Skip to content

DarynaAkkus/ai-news-automation

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 

Repository files navigation

AI-Powered News Publication Automation

Methodology Tools Course

Overview

This project designs an end-to-end automated news publication system for a social media editorial team using AI integration and the DMAIC methodology.

Business problem: Manual news monitoring is slow, costly, and error-prone.
Solution: An AI system that monitors competitors, detects viral content,
auto-generates posts, and publishes after editor approval.


Process Logic (TO-BE Model)

Competitor website
       ↓
AI monitors views in real time
       ↓
Views > 1,000 in 5 min?
   NO  → End process
   YES ↓
AI generates: text + title + image
       ↓
Editor review
   ├── Approve → Auto-publish to Facebook & Instagram
   ├── Needs fix → AI corrects → back to review
   └── Off-topic → End process

DMAIC Framework

1. Define

The current AS-IS process requires a journalist to manually monitor competitor sites (Onet.pl, WP.pl, Interia.pl), evaluate article popularity, write content, create visuals, and publish — all manually. This causes delays and missed trends.

Business goal: Reduce reaction time to trending topics, cut operational costs, and scale content production without increasing headcount.

2. Measure

Key performance indicators defined for comparison (before vs. after automation):

KPI Description
Reaction time Time from competitor publish to own post
Daily output Number of posts published per day
Engagement Likes, comments, shares per post
Cost per post Time and resources per publication
Rejection rate % of topics rejected by editor

3. Analyse — 5 Whys

Problem: The publication process is too slow and inefficient.

  1. Why is it slow? → Editor manually monitors competitor sites
  2. Why is monitoring manual? → No automated data collection system exists
  3. Why no filtering system? → Process relied on subjective editorial judgment
  4. Why subjective judgment? → No analytical tools or decision algorithms implemented
  5. Why no tools? → Process was never analysed for optimisation

Root cause: Lack of systematic content selection and over-reliance on manual work.

4. Improve — TO-BE Process Roles

AI System (automation):

  • Monitors competitor websites continuously
  • Detects viral content (threshold: >1,000 views / 5 min)
  • Generates unique text, title, and image
  • Submits to editor for approval
  • Auto-publishes after approval

Editor (quality control):

  • Verifies factual and linguistic accuracy
  • Evaluates title and image appeal
  • Approves, rejects, or requests corrections

IT Specialist (implementation phase only):

  • Configures monitoring system and social media integrations
  • Periodic technical oversight — not required daily

5. Control

Monthly KPI review covering: reaction time, daily post count, engagement rate, cost per post, and rejection rate. System parameters updated as market conditions evolve.

Expected benefits after automation:

Benefit Impact
Faster reaction to trends Higher reach and competitiveness
Reduced manual workload Editor focuses on quality, not repetition
Lower operational cost Less time per post
Scalability Multiple sources monitored simultaneously
Content standardisation Consistent quality based on defined criteria

Process Map (TO-BE)

automatyzacja_newsow_schemat drawio

Files

File Description
Projekt_Akkus.docx Full academic project report (Polish, WSZiB format)
process_map.drawio Editable BPMN-style process diagram

Context

Course: Automatyzacja i Robotyzacja Procesów Biznesowych
Institution: Wyższa Szkoła Zarządzania i Bankowości (WSZiB), Kraków
Year: 2026
Author: Daryna Akkus


Author

Daryna Akkus — Data Analyst | Business Analyst
GitHub · LinkedIn

About

AI-powered news publication automation system — DMAIC process design

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors