Yokeru

While digital technologies are emerging as a global solution to constrained social care budgets, many are still excluded because they don't have access to a digital device. Consequently, many suffer in silence and mental and physical illnesses go untreated until they become acute, causing a knock-on cost for the care systems that treat them. This, in turn, further depletes already strained resources. 

Yokeru has created an AI-enabled automated contact service that can identify those who are struggling the most. Currently supporting over 6,500 vulnerable individuals in London, Yokeru has so far identified and prioritised over 1,000 interventions. By enabling essential resources to be distributed efficiently at a fraction of the cost of current alternatives, Yokeru ensures that no one is left behind.

The digitisation of health services has moved many critical services online. Many of the most vulnerable do not have access to the internet or a smartphone and, consequently, are excluded from digital support. 

Indeed, in the UK alone, there are 12 million elderly individuals. Of these, fewer than 40% own a smartphone. The individual is required to actively seek help when their health deteriorates, causing many people to slip through the net of existing care systems. 

Due to the pandemic, the most vulnerable are required to self-isolate until a vaccine is available; making identifying who is struggling a challenge for public health systems. Traditional call centres are currently used to engage with vulnerable individuals; however, this requires a vast resource - something local authorities do not have. It is also time-consuming: calling 6,500 people (i.e. the over-70 population of a town of 36,000 inhabitants) once for 5 minutes takes an equivalent 67 working days.  

Furthermore, this is not a challenge unique to the UK (which has the highest rate of smartphone adoption in the world). In Ethiopia, the level of exclusion is higher, with 89% of people (90million people) without a smartphone.

Yokeru's automated call centre is a human-like engagement tool for collecting critical information from large groups of people. Yokeru provides the same outcomes as traditional call centres at one-fifth of the cost and is adaptable to many applications. 

Yokeru's service has four major components.

1. Outbound calls: Automated, AI-enabled call system which can call large numbers of people concurrently

2. Human-like questions: Leading AI synthesised voice to ask questions on the call 

3. Collect and interpret responses: Natural language understanding to intelligently record and interpret the responses 

4. Present actionable data: Collected responses are presented on a dynamic and secure portal which enables care entities to support those who are struggling within the community

The cloud-based platform has been built on Oracle cloud architecture and integrates a call fielding service with AI-enabled state of the art speech synthesis technology. 

Yokeru works with the customer to define a responsive call script. Calls are then run at the frequency determined by the customer. Service users (call recipients) receive informative calls that collect data on behalf of the customer.

Yokeru improves the safety and well-being of vulnerable individuals who are isolated and technologically excluded, including both those affected by the COVID-19 and those for whom isolation, and digital exclusion, is typical. This includes older people living alone, people with significant physical disabilities and people in very isolated areas, abusive relationships and precarious living situations.

The solution uses a human-like automated phone calling system to contact vulnerable individuals and collect information systematically. The data collection process is automated using market-leading tools to allow the system to identify individuals in distress. Yokeru works with users of its technology to suggest features, test the technology and give us constant feedback so we can develop technology in line with the actual needs of end-users. This is a continual process.

Yokeru is deployed to monitor 6,500 people on behalf of a London borough. The system has identified over 400 people in need of assistance in a month. There is a clear need for this type of service in the community, and Yokeru has proven technology that can provide it.

Coping with pandemics requires up-to-date and live data on population well being. Yokeru’s scalable solution offers rapid data collection, allowing for regular updates on the status of the community to governments and other support organisations. Yokeru also collects data without exposing key workers to danger (viral contamination).

Yokeru presents the health of a community as well as specific help requests from individuals suffering from symptoms, unable to get food or medication, or requiring other assistance. Yokeru has already helped to monitor 6,500 isolating people and flag over 400 people that needed help.

Monitoring of vulnerable individuals in their own home is traditionally done by a combination of automated sensors monitoring and of call centres. These services are costly to deploy and to operate and are not suitable for mass-monitoring. Support lines do exist, e.g. the 111 line in the UK, however, they are only inbound lines and don’t operate proactive calls. 

During the pandemic, these lines were saturated, and many individuals delayed their calls by fear of overloading the system. This caused unnecessary suffering. 

Yokeru’s solution is innovative because it delivers the same outcomes as a traditional call centre faster and at a fraction of the cost. The novel technology solves a real-world problem- that is one of efficient mass-engagement. The platform is built to be digitally accessible, being multilingual and not requiring a smartphone application to be used by service users. Importantly, for worldwide expansion, the system is also accessible to those who cannot read or write. 

