NGO Team Rubicon partners with Microsoft to move from reactive to preventive disaster zone response

Profile picture for user catheverett By Cath Everett November 4, 2020
Summary:
Team Rubicon, a US NGO that operates in disaster zones, is moving from a reactive to a preventative client services model, with the help of AI. CIO/CTO Raj Kamachee explains the how and the why.

Image of a team of Team Rubicon volunteers
(Image sourced via Team Rubicon website)

While the amazing humanitarian work being undertaken by NGOs around the world, particularly in disaster zones, is rightly widely recognized and lauded, most of us fail to consider the huge logistics challenges involved in ensuring that everything works smoothly and efficiently.

Justin Spelhaug, Vice President and Global Head of Tech for Social Impact at Microsoft, explains the challenge:

Responding to a disaster is a huge supply and demand problem. It's not just about getting a person to volunteer, but it's how does that organization leverages data and intelligence to get the ‘right' person on the ground based upon experience, skillset, proximity, availability and a number of other factors? To mobilize that many volunteers a day almost overnight, you have to have the systems and infrastructure to do it.

One NGO faced with just that issue is Team Rubicon, a US-based organization that uses the skills and expertise of military veterans to help rebuild underserved communities around the world hit by natural disasters ranging from hurricanes and wildfires - and lately Covid-19. To help it in its mission, it has been working on a digital transformation roadmap over the last two and a half years, which it is currently about half way through.

This digital transformation has resulted in it introducing a Microsoft Dynamics 365-based enterprise management and volunteer mobilisation system, which runs on the software vendor's Azure cloud platform. As Raj Kamachee, Team Rubicon's chief information and chief technology officer, points out though, the Covid-19 pandemic "threw us off course", not least because:

The model of a traditional event, such as an earthquake or storm, is that they're instantaneous and based in a single location - they don't generally happen at the same time. But our challenge in this instance was how to respond to a number of events all happening simultaneously.

These events at home in the US ranged from wildfires on the West Coast to Hurricanes Laura, Sally, Delta and Zeta as well as the pandemic itself. Such a complex situation meant that Team Rubicon needed to modify its existing centralised system, which was used by staff based at headquarters to deploy volunteers around the country.

The idea was to introduce a more distributed model in which 25 individual territories were given more control over mobilising the support required locally, a shift that was introduced in as little as five weeks. This mobilisation process includes vetting available skills and ensuring volunteers in vulnerable categories are not put in harm's way but are still able to make a useful contribution.

The benefits of adopting a decentralised model have already been considerable, says Kamachee:

Rather than deploy ‘grey shirts' [volunteers] to one location at a time, we can now deploy them simultaneously. So the volume of volunteers is now larger and their impact is higher and deeper than in the past, which is important as we're not just providing a single service - we're doing everything from providing emergency food and medicine to Covid testing sites.

The benefits of achieving "connectedness"

The NGO also partners with other organizations, including food banks, some of which can connect to its system to share data using an application programming interface, while in the case of others, data entry is a more manual process. Digitalisation is currently somewhere between 95% and 98%, but the idea of achieving "connectedness" is a key priority "to enable information to flow", Kamachee explains.

One means of achieving this information flow is by adopting Microsoft's Common Data Model, which will be used to link the organization's finance and delivery systems. The aim here is to track the journey each dollar makes from the time it leaves a donor's hand. Understanding the impact of that dollar on the ground will then make it possible to report back to donors what it has been used for and how it has benefited service users.

But there are other benefits to be gained from enabling "connectedness" too. Kamachee explains:

When all of your different data pipelines become a single conduit, it's better and faster for decision-making. It improves your operational efficiency and security, and means you can tap into data for things like customer and marketing analysis. That was the goal when we set up the system. We're about half way there and success will be when we have consistently flowing data that provides a feedback loop. The end goal is to ensure we can tap into the power of AI to augment our decision-making, whether we're talking about deployment, mobilisation, predicting training requirements, revenue analysis so the finance team can project what the budget needs to be for x number of projects, and asset tracking and logistics, so we know which goods are expiring or not.

To this end, the NGO already has access to a "huge data lake" on Azure, which takes raw structured and unstructured data in real time and helps staff make cost-effective rapid response decisions in emergency situations. Curated data is also siphoned off into a data warehouse, where it informs the body's enterprise resource planning and fundraising systems for organizational decision-making purposes.

As to where Team Rubicon intends to go from here, it has a number of plans in the offing. For example, it is currently adding a new Dynamic 365 module to provide omnichannel customer service and also launched an "Uber-style" emergency food assistance scheme for vulnerable people, although it unfortunately closed on 8 October due to lack of funds. 

In the pipeline for the first quarter of this year though is an AI-based program that assesses volunteers' skills and experience against the requirements needed to deal with a particular incident and contacts the most appropriate individuals to see if they are willing and able to help. Another system, which is also in testing, is an AI-based optical character recognition system that assigns a validity score to volunteers' documentation and certificates and alerts an administrator to any apparent discrepancies.

Over time though, the NGO's aim is to move beyond simply reacting to crises as they happen towards preventative work in order to help to build resilience within vulnerable communities with the help of local authorities and community leaders. This would mean providing post-disaster education, such as rebuilding homes, and creating networks of partners to "provide services we can't", says Kamachee. He explains:

The end goal is to ensure client services aren't single point in time-based but are provided continuously to ensure people's needs are met all the time. For that, we have a huge need for systems interoperability so we can work with our partners' data. It's what we're heading out to do because the more data we harness, the more AI and machine learning can come into the equation. The aim is to remove and automate mundane tasks, while interacting directly with the people we serve through our volunteers.

My take

Team Rubicon demonstrate that with a shrewd vision for the future, smart partnering and the judicious use of technology, NGOs can teach the private sector a trick or two about how to innovate with tight resources.