Silky Insights

Applied AI · Auckland · NZ, Australia & UK · est. 2021

AI capability, without building an AI department.

Silky Insights is a two-person AI implementation team working with clients across New Zealand, Australia and the United Kingdom. We help organisations add practical AI functionality to existing products, workflows and data environments.

We help you decide what to build, prepare the data needed to build it, and build the AI capability itself.

Talk to us about what’s worth buildingExplore our services
API calls served
20M+

API calls served

Estimated run-rate since October 2023, as at August 2026.

Documents indexed
100,000+

Documents indexed

Indexed across our search systems, as at August 2026.

Longest system in production
4+ years

Longest system in production

Still live, still supported.

Uptime SLA
99.9%

Uptime SLA

Our committed service level.

In their words

They are on our side and on our team.
Alan Young · Enerlytica
“Silky Insights’ approach is a kind of down-to-earth and pragmatic innovation, focusing less on the theoretical potential of artificial intelligence and more on the immediate, measurable impacts it can deliver.”
M2 Magazine
“Chris and Luke are a powerhouse duo who have an amazing complementary set of skills. Already we have seen a huge difference in what we are technically delivering to our customers…”
Annie Johnson · Co-Founder, Humaneer
“The team at Silky constantly evolve to fill in the gaps of our capabilities, with patience and a deep consideration for our commercial environment. We are grateful to have them, and you will be, too.”
Alan Young · Enerlytica
“Silky built us a tool that has enhanced our business capability and released our staff to concentrate on value add rather than mandrolic research…”
Paul Lewis · Director, FOS
  • Enerlytica
  • BusinessDesk
  • Humaneer
  • FOS

The problem

AI projects get stuck between ambition and implementation.

  • The use case is unclear

  • The data is not ready

  • The team is not resourced to build it

Many teams can see the potential of AI, but do not have the internal capability, time, or confidence to turn it into reliable software.

The use case seems promising, but the data is messy. The prototype looks impressive, but does not survive real workflows. Internal teams are busy, hiring is slow, and it is hard to know what is valuable, feasible, or safe to build first.

That is where Silky comes in.

What the work actually is

Messy, high-value information becomes a queryable layer.

Report archives, PDFs, data tables, transcripts, policies, recurring data flows. Ingested, extracted, structured, indexed, then wrapped in a tool that fits the workflow.

Archive as received: 45 years, no schema

Structured, indexed, queryable

Three ways in

Three ways we help

01

Add AI to your product or workflow

We design, build and integrate AI functionality into existing products, platforms, internal tools and customer experiences.

Search, recommendations, summarisation, document generation, classification, structured extraction, workflow automation, chat interfaces, user-facing features.

Typical shape

A development engagement through to a first integrated version. Ongoing support afterwards is optional: we can maintain it, or hand you the code, documentation and resources to run it yourselves.

Wrong for

Teams without a product or workflow to put it into yet. Start at line three.

02

Make your data AI-ready

AI is only as useful as the information underneath it. We clean, structure, connect and organise documents, reports, databases and knowledge sources so they can support reliable AI.

Document ingestion, structured extraction, foundational databases, data pipelines, metadata, vector search, APIs, monitoring, reporting reliability.

Typical shape

A scoped build on the archive or feed that matters. Pipeline support afterwards is optional: yours to keep and run if you would rather.

Wrong for

Anyone hoping to skip this step and go straight to a chat interface.

03

Get a practical AI roadmap

Not every AI idea is worth building. We start with a discovery workshop that digs into the root cause of what is actually going wrong, not just where to bolt AI onto an existing system, then assess feasibility and risk and define a practical roadmap.

That includes advice on data governance and the data-sharing agreements that constrain what you can actually do with your information, so the roadmap reflects what you are allowed to build, not just what is technically possible. The output is not a strategy deck for its own sake. It is a clear path to action, costed and sequenced.

Typical shape

A defined piece of work with one written output. No ongoing commitment attached.

