An ERP across four businesses: pipeline, finance and knowledge base in one system
The Task
The pipeline lived in one file, finance in another, clients in a third. The data drifted apart within a week and there was nothing to reconcile it against.
What We Built
We built Unika Hub on Next.js and Supabase: CRM, finance and P&L, a knowledge base and an ad-account registry in one database, separated by company. Plus deduplication of inbound leads by external id and phone.
Timeline
Guide for a similar project: 6-8 weeks for a base CRM and finance configuration.
Tools Used
Case section
Context
A familiar situation in a growing company: every process acquired its own file. With one business line that works. With four, the data starts contradicting itself and nobody knows which version is right.
Case section
What we built
CRM, finance, knowledge base and the ad-account registry now sit in one database. Separation by company is a field in the schema rather than separate systems, so you can look at one business or at the whole picture.
Case section
What the migration revealed
Half the procedures existed only in people's heads. Some figures contradicted each other across files — not because anyone erred, but because nobody had ever reconciled them. Digitising does not create these problems; it makes them visible.
Case section
The lesson
Before a process can be automated it has to be described. That is management work, and it pays off on its own whether or not agents ever follow.
What changed
Pipeline, finance and clients lived in separate files and drifted apart within a week
One database separated by company, where the data reconciles itself
The same lead was created again when the person came through another channel
Deduplication by external id and phone with a 30-day window
Key Results
- four businesses run in one system with separated access
- repeat enquiries no longer create duplicates
- finance and pipeline stopped contradicting each other
Related Articles
AI Agents for Business in Uzbekistan: Where to Start and When It Pays Off
A practical guide: which processes in Tashkent and the CIS are worth giving to AI agents, when a 2–3 week pilot makes sense, and when automation will not pay back. With links to real UNIKA cases.
How AI Automation Saves Businesses $48,000 a Year: The Full Calculation
A transparent calculation model showing how a 20-person company in Tashkent reaches $48K in annual savings. Every assumption is on the table — plug in your own numbers and check the math yourself.
5 Processes to Automate First
Not every process is equally worth automating. Here are the 5 tasks that deliver the most impact for the least investment, plus the prioritisation rule for picking your first project.