Algospark designs and supports projects that span a wide range of opportunities and challenges. We are specialists that design and deliver analytics and artificial intelligence solutions.
Applied AI Consultancy
We have a flexible approach to suit client requirements. An engagement involves realising new revenue opportunities or savings costs. We typically use the following process:
Products and Frameworks
Please get in touch for details about set-up and subscriptions:
We are a consultancy with skills that span data science, customer experience, service design, finance and technology. We partner with specialists to ensure solutions are value driven, fast to implement and iterative. Algospark is a network of associates led by Darren Wilkinson (CFA). He has many years experience leading and managing service innovation and analytics projects across numerous industries.
A sales forecasting tool that combines proprietary and third party data sets with machine learning. Delivers an efficient sales planning and feedback framework that targets 15-30% sales forecast error reduction.
Churn Prediction solutions to identify customers that are most likely to stop trading. The probability of a customer leaving is determined and "Value at Risk" calculated so that customer churn mitigation activities can be implemented. This helps revenue retention (1-3%) and reduce average spend on new customer acquisition (5-10%).
Product Recommender solutions to identify the products that customers are most likely to buy. Using hybrid collaborative filtering, products are prioritised based on the probability of customer purchase. This helps revenue uplift (1-2%) and determine product recommendations as part of churn mitigation activities.
Interactive product portfolio analytics tool that uses market basket analysis and a sales inference engine to determine product opportunities and their impact on the product portfolio. Targets 1-2% sales increase and 1-2% gross margin uplift from portfolio pricing efficiencies.
A product portfolio dashboard to align product investment decisions across multiple projects, businesses and geographies. Targets 25% reduction in analyst time and faster project approval processes.
Applied machine learning tool to optimise opening and closing of available pricing tiers. Targets 1-3% revenue uplift from optimising balance between increasing demand and increasing average sale price.
Machine learning to detect potential fraudulent activities. Optimises statistical detection techniques and balances against the business value of preventing fraud versus the cost of incorrect account deactivation. Targets 20-40% value add from fraud prevention and detection activities.
A machine learning framework to predict opportunities and challenges across business operations. Uses event classification and outcome rules to target increases in process efficiency.
A customer insights framework for understanding what matters, why it matters and what should be done. Targets increased sales and reduction in customer churn.
A verbatim insights quantification tool for capturing, aggregating, summarising and prioritising actions from verbatim feedback. Expected sales conversation rate increase of 5-10% and churn reduction rate of 10-15%. Forms part of the Xuuu Customer Insights Framework.
A customer personality profiling engine for designing customer experiences based on user preferences and personality types. Expected sales conversion rate increase of 5-10%. Forms part of the Xuuu Customer Insights Framework.
Use of artificial intelligence to recommend actions to customer service specialists. Targets 20% increase in customer response productivity.
AI driven image review tool. Used for delivering consistent and efficient review of advertising imagery, blog content, social media posts, ethics compliance, general corporate imagery, etc.
An AI driven image review solution. Clear Vision brings together a suite of computer vision algorithms that 1) read and interpret text, 2) identify and describe objects in pictures and 3) use content style filters to identify content themes. Targets consistent processes and 75% approval process reduction time. Get in touch for a username and password to try the solution!
A simplistic blog impact assessment tool to forecast the impact of blog posts on web traffic. Built as a foundation for web strategy and analytics projects that link site scrapes and Google Analytics.
Predicting Premier League football game outcomes. Uses an ensemble of AI models to predict a home win, draw or away win for upcoming Premier League games. See if your favourite team is likely to win, and if the bookmakers are offering favourable odds.
Generating amazing images using neural style transfer. Built for exploring new images and artwork using deep learning techniques to combine content images and style images.
A word prediction tool trained on tweets, blogs and news articles. Built as a foundation for Natural Language Processing (NLP) projects such as document classification and automated user interaction responses.
Location Spark is a decision support tool for new retail site investment. Given a postcode and size of store, it predicts sales, store characteristics and trading patterns for locations throughout the UK. Location Spark presents new site forecasts and key location metrics by bringing together numerous data sources and applied artificial intelligence. Insights are delivered in a simple mobile dashboard that is tailored to client requirements and can be up and running within a few weeks. Location Spark is a proven and powerful tool that helps make better and faster new site selection decisions.Launch Demo
Cluster analysis tool to determine store clusters and new store sales patterns using information from across an existing network of stores. It is also used for sales target setting and determining relative performance.
Map Explorer is an interactive visualisation tool. It has been designed for use on large screens and for presenting snapshots to compliment wider location analytics projects.
Deck GL hexagon view of eateries across the UK. Used as a framework to complement wider location analytics projects.
Crime Predictor pre-processes all reported crimes over 3 years in the UK to predict where and when crimes are expected to take place. This tool feeds investment decision frameworks to improve outcomes regarding new properties and new business locations.
Crime Explorer is a large screen interactive visualisation tool to explore UK crime patterns. It has been designed to present snapshot views to compliment wider location analytics projects.
Postcode Finder is a tool to find and explore the distribution and proximity of UK postcodes. It is used to find postcodes from approximate locations and is a useful companion application for geo-location intelligence tools.
Location dashboards help screen and evaluate potential new sales locations. Developed as early prototypes of Location Spark, location dashboards help fast track new site decisions. They reduce analysis time (20-40%) and improve location decision outcomes.
A classification and feedback tool to streamline adult social care needs assessments. Ensures consistent needs classification, ongoing wellbeing feedback and predictions of future needs. Targets 40% process time savings, and cost savings from reducing future hospital visits.
A prototype care supplier selection algorithm for care commissioners. Ensures suppliers are prioritized and selected faster to reduce care brokerage requirements. Expected 25% brokerage processing time savings.
A prototype Digital EHCP for children's social care specialists. Developed by Social Kemistri to deliver 25% process efficiencies. Brings together NLP algos, clustering and outcomes prediction.
An exploratory data tool covering children's special education needs in England using SEN2 data 2014-2019.
A scheduling and team resource optimisation tool for private tutoring, home care delivery and other time slot based services delivered to homes. Targets 10% cost efficiency savings.