trends and competitive landscape
3. Generate case study in English using OpenAI GPT-4o
   • Accepts Russian input but generates in English
   • Better AI quality and structure in English
   • Better handling of quotes and nuance
   • Follows comprehensive reference examples
4. Translate to Russian using specialized translation prompt
   • Professional translation with detailed editorial standards
   • Proper Russian formatting and terminology
   • СИТУАЦИЯ-ЗАДАЧА-РЕШЕНИЕ structure maintained
   • Eliminates канцелярит (bureaucratic language)
5. Humanize Russian text using AI post-processing
   • Removes AI patterns and robotic phrasing
   • Varies sentence structures naturally
   • Makes text sound human-written
   • Preserves all facts and DDVB branding
6. Validate Russian output for quality and standards compliance
7. Deliver via Telegram in Russian

Why English-First Generation with Humanization?

• Better Quality: AI generates more consistent, high-quality content in English
• Natural Input: Users write in Russian (their native language)
• Specialized Translation: Dedicated translation step ensures proper Russian formatting
• Terminology Control: Translation prompt enforces correct брендинг terminology
• Structure Preservation: SITUATION-TASK-SOLUTION translates cleanly to СИТУАЦИЯ-ЗАДАЧА-РЕШЕНИЕ
• AI Humanization: Post-translation pass removes AI patterns and robotic language
• Natural Output: Final text sounds human-written, not AI-generated
• Best of Both: Russian input comfort + English generation quality + human-like output

Telegram Bot Commands

Once your bot is active, users can:

• Send case study requests in Russian (plain text, natural language)
• Receive Russian case studies instantly (generated via English for quality)
• No special commands needed - just describe the project in Russian

Customization

Full, step-by-step instructions live in docs/customization.md. The guide covers:

• System prompt + persona: update DDVB services, methodology, awards, or switch the PR executive name everywhere the persona appears.
• AI behavior tuning: adjust OpenAI/Perplexity temperature, max tokens, or swap models (e.g., gpt-4o → gpt-4-turbo) directly in the relevant nodes.
• Publication targeting: add new publication profiles, detection logic in Language Detection & Parsing, and pass the selection into the prompt.
• Validation rules: extend the Validate Case Study node with extra checks, scoring, and automated revision triggers.
• Integrations & automation: push approved case studies to WordPress, Slack, Airtable, Google Drive, or custom endpoints after validation.
• Advanced features: multi-turn conversations, translation fallbacks, prompt A/B testing, feedback loops, and prompt version tracking.

Use the testing checklist in docs/customization.md after every change (Russian/English inputs, validation pass, formatting, costs, and quality score) before rolling updates into production.

Troubleshooting

Common Issues

Bot not responding:

• Check workflow is Activated (toggle ON)
• Verify Telegram credentials are correct
• Test bot with simple message first

API errors:

• Confirm API keys are correct
• Check API quota and billing status
• Verify GPT-4o access for OpenAI

Translation quality issues:

• Review the contents of prompts/russian-translation.md
• Adjust temperature (lower = more consistent, higher = more creative)
• Verify Russian formatting requirements

Case study structure problems:

• Review prompts/openai-generation.md
• Ensure SITUATION-TASK-SOLUTION structure is clear
• Check that quotes are being generated properly

Support

For issues or questions:

• Check n8n logs for error messages
• Review API quota and limits
• Verify all credentials are properly configured

License

Proprietary - DDVB Agency

Version

• Version: 2.2.0
• Created: November 2024
• Updated: November 2025 (Prompt externalization + workflow cleanup)
• n8n Version: Compatible with n8n 1.0+
• Platform: Telegram Bot + n8n Cloud
• Architecture: English-first generation with Russian translation and AI humanization