Home
{
explore
}
what we do
AI Services
Playbooks
Industries
{
Company
}
How we work
Case studIEs
insights
Book a call
{
or email us
}
[email protected]

Insights

Articles
On A Mission
MCP Audit: How Tool Definitions Eat an Agent's Context Window
{
AI Strategy & Readiness
}
October 19, 2026

MCP Audit: How Tool Definitions Eat an Agent's Context Window

MCP tool definitions can fill nearly three-quarters of an agent's context window. Our ten-minute MCP audit shows how to check yours and win that space back.

Read more
Three Levers That Make a Coding Agent Build What You Meant
{
AI Deployment & Integration
}
October 19, 2026

Three Levers That Make a Coding Agent Build What You Meant

A coding agent always builds something. Three levers in the specification raise the odds it builds what you meant, and a worked email and SMS example shows how.

Read more
Agentic UAT: Let an Agent Test Before Your Users Do
{
AI Strategy & Readiness
}
October 11, 2026

Agentic UAT: Let an Agent Test Before Your Users Do

An agent UAT pass runs your acceptance criteria in a real browser before user testing. We run it on UIs and APIs, and the prompt template is included.

Read more
Atomic: AI Coding Assistant Skills That Work in Claude, Codex and Copilot
{
AI Strategy & Readiness
}
September 30, 2026

Atomic: AI Coding Assistant Skills That Work in Claude, Codex and Copilot

Testing workflows built for one AI coding assistant rarely survive a switch. Atomic ships the same skills to Claude Code, OpenAI Codex and GitHub Copilot.

Read more
Spec Drift: Fix the Spec, Fix It Once
{
AI Strategy & Readiness
}
September 30, 2026

Spec Drift: Fix the Spec, Fix It Once

Hand-patching agent output fixes the bug once and lets the spec drift. Put the fix in the spec, regenerate, and it sticks. A worked example from ACDC practice.

Read more
Everyone's Chasing the Future. It lives in the past.
{
AI Strategy & Readiness
}
April 2, 2026

Everyone's Chasing the Future. It lives in the past.

The biggest defensible opportunity in enterprise AI is buried in forty years of COBOL and technical debt nobody wants to open.

Read more
Introducing ATOM - the Agentic Engineering Target Operating Model
{
AI Strategy & Readiness
}
March 31, 2026

Introducing ATOM - the Agentic Engineering Target Operating Model

Nick Martin on what changed late in 2025, when the cost of building software collapsed. ATOM is the operating model MISSION+ built in response.

Read more
Uncomfortably Simple in the Age of AI
{
AI Strategy & Readiness
}
March 30, 2026

Uncomfortably Simple in the Age of AI

We argued for shipping embarrassingly small. Agents now write working code from a spec in hours, so the temptation to build more is real. A revisit.

Read more
Introducing REX - our Innovation Agent
{
AI Strategy & Readiness
}
March 3, 2026

Introducing REX - our Innovation Agent

Every organisation sits on hundreds of ideas from its own people. REX structures them, spots the duplicates across departments, and scores what is left.

Read more
From Deployed to Adopted
{
AI Strategy & Readiness
}
February 9, 2026

From Deployed to Adopted

A Singapore bank had the licences and the governance in place, yet usage stayed low. Role-based learning changed that where generic AI training had failed.

Read more
What’s broken in crowdfunding (and how Web3 puts it right)
{
Web3 & Blockchain
}
December 23, 2025

What’s broken in crowdfunding (and how Web3 puts it right)

Crowdfunding promised creators access to capital and delivered clunky platforms and guesswork. Arjun Balraj on what blockchain changes about the model.

Read more
Why MGAs Are Redefining the Future of Insurance
{
Insuretech
}
December 17, 2025

Why MGAs Are Redefining the Future of Insurance

Managing General Agents do not carry risk on their own balance sheet, which lets them move faster than carriers.

Read more
The Journey
{
Product Practices
}
December 17, 2024

The Journey

You have an idea and you want to build it. Validate it first: founders can drive the first prototype themselves, and it should be thrown away afterwards.

Read more
GenAI PoCs: Proof of Complexity
{
Applied AI Solutions
}
September 2, 2024

GenAI PoCs: Proof of Complexity

Gartner expects 30% of generative AI projects to be abandoned after proof of concept. The failures usually come from how teams adopt what the PoC proved.

Read more
Estimates Not Commitments
{
AI Strategy & Readiness
}
July 30, 2024

Estimates Not Commitments

Teams get punished for treating an estimate as a promise, and Agile is less clear on the difference than people assume.

Read more
Shift The Curve
{
People Practices
}
June 30, 2024

Shift The Curve

You cannot hire the 10x engineer, and the best builder alive stalls inside a stifling process. Raising the whole team's output is the better bet.

Read more
Anti Agile - Introducing the Surge Sprint
{
Agile Practices
}
April 19, 2024

Anti Agile - Introducing the Surge Sprint

People complaining about Agile are usually complaining about Scrum. We go back to the Agile Manifesto and introduce the Surge Sprint we run instead.

Read more
Building Great Teams with MISSION+
{
Analytics & Insights
}
November 21, 2023

Building Great Teams with MISSION+

MISSION+ matches technical talent to what a company is actually trying to build. This is the day-to-day practice behind The Art of Strategic Product Building.