Yokeru is also designed to scale rapidly, making it uniquely capable of contacting large groups of people. It is also fast to deploy (typically 24 hours) and has no maximum call volume. Finally, using state of the art synthetic voice technology, Yokeru offers a human-like service that is unmatched. 

The output of the platform is aggregated for automated triage and insights generation. This is critical because many are excluded from accessing support and public health information.

Yokeru’s use of phone systems is the data gathering aspect of the system, but our key technology is the automation of day-to-day operations of the system. 

Our system revolves around a database that acts as the data storage, with several processes operating on the database to gather, maintain and report the data we collect. AI models can easily be used to estimate a person’s risk level and change the frequency of calls automatically. 

A key focus of the system is the lack of maintenance required by a client. All they need to do is upload the phone numbers to us, and Yokeru automatically delivers actionable data to a location for them to access insights at the same time every day. New systems such as audio processing can be integrated with the existing system in a modular way such that we can continuously add features as requested, including resources for under-resourced languages to help improve accessibility to any language. Yokeru is based on a combination of intelligent automation, data science and new cloud technologies to generate actionable insights to clients that will save resources and improve outcomes for the community it serves.

Our clients send us a list of the questions they wish to ask users. A human-like speech synthesiser then speaks these questions, and we collect their response. The answers are analysed in real time. The results of the survey are presented on an intelligent dashboard and deposited to a storage bucket.

Our solution has been deployed in May 2020 in the London Borough of Hammersmith & Fulham. Over 6,500 vulnerable households are being monitored by an automated call asking specific questions, such as:

• Whether they had symptoms of coronavirus

• If they needed assistance with a safeguarding issue

• If they wanted to be referred to a befriending scheme

• If they had adequate access to food and medication

• Whether they had any other issues, they needed help with from the council

After processing the incoming data, the solution was able to identify 1,00 help requests from individuals requiring support from the local authority, including individuals with: 

• Safeguarding issues

• Struggles with isolation and loneliness

• Concerns about access to food

• Concerns about access to medication

We know that around 40% of people answer the phone the first time. Combining this with redialing to catch people who didn’t pick up the phone, we can get an overview of the community as a whole as well as specific people who need help.

Yokeru observes that care and aid distribution were not benefitting from modern-day digital technological advances. Since the public sector and NGO’s have (infamously) constrained budgets, they turn to digital technology to bridge the communication gap with residents and citizens. 

Digital technologies and smartphone applications have enabled better communication with many demographics, however vulnerable, elderly, and economically more impoverished populations remain excluded because of lower levels of smartphone adoption.

Organisations need a tool that allows them to improve the quality (and distribution) of the support they deliver without increasing cost. 

We believe that proactive monitoring using the phone line is the answer, and Yokeru offers paradigm change in mass-engagement with the digitally excluded.

For the first time, the vulnerable population can be monitored at nation-scale using Yokeru’s technology. Yokeru not only enables local authorities and care services to distribute resources efficiently but also provides insights about what resources need to be made available in the future.  Regular monitoring enables the care/aid providers to identify and predict changes in household and community patterns, informing improved distribution. 

The outcome, for our target population, is improved understanding of excluded people to enable the improved distribution of resources (care/food/support). This method allows for isolated populations (vulnerable/elderly/3rd world communities) to be contacted and monitored to inform the actions of care/aid organisations. For these organisations, the outcome is an improved distribution of resources and cost savings. 

For social care services, the outcome is a better use of available resources, with the potential of savings on the long term.

Our solution currently serves 6,500 people in the London Borough of Hammersmith & Fulham. We are planning to deploy to 20 local authorities, which will bring us to serve about 80,000 people in the UK within the next 12 months.

We have recently started a collaboration with the World Food Programme, which will lead us to serve over 100,000 people in Ethiopia within the next 12 months.

We expect our solution to support about 10 million vulnerable individuals worldwide within 5 years. Its multilingual application enables us to add more languages to the solution, starting with Amharic because of the existing relationship with the World Food Programme. Our development plan involves deployment in the USA, where we already have a team, in other parts of Europe that have comparable health systems to the UK, and in Africa, where there is great health inequality.

