Typical approach
You pay for effort
The promise is kept once the agreed team has worked for the agreed period. Finding out why results didn’t come is left to you.
Services / Outcome-based AI-driven development
Services
We don’t stop at delivering code. We work backward from your business goals and outcomes, then build what achieves them. With AI-driven development (AIDD) built into our process, we build fast, validate early, and keep improving.
This service isn’t about supplying engineers by headcount. We work on the front line of your business, and we start by deciding with you what success looks like.
This service isn’t about supplying engineers by headcount. We work on the front line of your business, and we start by deciding with you what success looks like.
Sound familiar?
You explain what you need, but what comes back isn’t what you expected. Getting on the same page takes time, every time.
The product is live, but the business numbers haven’t moved. It isn’t even clear what to fix.
Turning a decision into a working product takes so long that, by the time you can judge it, the assumptions have changed.
You’ve tried AI and other new technology, but it hasn’t improved daily business decisions or development speed.
Problem: misalignment Press a button below to see each stage
Business and development aren’t aligned, decisions take a long time to reach the product, and the working system isn’t making money.
We’re accountable for agreed outcomes, not person-months. A dedicated lead works with you, and the KPIs and outcome metrics are set up front.
Business decisions (your company and a NAL project team) and the build (AI-driven development) move fast in parallel. Gaps are fixed at a sync point every sprint.
We build fast, validate early, and keep improving.
The goal is met in a closed loop, from defining the outcome to validating it. The work is accepted, and the value reaches your business.
Key terms Embedded team A setup where a dedicated lead joins your team. The lead understands your business goals and KPIs in depth and guides everything from agreeing on specs to directing development. AI-driven development (AIDD) A way to develop software fast. NAL’s own AI platform and our engineers work together on everything from writing specs to code generation, testing, and refactoring.
This is a conceptual diagram.
Service overview
We work with you to check whether the specs you give us will drive the outcome, rather than building them as received. Deciding and building run in parallel. The aim is to improve both the quality of decisions and the speed of delivery.
Note: We don’t recommend starting before the outcome is defined. We define it with you first.
01 What we’re accountable for
Software development is usually priced in person-monthsA unit of cost: the work of 1 engineer for 1 month. It is the usual way to price development, and you pay for the work done. or some other measure of how much work gets done. What we commit to is not the work that leads up to the outcome, but the agreed outcome itself. The difference shows when things don’t go to plan.
Typical approach
The promise is kept once the agreed team has worked for the agreed period. Finding out why results didn’t come is left to you.
This service
We agree with you up front on what success means, and we’re accountable for it. If we fall short, we use the numbers to find out why. Then we fix either the metrics or the approach with you.
In the typical model (person-month contracts), you pay for the amount of work done.
In NAL’s model, we agree with you first on what success means, and we’re accountable for it.
That’s why we don’t take a spec and build it as is. We start by working with you to check whether the spec will drive the outcome.
02 How we commit to outcomes
When either the business track or the development track stops, the other stops too. So we run the two in parallel and always compare them at set sync points. These sync points are what make our commitment to outcomes work in practice.
Business diagnosis → KPI definition → outcome review every sprint.
Together, we adjust what to build next as the market and users respond.
We always check what was decided against what was built. If they differ, we fix either the KPI or the way we build.
Strategic team setup → AI code generation and fast prototyping → automated testing and CI/CD.
AVATAR CODING, the AI platform we built in-house, speeds up implementation.
Use cases
01
We join before the specs are final and build first what most needs testing. Our design also makes sure we don’t build more than you need.
02
We use the numbers to find what is holding growth back, then make changes in order of impact. We also consider whether a rebuild is needed at all.
03
We work with your team and leave behind ways to make decisions and patterns for development. The goal is for the work to keep running after we step away.
We avoid writing more specs than needed. With AI-driven development, we quickly build a minimum viable product (MVP): only the features needed to test the outcome.
A dedicated lead stays with the project throughout and checks prototypes with you at each sync point. This makes it less likely that problems first appear right before delivery.
