Replacing a no-code workspace with an owned, private system
Emerald Consulting OS
The operating system for a student-run consulting club, replacing a monday.com workspace and its automations, with privacy built into how the data is stored.
The product story
- User
- The 5-person leadership team of INSEAD's pro bono consultancy, who run recruiting, clients and offers, and the applicants and clients who use its public pages.
- Problem
- The club ran on a monday.com workspace full of automations, but it holds applicants' CVs and clients' details, so it needs to own that data and control exactly who sees what. The aim was one system for the whole year, with no side spreadsheets.
- Options
- Keep adapting the no-code workspace, or build a system the club owns, with privacy in how the data is stored rather than in a setting.
- Decision
- Build the club its own system. One concrete choice: a reviewer sees a candidate's name, contact details and CV only after saving their own scores, so first reviews are fair.
- Outcome
- The public pages are live and the leadership team runs the club on it. On my CV I report that, against prior years, client relationship retention rose 40% and the share converting into projects rose 20%. The counts and comparison periods behind those figures are in the club's records and are not shown here, so this page does not claim the software caused all of the rise.
- Next
- Publish the comparison periods and counts behind the two figures, with the club's agreement.
How it works
How one module definition drives the club's whole system, with privacy in the data layer. Select a layer to open it.
Problem
A student consulting club ran its recruiting, clients and offers in a monday.com workspace full of automations. The club handles applicants' CVs and clients' details, so it needs to own that data and control exactly who sees what. The aim was for the leadership team to run the whole year in one system, with no side spreadsheets.
Approach
Build the club its own system, and make privacy part of how the data is stored rather than a setting. Each module is defined once, and its database tables, forms, views, imports, exports and permissions are all generated from that one definition.
Architecture
It runs on Postgres everywhere: hosted on Supabase behind Vercel in production, and a version that runs inside the app itself on a laptop and in tests. The app is built with Next.js. A change to how pages read the database cuts its estimated data transfer from about 15 GB to about 1.4 GB a month. Every table is locked down at the database level, so the hosted database returns nothing directly even if a public key leaked, and the test suite fails if a new table is added without that protection.
What I built
- Fair first reviews. A reviewer only sees a candidate's name, contact details and CV after saving their own scores. Interview rounds let applicants book their own slot from a public page, and interviewers score each interview against a rubric, with the results ranked on one page.
- A safe client view. Clients see a dashboard built from an approved list of fields, so internal comments, minutes, risks and files never reach them.
- Protection for client names. Invitations and feedback forms use single-use links, and saving a case study is blocked if it names the client, unless an admin overrides it with a recorded reason.
- A CRM and sales pipeline. Organisations, people and interactions sit in one place, and clients, deals and partners link to them. Deals move through stages from first lead to closed, with every change recorded, and a project request from the public site or a member's referral opens a deal automatically.
- Forms and a website the club can edit. Public forms and website content are edited as drafts and published as versions, and any earlier version can be restored. Applications only open during a set window, which the server enforces.
- An outreach email builder. Leadership members draft a first email to a prospect from the CRM's own records, and the draft is checked for length, a single clear ask and writing that sounds machine-made.
Results
Its test suite passes on the main branch, and the public pages are live.
| Check | Expected or threshold | Actual | Result |
|---|---|---|---|
| Test suite on main | All green | All green | Pass |
| Public apply page | Live and answering | Answering, form shown | Pass |
Stack
- Next.js 16
- TypeScript
- Tailwind 4
- shadcn/ui
- Drizzle
- Supabase Postgres
- PGlite
- Zod
- Vitest
- Vercel
How these figures are sourced
The outcome figure is agent-reported: it is taken from a test log or report written by the agents that built and ran the system. The status was read from the project's goal check. Every figure and claim on this page is tied to a source file or a CV line in the site's content.