1 year:

• 20 local authorities (UK) + Ethiopia

• Developing new technology to care for vulnerable people

  • With our team in communication with research partners

• Actively supporting communities in countries where we can help

  • Deploying our systems and liaising with clients to make sure the system is effective and their needs are met

• Working on expansion across most English speaking countries

  • Reaching compliance with HIPAA and other non-GDPR standards

  • Sales outreach through direct contact and press releases

5 years:

• Extensive penetration into countries in all continents, covering both widely spoken languages and smaller dialects in 3rd world countries, supporting elderly care, healthcare, humanitarian aid, charities and support organisations

  • Increasing links with the World Food Program, UK and International authorities

  • Making links with age related charities, domestic abuse charities, mental health charities

  • Development of specialised speech recognition software

  • PR and Legal push to make sure we’re ready and everyone knows it

• Deployment of Yokeru and its supporting products as part of the Yokeru Fleet as a complete care system for vulnerable people living alone

  • Extensive R&D as part of multiple research projects with our university partners etc

  • Piloting with organisations to validate the system, verify it can deliver actionable insight

The barriers that exist are mainly financial, cultural and technical.

Financial barriers include making sure we have enough funding to complete our research. Creating a profitable startup is a process that can take years so fundraising will be an important aspect of making sure Yokeru continues to innovate and change lives.

Culturally we will face barriers when it comes to language. Speech recognition is especially important when dealing with populations with high illiteracy rates. Many speech recognition software systems in English work reliably, but when it comes to other languages such as Amharic we find that they are very under resourced, language corpora are often small and specific to one use case in many languages.

Technically we have to overcome limitations in the phone networks. Phone audio is set to a lower quality than VOIP or other recorded audio standards, making it harder to get good quality audio classification work done. In Africa phone signals are more temperamental so we need to be robust to interruptions in service. We also need to be able to generate speech recognition models given low resources in certain languages.

Legal barriers do exist when it comes to data protection, but they can be mitigated by adhering to the right standards and making sure that all data is safe and used responsibly.

Competition in this space does exist, if another company providing similar services to us does grow rapidly then there is a barrier when it comes to finding clients.

To increase financial viability we will work to raise funding from several sources. Research projects will ideally be subsidised by grants agreeing to pay a fixed sum or proportion of the costs of research. We will also be aiming to partially self-fund via revenue from clients. The main way of raising funds will be a combination of grant applications, taking loans through startup schemes and selling equity in exchange for capital and investor expertise. 

To navigate the lack of language resources we are partnering with Imperial College London and the World Food Program to create new speech models that are accurate for the work we will be carrying out globally. This will not only benefit the communities we deploy in but will also serve to increase the general level of resources available to help organisations looking to use speech recognition in these languages.

In order to help deal with problems with the phone lines we will use data augmentation to incorporate robustness into our models.

We have good data protection consultants in the form of Pridatect, and they have helped us reach the standards necessary to operate in the UK and EU. They will also help us in our global expansion.

Market competition is less of an issue as we have a clear route to market with several clients, including the world food program. Penetration in the market is low so there is large room for growth.

At this point in time, Yokeru employs 3 full-time members of staff and 5 part-time colleagues, as well as a varying number of contractors depending on the company’s needs.

Our project lead, Monty Alexander, is the CEO and co-founder of Yokeru.  He studied Mechanical Engineering at Imperial College and previous roles have included  a Research, Design and Development (RDD) Engineer at Dyson for and researcher at Medtronic. His key skills are leadership, project management, entrepreneurship, mechanical Engineering, manufacturing, budgeting, computer programming, data science, and customer engagement.

Our technical lead is Ash Oldershaw, CTO and co-founder of Yokeru and formerly worked in robotics software engineering at Dyson. His key skills are software engineering, cloud development, data science, research, project management, and artificial intelligence. He studied Computer Science with Artificial Intelligence at the University of Leeds

Our operations lead is the co-founder and COO of Yokeru, Hector Alexander, who previously worked for TopHat for 3 years as a senior development manager and for Knight Frank as a surveyor for 2 years. His key skills are business development, project management, project coordination, problem solving, budgeting, systems analysis, and financial resource management.

Yokeru currently has the following group of partners: 

• Oracle - provided grant funding and engineering resources to help develop this solution. 

• Imperial College London - Department of Primary Care and Public Health - evaluating the impacts of the technology within the London local authority 

• World Food Programme & Imperial Business School - development of a project titled “AI solutions to fighting Malnutrition during Covid-19 and beyond” to support 100,000 Ethiopians and transform the method in which the impact of distributed aid is evaluated

Our primary target customers are public bodies such as local government and national health services. The service users of the product (the call recipient) are digitally excluded individuals who are typically vulnerable individuals, including the elderly (over 65) and disabled. Additional service users may include their friends/family. 