03 How we set KPIs
The way we set KPIsKey performance indicators: target values that measure how a business or a development project is progressing. They state, in numbers or conditions, what has to happen by when for the work to count as a success. as outcome metrics varies by project. Is the product already defined, is it still at the testing stage, or will we track your business numbers together? We agree on this in the first diagnosis, then start work.
Approach 01
Use this when the product and its completion criteria are already decided. We write down what must be met for the work to count as done. It suits changes or new features with a clear scope.
Approach 02
Use this when requirements change during the build. For each milestone, we set what should be usable and by when, then plan the next one again. It suits launching a new business.
Approach 03
Your business numbers, such as revenue or user count, become the metrics. Because our results move with your business growth, we can make bolder recommendations.
04 How contracts work
You choose how to pay based on what we commit to, not on how much we work. We offer three models, depending on your goals and how risk is shared.
Model 01
The base team fee is kept low, and we receive an outcome fee when the outcome target agreed in advance is reached. Both sides share the risk and the return.
Model 02
Each development milestone has its own acceptance criteria, and you pay when they are met. This model handles changing requirements well and keeps progress transparent.
Model 03
We receive a set share of the product’s revenue or profit. Upfront development costs are kept down, and we work alongside you as a long-term business partner.
We discuss the specific contract type and terms with you once we have confirmed the scope and timeframe.
Embedded team
We take part in decisions as your “external brain,” not as a contractor.
At NAL, an embedded product lead works inside your business to support your business decisions. The lead runs everything end to end, from analyzing and recommending specs to directing the AI-driven engineering team and validating outcomes. This lets you focus on business decisions and scope approval, with a team that is accountable for outcomes.
Strengths
We run AI-driven development while sharing the organization’s design principles, development rules, and technical knowledge with AI.
A lead who knows Japanese business practices and quality standards well runs the project. How we reach agreement, how detailed our reports are, and how acceptance checks run are all designed around the way Japanese clients work.
How we work
We work in four steps, from defining the goal to validating the impact after release.
We map where your business stands and the goals it must reach. Then we document what counts as an outcome and how to measure it, and both sides agree on it.
Note: Outcome metrics and success conditions stay fixed. Scope priorities (functional requirements) are adjusted every sprint.
Working backward from the agreed KPIs, we build a project team and a parallel process to achieve the outcome. Then we set the sync points between the 2 tracks.
Note: This shortens the lag between decisions and implementation.
Using NAL’s own AI platform, we generate code, write unit tests, and refactor at speed. We release working software early and keep aligning with you at each sync point.
Note: You can check progress on working software.
After go-live, we measure progress against the KPIs set in advance, in numbers. The project closes once both sides agree on the outcome, and we recommend improvements for the next stage of growth.
Note: Formal acceptance happens once the outcome is confirmed.
Our own context platform shares design principles and development rules with AI to support AI-driven development.
If you want a long-term development team of your own rather than an outcome-based engagement, we offer Active Labo, NAL’s dedicated development team service. You can choose either service or combine them, depending on your goals.
FAQ
Some factors, such as how the market responds, are outside our control, so we can’t guarantee the KPI will be met. If it isn’t, we use the numbers to find out why. At a sync point, we discuss with you whether to change the KPI or the approach. How payment links to outcomes, and what happens if a target is missed, are discussed and put in writing in the contract.
Yes. In fact, we recommend talking to us before the requirements are settled. We work out with you what counts as an outcome, and we work alongside you from the design stage so no unnecessary features get built.
We develop in parallel tracks the Agile way, so priorities can be reviewed flexibly at the sync point in each sprint. If a change does more for the target KPI, we take it on and keep going.
Active Labo is for building a dedicated engineering team of your own, and development runs flexibly under a time-and-materials contract. This service is different: we’re accountable for achieving specific business outcomes. You can also use both. For example, you can launch with this service and move to Active Labo for the operations phase.
Engineers always review AI output before it goes into the product. We also generate tests along with the code, and we run quality control to Japanese standards.