Skip to content
Flowscale

Projects and case studies

Four demonstration systems, with public code. Each project shows the problem, the automated solution and the targeted result.

These projects are demonstrations, not client cases. We will publish client results here as soon as a client allows it.

Internal knowledgeDemo project

AI agent and company knowledge base (RAG)

The problem

Your teams lose a lot of time hunting for a procedure, a document or an answer across scattered files. Onboarding new hires slows everyone down.

The automated solution

  • Automatic ingestion of internal PDFs and documents, chunked and indexed in a vector database.
  • An agent answers team questions grounded only in your documents (procedures, SOPs).
  • The same agent acts as a tutor for onboarding and employee training.

-80 %

search time

Estimated

Estimated result: Up to 80% less time spent searching for internal information

More about this project

Tools used

  • n8n
  • Supabase / Pinecone
  • OpenAI / Claude API
  • PDF and docs parsing

Workflow architecture

  • Documents (PDFs, procedures)
  • Content parsing
  • Vector database (Supabase / Pinecone)
  • AI agent (OpenAI / Claude)
  • Answer to the team
Real estateDemo project

Real-estate agent: WhatsApp lead qualification and matching

The problem

Buyers and renters message at all hours. Every late reply is a prospect contacting another agency.

The automated solution

  • A conversational AI agent captures the prospect's criteria (budget, area, property type).
  • Dynamic matching compares those criteria against the property catalogue (NocoDB).
  • Listing sheets and photos (Cloudinary) sent in real time on WhatsApp.

< 1 min

to answer the prospect

Target

Target result: Answer the prospect in under a minute, and get more qualified viewings

More about this project

Tools used

  • n8n
  • WhatsApp API
  • NocoDB
  • Cloudinary
  • OpenAI

Workflow architecture

  • WhatsApp message
  • AI agent (OpenAI): criteria
  • Property catalogue (NocoDB)
  • Photos (Cloudinary)
  • WhatsApp reply
HealthcareDemo project

Clinic assistant: appointment booking on WhatsApp

The problem

Front-desk staff are swamped with booking calls and messages, and no-shows leave empty slots.

The automated solution

  • The assistant handles consultation requests on WhatsApp.
  • It dynamically books an available slot in Google Calendar.
  • It sends automatic reminders to the patient.
Target

Target result: Fewer no-shows and a lighter load on the front desk

More about this project

Tools used

  • n8n
  • WhatsApp API
  • Google Calendar
  • Webhooks

Workflow architecture

  • WhatsApp request
  • AI assistant
  • Availability (Google Calendar)
  • Slot booked
  • Reminder on WhatsApp
Human resourcesDemo project

Interactive HR recruiting: application triage in Slack

The problem

Applications arrive via a form, then get copied, sorted and chased by hand. Recruiters lose hours on admin.

The automated solution

  • Applications received via Tally are processed and logged automatically in Google Sheets.
  • Each application is posted in Slack with Approve / Reject buttons (Block Kit).
  • The recruiter's decision updates the tracker with no re-entry.

÷ 3

of admin time

Estimated

Estimated result: Administrative processing time cut by 3x

More about this project

Tools used

  • n8n
  • Tally Forms
  • Slack Block Kit
  • Google Sheets

Workflow architecture

  • Application form (Tally)
  • n8n processing
  • Log (Google Sheets)
  • Slack message with Approve / Reject
  • Tracker updated

Results come from demonstration projects; your results depend on your context.

Frequently asked questions about our projects

How do you measure ROI?

We compare the cost of the automation (setup and running costs) with the cost of the human time saved. The calculator on this site applies that method with your own assumptions.

Why demonstration projects?

Because we prefer verifiable code to promises. Every demo is public: you can open it and test it.

Can I see the code?

Yes. Each project links to its GitHub folder.

A project like this? Let's talk about yours.

Book my audit