With transparent reports, a document library, and a managed queue of applications. A project that works both as a public point of trust and as an operational tool for intake.
A multi-page website for an organization that helps internally displaced people. For someone in stress, the site has to be simple and easy for finding information about available programs, participation conditions, required documents, the application mechanism, and how long the response takes. For the organization's team, the site is the operational entry point. Applications land in a queue where they can be sorted, assigned statuses and owners, and where completed requests are visible.
Mark is the head of a nonprofit organization that helps internally displaced people. The goal is to communicate with people, partners, and donors transparently and in an organized way. Mark knew that trust in nonprofits is built on specifics: understandable programs, clear conditions, a list of documents, transparent reports, and an easy application process. The organization had no technical team, so the solution had to be one they could operate on their own for both content and applications.
A site that calms people down
If a person arrives under time pressure and cannot quickly find conditions, the document list, or the submit button, they leave for chats and phones.
Application types and timelines
Help can be one-off or long-term, with different sets of documents, contacts, and reasons. The form and the back office had to work as structure, not walls of text.
Application queue
What matters is statuses, owners, priorities, repeat applications, and contact history. Otherwise an application turns into a free-form email that can be forgotten.
Public visibility and transparency
A clear structure in reports and documents — period, program, file, confirmation — matters both for donors and for people who want to see that the organization is working.
Team autonomy
News, distribution schedules, changes to program conditions, new documents — this is daily work the team has to handle themselves, consistently and without waiting.
AI assistance
In humanitarian work, inaccuracies and guesswork are a hit to reputation, so AI can only be used for compression, structuring, and highlighting gaps.
We designed the public part on Next.js to show programs, news, materials, and reports in a strict structure so that a user can quickly find answers and evidence. Meanwhile, Directus became the operational and editorial back office with roles, publication statuses, reference data, and a document library.
The website as a help route
We designed the site structure so a person reaches the right program or application submission in 1–2 clicks, without reading a generic copy or getting lost in the menu.
Aid programs as managed scenarios
We built clear program cards in the CMS with conditions, document lists, response times, restrictions, and next steps, so the team does not invent a new format every time.
A managed application queue
We designed application forms that immediately create records with statuses, application types, contacts, location, attachments, and an assigned owner, so the whole team can see the request.
Directus as the operational back office
We gave editors and operators understandable working screens, separate access rights, and introduced drafts and moderation so that content and applications go through a minimum quality check.
An AI assistant for triage of applications
We made an AI assistant that condenses an application into a short summary, highlights what is missing for processing, and suggests clarifying questions. For content, it flags where dates, links, or files are missing before publication.
A Next.js frontend for mobile UX and weak internet
We built pages as lightweight templates for programs, news, documents, and reports with good reading on a phone, a fast first screen, and convenient forms.
We started with typical requests to the nonprofit and identified the most frequent needs of the people, what they forget to include, the causes of conflicts, FAQ, and the steps where the team most often loses contact with the applicant. Then we got specific scenarios from the client: who can be helped and who cannot, what restrictions apply, what counts as a minimum data package, and which decisions the team is allowed to make immediately. At this stage we also fixed the limit: what can be automated, where AI can only suggest, and what AI must not touch at all.
The result is a working tool that fully covers the tasks assigned to it.
we launched a managed editorial process
manual clarifications on applications
in the nonprofit's corporate mailbox
to process one application
with no issues in public reporting
where users find what they need without email exchanges
Or send us a message and we'll get back to you within 15 minutes during business hours