Customers want to use Yokeru’s platform because they need more cost effective and cheaper ways of communicating (or engaging) with large groups of people. 

When deployed in the context of a pandemic, Yokeru enables mass-engagement and improves digital oversight of vulnerable individuals. Consequently, care costs can be reduced while increasing access to, and quality of, care. 

Ultimately, positive quality of life outcomes are expected for our service users, and significant reduction in costs of social care for our customers. A study into the quality of life outcomes from using the technology is being undertaken by Imperial College, London, and a paper will be published demonstrating the results. 

Our solution is :

• Scalable and simple to install. 

• Reliable: provides comprehensive oversight.

• Simple to use: presents actionable insights keeping the burden of data interpretation from the ASC provider or family.

• A long-term solution.

Customers pay for Yokeru’s service, rather than service users. Yokeru is currently operating on a pay-per-minute basis with its existing clients. The price point is significantly lower than the competition (i.e. traditional call centres), and this saving to the customer is made possible by automation. 

Yokeru is currently privately funded, and is bringing in revenue from grants to sustain the business until it is forecast to become economically viable at the end of 2020 (when 6 local authority customers are projected to be secured). 

In the interim, grants and further private funding sources are being sought. 

Once three local authority customers are secured, Yokeru will look to raise investment capital from venture capital sources.

Solve will provide an invaluable opportunity to join the talented cohort of funders, startup businesses and (most importantly) experts that we can work collaboratively with to advance the Yokeru project. The further out technology can spread, the more significant the positive human impact can be.

One of the biggest challenges Yokeru will need to overcome is the delivery of the service outside of the UK. Global expansion presents regulatory (data protection), legal, and linguistic barriers that we would look to Solve partners for assistance and support. One particular challenge is that of obtaining HIPPA compliance in the USA; a regulatory hurdle that is critical to the success of the business there. We hope the connections formed through Solve would help us overcome this.

We believe that digital exclusion is one of the world's most pressing challenges. Yokeru is working to solve this problem. To deliver this vision, Yokeru needs funding and expertise. Third-party financing is required until the company reaches viability (forecast at the end of 2020). The $10,000 funding that Solve offers will be helpful to sustain the project to this point.

To reach the goals mentioned above, Yokeru seeks partnership to maximise its impact.

The MIT Solve network would help our technical team by mentoring and advising technical development and growth strategy, both in increasing scalability of processes as well as developing our new AI systems.

MIT Solve has partners from every stage of maturity, access to talk to these partners means we would be able to share experiences and ideas of how to progress the company, with the common goal that we are all working on projects that benefit the world.

We will also be able to build connections to aid/care providers who could benefit from this technology through outreach as being a partner of MIT Solve.

We would be able to provide a benefit as well, as we will freely share our experiences with other partners, and introduce them to organisations we have been working with as appropriate.

MIT Solve have incredible partners and Yokeru would be incredibly grateful for connections to: 

• Vodafone Americas Foundation. Their affiliation with Vodafone, which provides telecommunication services in more than 27 countries, would be invaluable for Yokeru as our technology is based on the phone line. Their expertise and connection to Vodafone business would enable Yokeru to support communities globally. We would be looking to talk to Meena Palaniappan, the CEO. 

• Flagship Pioneering. The broad engineering and business expertise which they have would, again, be invaluable for the development and wide scale deployment and growth of Yokeru. Their approach to growth and their incredible track record for supporting and accelerating business growth would massively help to accelerate the number of isolated/vulnerable individuals supported by our technology.

Our team has experience in AI through work in industry and academia, partnerships with Oracle, Imperial College London and the UN’s World Food Program give us access to experts in speech recognition, data analysis and other areas of machine learning. We have a market validated solution in pandemic assistance addressing a problem faced by millions.

We would use the prize specifically to develop robust AI models for speech recognition in low-quality audio over the phone. To do this we would have resources to research data augmentation to build robust models for under-resourced languages. We would also use AI to create effective methods of prioritising users depending on the urgency of their situation. We would also use AI to work with creating effective speech to text models for under-resourced languages with small corpora. The prize would also enable us to advance our analysis of the data we gather to better understand how pandemics affect communities globally.