Most of our Central engagements start the same way: a 20-minute call where the owner describes a workflow we've heard before in shape but never in detail. Semantic search (RAG) is then designed against the detail, not the shape.
Built and supported here - the way a Central business would actually use it.
Our goal is to give Central businesses a three-day weekend, so people can spend more time with their families and the people they love :)
What semantic search (RAG) actually does
Search that understands intent, not just keywords. Your team types what they mean - and gets the right document, ticket, or product from across every system, with citations.
- 01 Indexes Drive, SharePoint, Notion, Slack, your CRM
- 02 Returns answers with source links - no hallucinations
- 03 Permissioned so staff only see what they should
- 04 Re-indexes nightly so results stay fresh
Built on: Pinecone Claude Postgres pgvector Vercel AI SDK
Without Kiwi Dynamics
With Kiwi Dynamics
What you actually get
Every engagement is scoped and quoted up front. This is what is in the box.
- A working pilot, in productionNot a prototype on someone's laptop. The first slice of semantic search (RAG) runs against real work within weeks.
- Your data stays yoursIt runs on your accounts and your tools. If we part ways you keep the system and everything in it.
- The workflow mapped before codeWe write down what good looks like for Central businesses first, so nobody is guessing at handover.
- Support after it landsThe people who built it stay reachable when the business changes shape around it.
How semantic search (RAG) compares
The two things most businesses do instead, and where each one runs out.
| Hiring for it | An off-the-shelf tool | Kiwi Dynamics | |
|---|---|---|---|
| Fit to how you work | Fits perfectly, costs a salary | You bend your process to suit the tool | Built around the workflow you already run |
| Time to something useful | Immediate, and permanent | Quick to switch on, slow to make fit | A working slice in weeks, then hardened |
| Who owns the data | You do | The vendor, on the vendor's terms | You do, in your own accounts |
| When it breaks | That person sorts it, if they are in | A support queue and a ticket number | The people who built it |
| What it costs | A salary, every year, forever | Per seat, forever, used or not | Scoped and quoted up front |
The Central context, plainly.
- Central is Hong Kong's financial core, home to the regional headquarters of most major global banks - AI here has to meet a market built on precision, compliance and speed of execution.
- Hong Kong's stock exchange, the regional HQs of the world's biggest banks, and a dense wealth management and private banking sector all sit within a few blocks. Businesses here expect AI that's compliant-by-default and genuinely production-ready, not a pilot.
We work with teams across Central: Admiralty · Sheung Wan · IFC · Mid-Levels · Wan Chai.
Talk to us about this →
How we build semantic search (RAG) for a Central team.
We scope narrow, ship a working pilot, then harden it into production. The first slice is the highest-leverage workflow for your Central business, so value lands before the build is finished. AI search over your data.
Talk to usHow the work runs
- Find the one workflowWe look for the job costing Central businesses the most hours or the most leads, and deliberately ignore the rest for now.
- Ship a working sliceA narrow version goes into production in weeks, against real work, so the value shows up before the build is finished.
- Prove it, then widenWe measure it against what the work cost before. If it does not pay for itself, we say so rather than scaling it.
- Harden and hand overLogging, fallbacks and a real handover, so it keeps running when we are not in the room.
The outcome for Central teams
What changes for Central teams after this lands: the work that used to need a person stays done, the work that needs a person gets done with their attention undivided. Average search time drops from 6 minutes to 12 seconds.
Not your typical AI agency.
Honest about what AI can and cannot do
Ships the one workflow that pays for itself
Hours given back, never the size of the invoice
*Every engagement is scoped and quoted up front. Results vary by workflow and business.
FAQ
When does semantic search (RAG) actually pay back?
Inside the first quarter, in our experience. We pick the first slice specifically because it's the highest-leverage workflow for a Central - so the savings start landing before the rest of the build is finished.
Do you do hourly billing or fixed price?
Fixed price for the pilot, every time. After that it's your call - fixed price per milestone or a small monthly retainer for ongoing iteration. We don't run open-ended T&M because it disincentivises us from finishing.
Do you have proof this works for Central businesses?
Direct case study: Average search time drops from 6 minutes to 12 seconds. Happy to walk you through full numbers on a call.
Will this run on our own infrastructure?
Yes, where it makes sense. Semantic search (RAG) can sit entirely in your cloud account, with model calls routed through endpoints you control. We default to Pinecone, Claude, Postgres pgvector, Vercel AI SDK but the architecture supports your existing platform choices.
Talk to us about this
Tell us what you're trying to do and we'll reply with how we'd build it - no obligation.