Wrong for

Teams who already know what to build. Go straight to line one.

Proof

Working AI systems, not just advice

We are strongest when we can point to systems that are already in use. Across energy, media, HR and compliance, the pattern is similar: valuable knowledge is trapped in documents, workflows, databases or archives, and teams need a reliable way to turn it into action.

Market reports, PDFs and data tables as received. No schema.

  • AI-Ready Data Foundations
  • AI & Data Strategy Advisory
  • AI Product Implementation
  • Document Intelligence

Enerlytica

Energy market intelligence · New Zealand

Enerlytica works with dense, high-value energy market information: reports, PDFs, data tables and recurring data flows. The opportunity was not simply to “add AI”, but to create a more reliable information layer that could support faster research, better reporting and future AI-enabled products.

We have worked with Enerlytica continuously since 2023: document ingestion, table extraction, foundational database design and semantic search first, then automating their daily market reports end-to-end and rebuilding their website on top of the same foundation. Alongside the build, we advise on Enerlytica’s ongoing AI and data strategy: what to prioritise next as the market, the data sources and the technology change.

Read the Enerlytica case study →

Raw meeting conversation, context and user input.

  • Product Implementation
  • Product Strategy & UX Advisory
  • Embedded AI
  • Workflow Copilots
  • Document Generation

Humaneer

HR and compliance product · Australia

Humaneer needed practical AI capability layered into Mak, its HR and compliance product: turning transcripts, context, policies and user inputs into outputs that support real HR and compliance workflows.

Alongside the build, we advised on UX and product strategy and direction throughout the relationship, how the compliance intelligence should surface, what to prioritise next, how Mak should be positioned. Humaneer’s team made every product and strategic call; ours was advice, not the decision.

This is the type of AI implementation where trust, traceability and workflow fit matter. The engagement has continued since, extending Mak from a chat assistant into a system that listens to meetings directly and recommends what to do next, not just what happened.

Read the Humaneer case study →

Years of articles, company coverage and journalist profiles, sitting as inventory almost no reader reaches.

  • Archive Discovery
  • User-facing AI Features
  • Product Integration
  • Semantic Search

BusinessDesk

Business journalism · New Zealand

A large and growing archive of business journalism, market information, company coverage and journalist profiles. The opportunity was to improve how readers discovered relevant information, and how the platform connected users with the right content at the right time.

Silky built AI-powered search and recommendation that goes beyond keyword matching: understanding user queries and article context to surface relevant articles, journalist profiles and market pages across the archive. It has been running in production since 2023.

Read the BusinessDesk case study →

The set of companies FOS identifies as worth tracking. Their list, not ours.

  • Workflow Automation
  • Competitor Intelligence
  • Custom Scrapers
  • Open-Source Monitoring
  • Hosted & Supported

FOS

Business growth consultancy · United Kingdom

Facilities Optimised Solutions advises facilities management and construction businesses on growth, acquisition, bid strategy and pricing. Knowing what a defined set of competitors is actually doing is core to that advice, and it used to mean staff searching the internet for updates by hand.

We selected the archives and news sites that carry real signal, built scrapers for the sources with no usable API, and turned the result into a monthly update on company activity across a watchlist FOS defines. It has run for two years, letting them prioritise which future customers to pursue and advise existing ones on how their competition is performing.

Read the FOS case study →

Why Silky

Senior AI capability without the hiring risk

Hiring an internal AI team can make sense once AI becomes core to your business. But hiring before you know what you need can be slow, expensive and risky.

Silky helps you prove the use case, define the architecture, build the first working system, and create a clearer basis for future investment.

You work directly with the people designing and building the system. No large consulting team. No endless strategy programme. No handover gap between advice and implementation.

