Results of the PI requirements Analysis Phase

At last the results of the requirements analysis for the PI platform are ready! We’d like to thank our various partners, friends and followers whose participation and support was vital for completing our analysis. Click to download the PI Features and User stories.

In total we ran four requirement elicitation workshops to inform the features of the PI platform and document a series of topical user stories to inform the subsequent development phase.

1. In Geneva, Switzerland, on the 9th – 10th of October 2014 with the International Committee of the Red Cross (ICRC), the International Organization for Migration (IOM) and the United Nations High Commissioner for Refugees (UNHCR).
2. In Delft, The Netherlands, on the 29th of October 2014 with our media partner Free Press Unlimited (FPU). The workshop was hosted by the Delft University of Technology (TUD).
3. Via Skype on the 5th of March 2015 with Amnesty International (AI) based out of Washington DC, USA. Conflicting agendas within AI and ongoing crises made it impossible to meet face to face in London as planned.
4. In Monrovia, Liberia, on the 19th – 20th of March 2015 with the Liberian Peacebuilding Office (LPO) and civil society members of the Liberian Early Warning and Early Response (EWER) Working Group.

Features:  We elicited a great number of features for the PI platform with short and repetitive brainstorming processes, both during the workshops and afterwards by using Agile Software Development methods. With the help of the participants the features were regrouped and organized according to the following categories:

• Dialogue for Data Collection
• Identify and Evaluate Sources
• Evaluate Information
• Verify Information
• Feedback to Sources
• Output for Organizations
• Sharing Data
• Privacy and Security
• Software and Design

User stories
Subsequently the participants were asked to define “user stories” or use cases, describing a topical use of some of the features elicited in their respective domains of intervention. These user stories are particularly useful as they allow developers to better understand the contexts in which the technology will be deployed. They also help the PI team prioritize the development of features that are in high demand and create value at an early stage of development. These user stories will inform recurrent exchanges between PI’s developers, as well as organizations’ focal points and testers in the course of the iterative development and testing phase of the PI platform. Participants defined uses of the PI platform to assist the following types of interventions:

• Family Tracing (ICRC)
• Volunteer management (ICRC)
• Camp management (IOM, UNHCR)
• Incident reports (FPU)
• Rights monitoring (AI)
• Peacebuilding (LPO, EWER)

We also discussed a series of user stories that span all fields of intervention:

• Syntax errors
• Unknown location
• Communication breakdown
• Informing about risks

Outcomes
So far the PI project has yielded the following outcomes:

• The initiation of a discussion around the benefits and challenges of the automation of a series of critical information processes across interlocking fields of intervention (humanitarian, human rights, media and peacebuilding);
• Exchanges between organizations who often (co-)operate with or alongside one another about the opportunities and challenges of ICT solutions to manage and share information;
• The reaffirmation that ethical principles and protection standards must be at the core of any ICT solution during its entire project life-cycle starting with the requirements analysis phase, the development phase as well as during and after operational deployments of a mature solution;
• The understanding that ICT solutions are but a means to an end among many other means, and that for all their benefits, human interactions and interventions remain of primordial importance.

Lessons learned
In retrospect and to summarize, the exchanges and conversations we had were extremely valuable. Together, we discussed and reflected on the best ways to automate the collection, evaluation, verification and sharing of information as well as providing access to valuable information and actionable feedback, while guarding against risks for the users. During these exchanges the PI team also learned a lot about the challenges faced by the humanitarian, human rights, media and peacebuilding community.

From the beginning we decided to embrace multiple fields of intervention to test the hypothesis that many of the core features of PI – the automation of a series of critical information processes – could be of use to a series of interlocking organizations. Informed by our interdisciplinary discussions, we believe that the hypothesis holds true provided that we develop a generic platform around these core features which also provide its users with the possibility to customize its use to their domain and specific needs.
Overall, while remaining aware of the many challenges ahead, participants provided positive feedback about the process, the results, and appeared enthusiastic about testing a prototype of the PI platform in the future. Nevertheless, retaining the engagement of all our humanitarian stakeholders during the subsequent development phase proved to be more difficult. While the UNHCR and IOM indicated that they would follow the development phase with interest, at this stage only the ICRC is willing to contribute time and resources during the development phase, together with all of our other partners in the fields of human rights (AI), media (FPU) and peacebuilding (LPO).

Next Steps
Based on a more detailed version of the results, the PI team will prioritize a core number of the elicited requirements and develop a first prototype. Given the availability of resources, the envisaged prototype will demonstrate:

1. Automated dialogues between PI and a source in the form of questions and answers to collect and verify information and send feedback to the source
2. Basic credibility analysis by means of triangulation of the information received
3. Possibility to attribute a reliability score to sources.

Thank you for your continuous support!