Read more
Unlocking Agility
{
AI Deployment & Integration
}
July 12, 2023

Unlocking Agility

Companies say they want to behave like a tech company, then the transformation stalls. Specific changes you can make this week to shift actual behaviour.

Read more
The Art of Strategic Product Building
{
Applied AI Solutions
}
June 27, 2023

The Art of Strategic Product Building

Jumping straight into building is tempting and rarely delivers what you expected. MISSION+ cuts the wasted effort before a line of code gets written.

Read more
Creating a Simple yet Effective Engineering Culture
{
ESG & Sustainability
}
April 3, 2023

Creating a Simple yet Effective Engineering Culture

Engineering culture is easiest to build at the start and awkward to retrofit. The basics to set up early, and how to recover if you missed that window.

Read more
Uncomfortably Simple: A Framework for Builders To Tighten the Feedback Loop
{
Analytics & Insights
}
March 7, 2023

Uncomfortably Simple: A Framework for Builders To Tighten the Feedback Loop

Gall's law says complex systems that work evolved from simple ones that worked. A framework for shipping something embarrassingly small and learning from it.

Read more
Show more
From AlphaGo to Abundance
{
On A Mission
}
October 12, 2026

From AlphaGo to Abundance

AlphaGo was beaten 100-0 within 18 months. Ned on how compounding moves the constraints, plus our new partnerships with Anthropic and OpenAI.

Read more
Hot Chips and Robot Sports
{
On A Mission
}
September 11, 2026

Hot Chips and Robot Sports

Custom inference chips arrived, robots broke two human world records, and a walking biped now costs $399. Agentic engineering grew up too.

Read more
“My model committed more crimes than yours!”
{
On A Mission
}
August 11, 2026

“My model committed more crimes than yours!”

OpenAI hacked Hugging Face, Anthropic claimed three near-felonies, and Felony Bench was born. The labs are now competing to look the most dangerous.

Read more
Fable came back, but the bigger question stayed
{
On A Mission
}
July 13, 2026

Fable came back, but the bigger question stayed

Fable came back after a blip. Enterprises still have to ask whether American technology can be pulled without warning.

Read more
From Tokenmaxxing to Outcomes
{
On A Mission
}
June 11, 2026

From Tokenmaxxing to Outcomes

Nobody asks whether agentic engineering works any more. They ask how to do it without blowing the budget. Token spend has become the number one question.

Read more
Specs, Code, and the Closing Frontier
{
On A Mission
}
May 14, 2026

Specs, Code, and the Closing Frontier

Qwen 3.6, Kimi K2.6 and DeepSeek V4 landed in one week. The gap between open-weight Chinese models and closed American ones is now low single digits.

Read more
Agents, Insurance, and the End of Vibe Coding
{
On A Mission
}
April 9, 2026

Agents, Insurance, and the End of Vibe Coding

Letting an agent make its own decisions raises velocity and a pile of questions about quality and security. Two frameworks we use to answer them.

Read more
Friction, Fiction, and Prediction
{
On A Mission
}
March 6, 2026

Friction, Fiction, and Prediction

OpenClaw's GitHub stars went vertical, and for many people it was their first brush with an autonomous assistant. Ned reads the curve, and its limits.

Read more
When Speed Breaks Old Rules
{
On A Mission
}
February 5, 2026

When Speed Breaks Old Rules

Five minutes for a model to write good code felt like lightning a year ago. Now it feels slow, and faster models are breaking Agile's old assumptions.

Read more
Build With Intent
{
On A Mission
}
January 16, 2026

Build With Intent

A year ago the call was Google and China, and Gemini 3 says half of that landed. Ned's notes for builders heading into 2026.

Read more
Beyond Prompt Engineering
{
On A Mission
}
November 6, 2025

Beyond Prompt Engineering

AWS went down, then Azure followed. Ned on infrastructure dependence, and why prompt engineering stopped being the interesting problem.

Read more
Building the Future, One Step at a Time
{
On A Mission
}
October 7, 2025

Building the Future, One Step at a Time

Sora 2 understands physics well enough to change digital content production. Plus OpenAI moving into social with cameos, and what Meta is doing about it.

Read more
The Noesis and the Noise
{
On A Mission
}
September 5, 2025

The Noesis and the Noise

Google DeepMind's Nano Banana got called a Photoshop killer, which undersells it. It has a working understanding of how the world fits together.

Read more
A Tax on the Good Guys
{
On A Mission
}
August 8, 2025

A Tax on the Good Guys

Two robotics projects from this month, including K-Scale Labs and its plan for a cheap open source robot that anyone can build and improve.

Read more
Driving the Engine Forward
{
On A Mission
}
July 4, 2025

Driving the Engine Forward

Every stand at SuperAI was an AI platform, and platforms are hard to sell. The opening is vertical: maritime law, SE Asia corporate secretarial, pick a domain.

Read more
Ship Fast, Share More
{
On A Mission
}
March 7, 2025

Ship Fast, Share More

Vibe coding is the new normal and Claude Code is the current contender. Try it, but expect patchier results than the hype suggests.

Read more
{
Work with us
}

Start with one contained project.

Bring us the part that is stuck, whether in a team or a workflow.
Book 30 minutes with an engineer who has taken agentic delivery into production inside a regulated business.
Let's connect