Academic Conference Presentation Deck from Research Paper

Introduction

Converting a research paper into a conference presentation deck is one of those tasks that feels like it should be simple but rarely is. You're juggling multiple formats, styles, and complexity levels; extracting the key insights without losing nuance; and then arranging everything into something that actually works on screen. Most researchers end up doing this manually, copying sections between tools, re-writing explanations, and rebuilding visual layouts multiple times. It's exhausting and prone to inconsistency.

The good news is that four specialised AI tools can handle this entire process with almost zero human intervention. By connecting them through an orchestration platform, you can send in a research paper and receive a polished, structured presentation deck. Each tool does exactly what it was designed for: one extracts and clarifies the research concepts, another turns data into visuals, another structures slides, and another adds the final design layer.

This workflow is worth setting up even if you only run it a handful of times per year. Once configured, it saves hours of tedious work and produces more consistent results than manual assembly.

The Automated Workflow

Overview of the Process

The workflow moves through four distinct stages: paper analysis, concept clarification, visual generation, and slide assembly. Data flows from one tool to the next using a central orchestration platform. Here's the sequence:

1. Upload your research paper to bhava-ai for intelligent content extraction.
2. Send the extracted sections to explainpaper for technical concept simplification.
3. Pass key findings and statistics to text2infographic for visual design.
4. Assemble everything into a structured presentation using preswald-ai.

Choosing Your Orchestration Tool

For this workflow, I recommend n8n over Zapier or Make. Here's why: n8n allows local execution, handles long-running processes better, and doesn't have strict timeout limits on free accounts. Since your research paper might take a few minutes to fully analyse, n8n's flexibility is valuable. Additionally, n8n's node-based editor makes it simpler to see how data moves between tools.

If you're already deep in the Zapier ecosystem or prefer cloud-only solutions, Make (Integromat) works well too. Avoid basic Zapier for this workflow; you'll hit limitations on data transformation.

Setting Up the Workflow in n8n

Start by creating a new workflow in n8n. You'll need API keys for each tool:

• bhava-ai: Request API access from their developer portal.

• explainpaper: Generate an API key from account settings.

• text2infographic: Use their REST API (available on pro plans).

• preswald-ai: API available for enterprise accounts; contact their sales team.

Step 1:

Trigger and Paper Upload

Your workflow begins with a manual trigger or webhook. For simplicity, start with a webhook that accepts a paper upload:

Copy
POST /webhook/paper-upload
Content-Type: application/json

{
  "paper_url": "https://example.com/research-paper.pdf",
  "paper_title": "Advances in Neural Architecture Search",
  "author_name": "Dr. Sarah Chen",
  "conference_name": "NeurIPS 2024"
}

In n8n, add a Webhook node to listen for this POST request. Then add an HTTP Request node to download the PDF:

Copy
Method: GET
URL: {{$json.paper_url}}
Response Format: File

Store the file in a variable for the next step.

Step 2:

Extract Content with bhava-ai

bhava-ai specialises in understanding academic papers. Its API endpoint for document analysis is:

Copy
POST https://api.bhava-ai.com/v1/analyse-paper
Authorization: Bearer YOUR_BHAVA_API_KEY
Content-Type: application/json

{
  "document": "base64_encoded_pdf_content",
  "extraction_type": "academic_paper",
  "sections_needed": [
    "abstract",
    "introduction",
    "methodology",
    "key_findings",
    "conclusions"
  ],
  "output_format": "json"
}

In n8n, add an HTTP Request node configured as follows:

Copy
Method: POST
URL: https://api.bhava-ai.com/v1/analyse-paper
Headers:
  Authorization: Bearer {{$env.BHAVA_API_KEY}}
  Content-Type: application/json

Body (raw):
{
  "document": "{{$json.file_content_base64}}",
  "extraction_type": "academic_paper",
  "sections_needed": ["abstract", "introduction", "methodology", "key_findings", "conclusions"],
  "output_format": "json"
}

The response will be a JSON object containing structured sections:

Copy{
  "abstract": "A concise summary...",
  "introduction": "Background and motivation...",
  "methodology": "The approach used...",
  "key_findings": [
    {"finding": "Result A", "significance": "high"},
    {"finding": "Result B", "significance": "medium"}
  ],
  "conclusions": "Overall insights..."
}

