Co-founder Meha Nelson tells us about Predina’s aspirations for Data Pitch and beyond.


Describe your Data Pitch challenge idea


More than a quarter of the 1.25 million deaths on the road are attributable to work-related driving, the direct cost of which is over $500 billion per year. It is a particularly big problem for heavy goods commercial fleets. Most companies, from Amazon to Walmart rely on road haulage. Major accidents are often catastrophic with regards to the number of lives lost, damage to the brand reputation and the direct costs of which can exceed £1m per major accident. There is currently no solution that helps these companies to predict and mitigate the risk of road accidents to their employees and contractors.

At Predina Tech we are developing an AI platform that analyses large quantities of commercially-sourced proprietary and open source data to predict and mitigate the risk of road accidents for any driver, in any place, at any time. Think of our system as Google Maps for safety. Instead of just providing the fastest route to get from A to B, Predina gets you there using the safest route possible in a complex urban environment. Our easily scalable AI software uses historical accident data, drivers’ behavioural and contextual data to accurately predict the risk, severity and the potential cause of an accident, and help mitigate these risks.


What does the idea set out to achieve?


Aligning itself with European Union objectives for addressing Societal Challenges of Health, Demographic Change and Wellbeing, the goal of this innovation business project is to empower logistics professionals to analyse complex systems, enabling long-term efficiency gains through the use of the Predina system, based on the world’s first proprietary road accident prediction algorithm. This will be achieved by establishing a robust approach for the logistics sector to implement efficiencies based on quantitative data and tailored analysis, by establishing Predina as the gold standard for accident prediction process analysis in the EU and by enabling benefits to be adopted by existing logistics providers and new market entrants based on affordability, scalability and availability.

There are four specific objectives.

  1. The first objective is to develop and improve the software components of the Predina platform based on outputs of the feasibility study by May 2018.
  2. The second objective is to perform testing and verification in real-time in Germany and France by June 2018.
  3. The third objective is to finalise the Predina platform’s localisation and readiness for the EU market launch by July 2018.
  4. The fourth objective is to develop commercialisation and sales strategies that support the deployment of the Predina platform, to encourage market adoption and promote further investment by the fourth quarter of 2018.


What makes your idea different or unique?


The Predina system is in a unique position. It has sourced historical and real-time proprietary data from our client, a multinational chemicals company, as well as successfully finished a dedicated UK pilot for their fleet. We have also used our novel machine learning platform to process commercial data with data from public sources to successfully combine 28 contextual variables to accurately predict the likelihood, severity and cause of a road accident in real time. Finally, we used geospatial extreme value modelling, collaborating with the Alan Turing Institute.

Our expert team built the self-learning and adjusting algorithm system to be able to predict accident risk accurately. It is valuable because by having the full picture on what increases and decreases the risk of an accident, our model is able to become more accurate and refined with access to more data over time. By gaining a complete understanding of the drivers’ journeys, fleet management practices are able to improve processes, unlocking a wide range of benefits: increased safety on the roads, enhanced driver experience, increased travel and cost efficiency, positive environmental impact and reduced administrative burdens.


Where did the idea come from?


Bola is an automotive domain expert and I am a risk analytics, machine learning domain expert. We were pondering how vehicles interact with their environment to quantify risk. We were looking at applications around the autonomous vehicle space and how machine learning can be used for it. We did a small pilot with a road construction company to predict the risk of onsite accidents using the same premise.

During this time we met a board member of our current client in a networking event. She showed strong interest and introduced us to a new application for our solution (i.e. to predict road accidents for heavy goods vehicles dynamically). They put all the procedures and stakeholders in place to provide us with full support in building the application, including interviewing drivers, fleet managers, and the scheduling, safety and transport teams. Bola even spent an entire day in the truck with one of their drivers, in order to understand the gravity of the problem at ground level. They paid £30k for the pilot and the rest is history.


What excites you about the challenge you applied for?


Britain alone, loses £30 billion (which is 2.3% of its GDP) annually in direct and indirect cost of road crashes, most of which falls on busy motorways and main roads. Over the last decade 30,000 people have been killed, and a further 300,000 people seriously injured in crashes. In Europe, road crashes remain the leading cause of death among young adults. Improving road safety by implementing and standardising new AI software infrastructure that can accurately prevent road accidents is our main motivation. We believe in using AI for a better tomorrow!


How did your team meet?


We met through a mutual friend in the Entrepreneur First programme (cohort 7).


What’s the best thing about working with data?


Data lets you get into the finer details of the problem and come up with useful insights (which a human being wouldn’t have been able to decipher) through numerical computations. This is critical in solving these problems.  Accidents are considered to be an inevitable part of our lives but we believe they are avoidable. In a nutshell, we are saving lives by working with data.


Anything else you want to tell us about your startup and why you do what you do?


Predina is still at the pre-revenue development stage, and there is continuous interest in the technology from consumers, early adopters, investors, our first client and the government. We have conducted a successful UK pilot and received very encouraging feedback from the partner client, as well as seed investment and a number of prestigious awards.

United by a common goal, governments, companies of all sizes and individual citizens are working hard to increase road security. Our system delivers what nobody on the market has yet managed: clear and accurate road accident prediction at a fraction of the cost of an insurance premium. We now need to rev the engines and take it to the next developmental level.

Bola’s girlfriend was involved in a road accident with a truck so this project is close to his heart. I have a background in risk and AI, and I want to use my knowledge to have a positive impact on the world. Since transport is an industry that affects people from every walk of life, I feel that I can have the biggest impact, and not just in the UK but in the world. At Predina, we believe in using AI to create a better and safer future for all of us. Let’s change road safety forever!