Semantic search (RAG) in Kaikōura.
For Kaikōura operators who want a working pilot in weeks, not a year-long programme.
Kaikōura welcomes around a million visitors a year. We automate bookings, enquiries and reviews so small teams keep up with big seasons.
Business in Kaikōura.
Tourism, directly and indirectly, is the major contributor to Kaikōura's GDP, built on whales, the coastline and crayfish, alongside the district's fishing and farming heritage.
A deep undersea canyon brings Giant Sperm Whales close to the coast, and Whale Watch has run since 1987.
An AI agency, minus the theatre.
*Every engagement is scoped and quoted up front. Results vary by workflow and business.
Already buried in admin? We'll sort the first workflow for you :)
Nine hours a week of admin is 468 hours a year. That could drop to about 52.*
*Illustrative, based on 9 hours a week of admin at a typical 80% automation rate. The calculator uses your own numbers.
Knows the New Zealand rules.
Ask questions across your manuals, contracts and records and get answers with the source attached, not a guess.
- Data that stays in New Zealand
- Serious breaches get reported
- What the Privacy Commissioner expects
Small team.
Real systems.
Zero fluff.
What semantic search (RAG) covers
What changes for Kaikōura 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.
- Indexes Drive, SharePoint, Notion, Slack, your CRM
- Returns answers with source links - no hallucinations
- Permissioned so staff only see what they should
- Re-indexes nightly so results stay fresh
- No shelfware or licences you do not use
- No annual contracts
On the tools or running at scale, the admin still eats the week :)
The questions a public sector buyer asks, answered up front: privacy impact, the NZISM, where the data lives and how an algorithm is held to account.
*Illustrative example based on nine hours a week of admin. Actual results vary by workflow and business.
Questions Kaikōura businesses ask
Can you work with our existing systems?
Yes. The default semantic search (RAG) stack we reach for is Pinecone, Claude, Postgres pgvector, Vercel AI SDK, but we'll bend it around whatever you already run - Xero, HubSpot, Shopify, Cin7, your own in-house apps. The discovery week maps every data source before any build starts.
Are there hidden costs we should plan for?
Three to know about: model/API spend (which we set up under your own account, not ours, so you see and control it), any new SaaS subscriptions we recommend, and your team's time during rollout. We surface all three in the quote so there are no surprises.
Anyone else in this space using semantic search (RAG)?
Plenty. Average search time drops from 6 minutes to 12 seconds. The interesting question is rarely "does it work" - it's "is your team ready to use the output." That's what we'd scope on the call.
How does this fit the Public Service AI Framework?
The framework, led by the Government Chief Digital Officer, asks for safe, transparent and accountable AI. Agencies that sign the Algorithm Charter also commit to transparency in how they use algorithms.
Do you respond to government tenders?
Most large government contracts are advertised on GETS, the Government Electronic Tender Service, where suppliers register with RealMe.
Talk to us about semantic search (RAG)
StartContact us
Tell us what you're trying to do and we'll reply with how we'd build it, no obligation.
Was this helpful?
Thanks, your comment is with our team.