Step 3:

Simplify Technical Concepts with explainpaper

Academic papers use dense, technical language. Explainpaper rewrites sections for clarity without losing accuracy. Use their API to process each section:

Copy
POST https://api.explainpaper.com/v1/simplify
Authorization: Bearer YOUR_EXPLAINPAPER_API_KEY
Content-Type: application/json

{
  "text": "The methodology employs a transformer-based architecture with multi-head attention mechanisms...",
  "target_audience": "general_academic",
  "preserve_technical_terms": true,
  "output_length": "medium"
}

In n8n, add a Loop node to iterate through each extracted section (abstract, introduction, methodology, etc.). For each section, add an HTTP Request node:

Copy
Method: POST
URL: https://api.explainpaper.com/v1/simplify
Headers:
  Authorization: Bearer {{$env.EXPLAINPAPER_API_KEY}}
  Content-Type: application/json

Body (raw):
{
  "text": "{{$json.current_section_text}}",
  "target_audience": "general_academic",
  "preserve_technical_terms": true,
  "output_length": "medium"
}

Store the simplified versions in a new variable. The response structure is straightforward:

Copy{
  "original": "...",
  "simplified": "...",
  "key_concepts": ["concept1", "concept2", "concept3"]
}

Step 4:

Generate Infographics from Data with text2infographic

Key findings, statistics, and comparisons work better as visuals. text2infographic converts text descriptions into structured infographic designs:

Copy
POST https://api.text2infographic.com/v1/generate
Authorization: Bearer YOUR_TEXT2INFOGRAPHIC_API_KEY
Content-Type: application/json

{
  "content": "Our research shows that neural architecture search reduces training time by 45% compared to manual design. Across three datasets, the improvement averaged 42-48%.",
  "style": "academic",
  "format": "vertical_chart",
  "colour_scheme": "professional_blue"
}

In n8n, add another HTTP Request node. Extract the key findings from the bhava-ai output and format them as descriptive text:

Copy
Method: POST
URL: https://api.text2infographic.com/v1/generate
Headers:
  Authorization: Bearer {{$env.TEXT2INFOGRAPHIC_API_KEY}}
  Content-Type: application/json

Body (raw):
{
  "content": "{{$json.key_findings_formatted}}",
  "style": "academic",
  "format": "vertical_chart",
  "colour_scheme": "professional_blue"
}

The response includes a URL to the generated infographic:

Copy{
  "infographic_url": "https://cdn.text2infographic.com/output/abc123.png",
  "format": "png",
  "dimensions": "1920x1080",
  "alt_text": "Comparison of neural architecture search results"
}

Download and store this URL.

Step 5:

Assemble the Presentation with preswald-ai

preswald-ai is designed specifically for building presentation decks from structured content. Its API accepts content blocks and produces a ready-to-present slide deck:

Copy
POST https://api.preswald-ai.com/v1/build-deck
Authorization: Bearer YOUR_PRESWALD_API_KEY
Content-Type: application/json

{
  "deck_title": "Advances in Neural Architecture Search",
  "author": "Dr. Sarah Chen",
  "conference": "NeurIPS 2024",
  "slides": [
    {
      "slide_number": 1,
      "type": "title",
      "title": "Advances in Neural Architecture Search",
      "subtitle": "Dr. Sarah Chen",
      "metadata": {"conference": "NeurIPS 2024"}
    },
    {
      "slide_number": 2,
      "type": "content",
      "title": "Abstract",
      "body_text": "{{simplified_abstract}}",
      "layout": "text_only"
    },
    {
      "slide_number": 3,
      "type": "content",
      "title": "Key Findings",
      "image_url": "{{infographic_url}}",
      "layout": "image_dominant"
    }
  ],
  "theme": "academic",
  "output_format": "pptx"
}

In n8n, assemble the final payload by combining all previous outputs. Use a Function node to construct the slides array:

