Case studies
Data based engineering and human centered design
Client
UK based books re-seller
Ineffective marketplace pricing strategy and warehouse receiving strategy were causing financial loss. Due to the lack of a proper warehouse management system connected with receiving and pricing strategies, stock management had to be done by the team of stock managers
Solution
Automated pricing and receiving strategies using statistical modeling and machine learning to adjust the book price and decide which book should be accepted into a warehouse. Automated books listing across countries an channels.
Sales increased over 10 times for selected marketplaces
Reduced invetory age of product by 50%
Increased stock quality – 50% less books delisted
Sales increased over 10 times for selected marketplaces
Reduced invetory age of product by 50%
Increased stock quality – 50% less books delisted
Client
Debt recovery law firm
Problem Description
Processing of 200.000+ legal documents from incoming mails on a monthly basis. Applying a standard OCR tool wasn't enough to bring real process automation as all details still had to be transferred to proper systems and OCR errors had to be corrected. Periodically, batches of 500.000 + files had to processed in the shortest possible time.
Solution
Set of text classifiers – recognizing a document type
Named Entity Recognition models – mining the information
Image Processing Pipelines – improving OCR input quality
Quality Assurance Framework - assessing if extracted details hold with business logics
Reduced processing time from months to hours
Cost reduction – 70% per document
Reduced risk of leaking PII data
Client
Debt recovery law firm
Problem Description
Processing of 200.000+ legal documents from incoming mails on a monthly basis. Applying a standard OCR tool wasn't enough to bring real process automation as all details still had to be transferred to proper systems and OCR errors had to be corrected. Periodically, batches of 500.000 + files had to processed in the shortest possible time.
Solution
Set of text classifiers – recognizing a document type
Named Entity Recognition models – mining the information
Image Processing Pipelines – improving OCR input quality
Quality Assurance Framework - assessing if extracted details hold with business logics
Reduced processing time from months to hours
Cost reduction – 70% per document
Reduced risk of leaking PII data
Client
One of London Councils
Problem Description
The resident support team focused on replying to the repetitive questions. Council offered a good quality online self - service via the website however, many residents still prefer to use the messaging platform.
Solution
Chatbot integrated with backend systems of the Council and other services providers.
Automation of repetitive questions
Shortening resident journey time
Gaining new insights
Automation of repetitive questions
Shortening resident journey time
Gaining new insights
Client
Medical Insurance company
Problem Description
Ineffective reporting and price calculation. Data was spread across multiple sources. Multiple KPIs calculated in different tools. This caused a long re-calculation time as creating a new offer took more than 10+ hours of computation and engaged several analysts. Due to the data sensitivity, the process could not be moved to the cloud environment.
Solution
Transferring the process to the one consistent app hosted on dedicated server.
Computing time reduced to real time
Whole process can be managed by single person using simple GUI.
Improved process security
Client
Medical Insurance company
Problem Description
Ineffective reporting and price calculation. Data was spread across multiple sources. Multiple KPIs calculated in different tools. This caused a long re-calculation time as creating a new offer took more than 10+ hours of computation and engaged several analysts. Due to the data sensitivity, the process could not be moved to the cloud environment.
Solution
Transferring the process to the one consistent app hosted on dedicated server.
Computing time reduced to real time
Whole process can be managed by single person using simple GUI.
Improved process security
Client
Parcel delivery company
Problem Description
Customer service used to spend their time answering 5 – 10 repetitive questions raised by users.
Most of the users were interested just in parcel delivery status information, pricing calculations, and ETA estimations.
Solution
Chatbot integrated with backend systems.
Automating almost 80% of all queries.
New channel for lead collection.
Increased visibility of the company
Automating almost 80% of all queries.
New channel for lead collection.
Increased visibility of the company
Client
Books retailer
Problem Description
Searching for books misplaced by customers is time consuming.
Solution
Template matching algorithm.
Reduced time of localizing misplaced books.
Optimizing crew’s tasks
New insights about customers behavior
Reduced time of localizing misplaced books.
Optimizing crew’s tasks
New insights about customers behavior
Client
Books retailer
Problem Description
Searching for books misplaced by customers is time consuming.
Solution
Template matching algorithm.
Reduced time of localizing misplaced books.
Optimizing crew’s tasks
New insights about customers behavior
Reduced time of localizing misplaced books.
Optimizing crew’s tasks
New insights about customers behavior
Client
FMG retailer/wholesaler
Problem Description
Bad design of on-premise desktop app caused multiple issues with uptime, scaling and little usability.
Solution
Redesigning the architecture, UI/UX and business context of the app.
Developing cloud-based platform accessible for all the clients via browser + PWA (Progressive Web Apps)
Launching a recommender system
Increased sales by 30%
Mobile app (pwa) yearly revenue – 40 M PLN
Data insights for every store owner
Increased sales by 30%
Mobile app (pwa) yearly revenue – 40 M PLN
Data insights for every store owner
Client
Polish energy supplier
Problem Description
Poor UX of the current solution.
Low conversion rate.
High load of service staff forced to contact and explain incorrectly completed applications
Negligible percentage (less than 1%) of sales with additional services (with the highest margin)
Solution
A completely new remodeled shopping path consistent with the actually diagnosed intentions of users with support from the recommender system. Launch of forms in the mobile version.
Upsale of energy planes boosted from 1% to 50%
Decreased error rates in contract forms by 28%
Technobusiness award : Leader in the Industry 4.0 category
Client
Polish energy supplier
Problem Description
Poor UX of the current solution.
Low conversion rate.
High load of service staff forced to contact and explain incorrectly completed applications
Negligible percentage (less than 1%) of sales with additional services (with the highest margin)
Solution
A completely new remodeled shopping path consistent with the actually diagnosed intentions of users with support from the recommender system. Launch of forms in the mobile version.
Upsale of energy planes boosted from 1% to 50%
Decreased error rates in contract forms by 28%
Technobusiness award : Leader in the Industry 4.0 category
Client
International construction materials manufacturer
Problem Description
Lacking proper products presence.
Missing good content and branding.
Solution
Creating new Key Visuals
Developing product-related landing pages
Digital campaigns at : Google ads,
Social media and Youtube
Automating leads collection
SEO optimization