Predina

Being the CEO can come with weighty responsibility. On meeting at a Digital Catapult event, Predina Co-founder and CEO Bola Adegbulu asked the newly appointed CEO of a large transportation company “What’s the worst thing about your job?” “I dread receiving an early morning call to say one of our delivery drivers has been killed in a road accident.” responded the transportation CEO.

Neither could have predicted the eventual outcome of this simple conversation would have life-saving consequences.

When a large transportation company’s new CEO met Predina’s co-founder and CEO at a Digital Catapult event, they had no idea that the conversation they were about to have would lead to a project with the potential to dramatically reduce road accidents for thousands of drivers.

Barriers to market for AI startups

Predina is an innovative startup using AI and machine learning to use contextual data to dynamically predict road transport risk in real time. Its vision is to use AI to dramatically reduce the incidence of road accidents for the transportation industry.

Predina’s solution acts like a very experienced road safety expert. Using machine learning it links 28 contextual variables such as road layout, temperature, real-time weather conditions and satellite images to data on historic road accidents. The knowledge from this model is then made accessible in two ways:

Firstly, fleet managers can see a dashboard that predicts the risk of accidents allowing them to proactively alter routes and journey times to reduce accidents. Secondly, the application is integrated with on-board sat nav systems to give drivers an audio alert when they enter and leave a high risk zone.

As a startup in the machine learning space, developing its product came with some significant barriers to market entry. Predina’s Co-founder and CEO, Bola Adegbulu explains “Machine learning companies have unique challenges that not many startups face. Running machine learning is very expensive and you need to be able to run it quickly as customers expect fast answers.”

Machine learning algorithms have an insatiable appetite for compute power which leaves startups with unavoidable monthly costs in the thousands.

Machine Intelligence Garage removed the barriers to entry for Predina

Predina’s initial contact with Digital Catapult was simply signing up to its newsletter as a way of keeping its ear to the ground for exciting developments in tech. This led to attending events and the fateful conversation with the transportation company’s CEO.

When Digital Catapult launched its Machine Intelligence Garage programme in 2017 with a particular focus on the problem of computation power, it presented a great opportunity for Predina.

As a successful applicant to the programme, Predina got access to valuable development expertise which helped them to configure the architecture in a future-proof and scalable way using Google Cloud Platform (GCP) compute power. Machine Intelligence Garage also provided Predina with $100k GCP vouchers significantly reducing the cost of market entry.

Training an algorithm on a data set using cloud compute power is not a straightforward process. It can involve the configuration and integration of multiple pieces of software.

“Computation power can be a real issue for machine intelligence startups. In addition to the cost, there is a high level of complexity in choosing the right resources and configuring them for your purposes.” Libby Kinsey, Head of AI and ML, Digital Catapult.

Digital Catapult helps seeks out best in class startups for Corporations

Digital Catapult has an enviable network of startups, academics and larger organisations that are developing AI and machine learning solutions. This combined with its technology provider agnosticism, gives Digital Catapult a unique position to help corporations, like this transportation company, to find the right startups to help them tackle the most pressing business challenges.

Through an active annual calendar of over 200 engaging events attended by a total of 10,000 delegates, Digital Catapult has become the place to meet noteworthy startups. By bringing together larger organisations, startups and academia, Digital Catapult is a catalyst for game-changing conversations.

Predina gets additional boost from Data Pitch Accelerator

Data Pitch is a EU-funded open innovation programme bringing together corporate and public-sector organisations that have data with startups and SMEs that work with data.

Last year Data Pitch ran its first competition for SMEs and startups, asking them to propose data-enabled solutions to a range of challenges across sectors including transport. Predina was one of 18 successful startups from across Europe selected to join the accelerator on the strength of its proposals. Data Pitch provided €100,000 in equity-free funding towards engineering and project costs, introductions to investors which helped them raise a further six-figure sum and ongoing mentoring and support, including legal advice on GDPR and how to handle data.

Ryan Goodman, Startup Programme Manager for Data Pitch, says, “Predina’s success illustrates the benefits of collaborative working to enable the development of ideas and solutions that would otherwise have been impossible to get off the ground. By bringing together startups with larger organisations willing to share data, and with assistance from programmes such as Data Pitch and Digital Catapult’s MiGarage programme, startups can realise their potential and create products and services that bring real economic and social value.”
 

Engaging with Digital Catapult helps Predina develop its product and find a key customer to trial its product.

Participating in the Machine Intelligence Garage programme gave Predina access to valuable expertise and computation power. This enabled Predina to train its algorithm on large data sets and accelerate product development to a degree that would have not been possible otherwise.

The connections made at Digital Catapult and Data Pitch led to the initial trial and subsequent rollout of the solution for the transportation company and connections that have led to six-figure investment in Predina’s growing business.

Successful trial accurately predicts the risk of accidents at 70% accuracy

During the initial trial, the Predina solution was able to accurately predict the risk of accidents at 70% accuracy. Using this data to alert drivers to high-risk areas and re-routing them when possible has the potential to saves lives and reduce injuries as well as decreasing the business costs associated with repairing vehicles and missed deliveries.