A lower-risk way to start

The usual path

  • Hire before the role is clear
  • Run broad AI workshops
  • Build demos that do not fit real workflows
  • Leave internal teams to figure out AI patterns alone
  • Commit budget before value is proven

The Silky path

  • Start with the workflow, product or decision that matters
  • Assess value, feasibility, data readiness and risk
  • Build a useful first version with real data
  • Integrate it into the existing environment
  • Scale what works, once value is clear

How we work

From uncertainty to working capability

We usually start small, prove value quickly, and scale what works.

  1. 01

    Understand the real problem

    We start with a discovery workshop that gets to the root cause of the workflow, product, decision or information problem you are trying to improve, not just where AI could be layered on top.

  2. 02

    Identify the right AI opportunity

    We assess value, feasibility, data readiness, risk, and the fastest path to a useful first version.

  3. 03

    Build with real data and real users

    We create practical prototypes and MVPs that can be tested in the environment they are meant to serve.

  4. 04

    Integrate into the workflow

    We design AI to fit into existing products, systems, permissions and user behaviour.

  5. 05

    Improve and scale

    We harden what works with evaluation, monitoring, feedback loops, documentation and production support.

Under the hood

Built with, not locked to

Tool-agnostic and model-agnostic by design. This is the stack we reach for most.

Models

  • OpenAIOpenAI
  • AnthropicAnthropic
  • GeminiGemini

Cloud

  • AWSAWS
  • AzureAzure
  • Google CloudGoogle Cloud

Data & retrieval

  • PostgresPostgres
  • MySQLMySQL
  • SupabaseSupabase
  • ElasticsearchElasticsearch
  • LangChainLangChain
  • HaystackHaystack
  • LangSmith
  • DirectusDirectus
  • StrapiStrapi

App & delivery

  • Next.jsNext.js
  • NuxtNuxt
  • VercelVercel
  • RailwayRailway
  • PythonPython

Automation

  • ZapierZapier
  • Inngest

Who we are

Two people. Both senior. Both on your project.

Chris Matthews, co-founder of Silky Insights

Chris Matthews

Co-founder · Strategy, Product, Commercial

Eight years running client-facing AI discovery and carrying early-stage ideas through to governed production systems.

Luke Doyle, co-founder and CTO of Silky Insights

Luke Doyle

Co-founder · Chief Technology Officer

Builds the systems: RAG, document ingestion, vector search, structured outputs, and delivery across Azure, GCP and AWS.

Questions buyers actually ask

The objections, answered plainly

What does an engagement cost?
It is scoped to the work, and you get a number before you commit to anything. A development engagement covers the build; support afterwards is optional: we can maintain the system, or give you the code, documentation and resources to run it in-house.
Where does our data go?
Into infrastructure you control, in the region you choose. We are tool-agnostic and model-agnostic: nothing is pinned to one vendor, and your data is not used to train anything. Azure, GCP and AWS are all fine.
Why a two-person firm over a large consultancy?
Because the people who advise are the people who build. No bench, no juniors learning on your budget, no handover gap. If that trade-off does not suit the scale of your programme, we will tell you.
What if the first build does not work?
You find out in weeks, not quarters, at a scope small enough that finding out is cheap. Feasibility and data readiness are assessed before anyone writes production code.
Who owns the IP?
You own the developed IP on payment. We retain our background IP: the tooling and patterns we bring with us. That is in the contract, not just the sales conversation.
What happens if Silky stops operating?
Every contract carries a source-code continuity clause. You keep the code, the infrastructure and the documentation, and you can hand it to anyone.

Let’s work out what is worth building.

Whether you are trying to add AI into an existing product, make your data usable for AI, or understand where AI should fit in your business, we can help you find a practical next step.

Send us a noteEmail us directly

What happens on the call

  • Thirty minutes. No deck, no discovery phase.
  • You describe the workflow or the data problem.
  • We say whether AI is the right tool, and what a first useful version looks like.
  • If it is not worth building, we will say so.

We only use this to reply to you. See our privacy policy.

Talk to us directly

chris.matthews@silkyinsights.com+64 21 432 321

We reply within one business day.

All case studies →