Copy// Combine extracted content, simplified text, and infographic URLs
const slides = [
  {
    slide_number: 1,
    type: "title",
    title: $json.paper_title,
    subtitle: $json.author_name,
    metadata: { conference: $json.conference_name }
  },
  {
    slide_number: 2,
    type: "content",
    title: "Abstract",
    body_text: $json.simplified_abstract,
    layout: "text_only"
  },
  {
    slide_number: 3,
    type: "content",
    title: "Introduction",
    body_text: $json.simplified_introduction,
    layout: "text_only"
  },
  {
    slide_number: 4,
    type: "content",
    title: "Key Findings",
    image_url: $json.infographic_url,
    layout: "image_dominant"
  },
  {
    slide_number: 5,
    type: "content",
    title: "Conclusions",
    body_text: $json.simplified_conclusions,
    layout: "text_only"
  }
];

return { slides: slides };

Then send this to preswald-ai:

Copy
Method: POST
URL: https://api.preswald-ai.com/v1/build-deck
Headers:
  Authorization: Bearer {{$env.PRESWALD_API_KEY}}
  Content-Type: application/json

Body (raw):
{
  "deck_title": "{{$json.paper_title}}",
  "author": "{{$json.author_name}}",
  "conference": "{{$json.conference_name}}",
  "slides": "{{$json.slides}}",
  "theme": "academic",
  "output_format": "pptx"
}

The response provides a download link:

Copy{
  "deck_id": "deck_xyz789",
  "download_url": "https://api.preswald-ai.com/v1/download/deck_xyz789.pptx",
  "status": "ready",
  "created_at": "2024-01-15T10:32:00Z"
}

Final Step:

Store and Notify

Add one more HTTP Request node to download the final PPTX file, or simply store the download URL in a database or send it via email:

Copy
Method: POST
URL: https://api.sendgrid.com/v3/mail/send
Headers:
  Authorization: Bearer {{$env.SENDGRID_API_KEY}}
  Content-Type: application/json

Body (raw):
{
  "personalizations": [{"to": [{"email": "user@example.com"}]}],
  "from": {"email": "noreply@myservice.com"},
  "subject": "Your presentation deck is ready",
  "html": "Your presentation is ready for download: {{$json.download_url}}"
}

The Manual Alternative

If you want more control over the final deck or prefer not to set up an orchestration platform, you can run these tools sequentially by hand. Upload your paper to bhava-ai, copy the extracted sections to explainpaper, select key statistics and paste them into text2infographic, then manually arrange everything in preswald-ai or PowerPoint. This takes roughly 90 minutes for a typical research paper, compared to 10-15 minutes once the automation runs.

The manual approach also lets you review each stage and make corrections before the next tool processes the content. For a critical presentation, this review step may be worth the extra time.

Pro Tips

Monitor Rate Limits. Each API has request limits. bhava-ai allows 100 requests per minute on their standard plan; explainpaper and text2infographic allow 50 per minute. If you're processing multiple papers in quick succession, add a Delay node between requests to avoid hitting limits. A 2-second delay is safe for all four tools.

Handle Long Papers. If your research paper exceeds 50 pages, bhava-ai may time out during extraction. Split the paper manually into two sections, run them through the workflow separately, and then use preswald-ai to merge the decks. Alternatively, ask bhava-ai to prioritise the first 40 pages and methodology sections.

Customise the Theme. preswald-ai offers several built-in themes. If you're presenting at a specific conference that provides a branded template, contact preswald-ai support to request a custom theme, or manually apply the conference branding to the final PPTX file after download.

Test with a Small Paper First. Before running this on a critical paper, test the entire workflow with a shorter research article (10-15 pages) to validate that all API connections work and that the output quality meets your standards. Adjust settings like output_length in explainpaper or colour_scheme in text2infographic based on the results.

Cost Savings Through Batch Processing. If you're preparing multiple presentation decks, run them in sequence within a single n8n workflow. Most tools offer volume discounts, and batching reduces overhead. A batch of five papers costs roughly 30 per cent less than processing them individually.

Cost Breakdown

Total monthly cost (using all tools actively): £352 (self-hosted n8n) to £392 (cloud n8n).

If you run this workflow just once per month, the cost per presentation is quite reasonable. If you're generating decks weekly, the cost per output drops significantly. For teams of researchers, splitting the subscription cost across multiple users improves the value proposition.