Here is a number that should bother you if you are spending money on AI. According to the UK government’s own AI adoption research, published in February 2026, 77% of businesses using AI have not seen any change in revenue. Not a small change. Not a delayed change. No change at all.
That is not a rounding error. That is four out of five businesses spending time and money on AI and getting nothing back on the top line.
I have been building software for 27 years. I have watched this pattern repeat with every technology wave. Cloud. Microservices. Big data. Mobile. Every time, businesses rush to adopt the technology without first asking the only question that matters: what are we actually trying to achieve?
They build AI because everyone is building AI. They run a pilot because the board wants a pilot. They chat with a model and call it a strategy. And then they wonder why 77% of them see no revenue impact.
The problem is not the AI. The problem is that they started with the technology instead of the objective.
This Is Not a New Idea (But It Needed One)
The smart people at Atlassian and Thoughtworks have been writing about something called outcome-driven development for a few years now. The core idea is sound: stop measuring teams by what they ship, start measuring them by what changes in the business because of what they ship.
Output is features delivered. Outcome is churn down 3%, conversion up 10%, support tickets halved. Outcome-driven development says the second list is the only one that matters. It rewrites roadmaps as goals instead of feature lists. It uses feature flags, continuous deployment and analytics to test whether something actually moved the needle before you commit to it.
It is the right philosophy. But it stops short in one important way. It treats the team as the hunter and the technology as the tool. The team decides what to build, builds it, measures it, and adjusts. That is good. But it is still a human team doing the finding, the building and the measuring, just with better metrics.
What I am proposing takes that philosophy and adds something the original authors could not have predicted: AI agents that do not just measure outcomes but actively go after them.
I am calling it OKR-driven development.
What Is OKR-Driven Development
OKRs, or Objectives and Key Results, have been around since Andy Grove popularised them at Intel in the 1970s. John Doerr took them to Google in 1999 and they spread from there. You set an objective, which is what you want to achieve. You attach key results, which are how you know you got there. Simple, proven, effective. The OKR framework is already the goal-setting engine behind outcome-driven development. Atlassian uses it exactly that way.
What I am proposing is making OKRs the starting point for technology development, not just the measurement framework.
In most businesses, technology development happens in one of two ways. Either IT responds to tickets, fixing problems and patching issues as they arise. Or a team picks a technology, usually because someone read about it on LinkedIn or saw a competitor doing it, and builds something with it to see what happens.
Both approaches start with the technology or the problem. Neither starts with the objective.
OKR-driven development starts with the objective. You decide what your business is trying to achieve. Build revenue. Reduce churn. Enter a new market. Cut operational costs. Whatever it is, you set the OKR first. Then you build the technology to serve that objective. And the technology, AI included, is measured against whether it moves the key result.
This sounds obvious. It is obvious. But almost nobody does it, and the 77% revenue failure rate proves it.
The Difference: Active Pursuit, Not Passive Measurement
Here is where OKR-driven development goes beyond outcome-driven development.
Outcome-driven development says: “measure whether what you shipped moved a business metric.” That is a measurement philosophy. It tells you whether your work mattered, after the fact.
OKR-driven development says: “set the objective, then let AI agents actively hunt for ways to move it.” That is a development methodology with a mechanism. The AI does not just report on whether outcomes happened. It proactively looks for opportunities to make them happen.
The distinction matters. Outcome-driven development is retrospective. You ship something, then check if it worked. OKR-driven development is prospective. You set the target, and the AI goes looking for ways to hit it before you have even decided what to build.
Think of it this way. Outcome-driven development gives the team a scoreboard. OKR-driven development gives the AI a mission.
Why It Changes Everything
When you lead with OKRs, the questions change.
Instead of “what can we do with AI?” you ask “what is our objective and can AI help us reach it?” That is a completely different conversation. One is technology-led. The other is business-led.
Instead of “let us pilot a chatbot and see what happens” you ask “our objective is to reduce customer support response time by 40%. Can AI help us do that, and if so, how?” Now you have a measurable target, a clear use case, and a way to know if it worked.
Instead of building tools that might be useful someday, you build tools that have a job to do. And you can measure whether they did it.
The shift is from reactive to deliberate. From “what problems can we fix?” to “what outcomes are we going after?” From patching and improving to building and measuring.
This matters for AI specifically because AI is not like other technology. A CRM does what you tell it. An API does what it is designed to do. AI is different. AI can hunt. It can find opportunities you did not know existed. It can analyse data, spot patterns, suggest actions and even take actions on its own.
But only if it knows what it is looking for.
Point AI at your business without an objective and it will find things. None of them will be the things you need. Point AI at a clear objective with measurable key results and it becomes a focused, accountable member of your team.
What OKR-Driven Development Looks Like in Practice
Let me make this concrete. Here is how it works across different areas of a business.
Revenue growth. Objective: increase average customer spend by 15% this quarter. Key result: 40% of existing customers buy a second product line. The AI does not wander around looking for things to do. It analyses purchase patterns, identifies cross-sell opportunities, scores leads by likelihood to convert, and surfaces the top 20 accounts to target this week.
Customer retention. Objective: reduce monthly churn from 5% to 3%. Key result: identify and save 30 at-risk accounts per month. The AI monitors usage data, support ticket patterns and engagement signals, flags accounts showing churn indicators before they leave, and drafts retention outreach for the team to review.
Sales pipeline. Objective: increase qualified pipeline by 25%. Key result: 50 new SQLs per month from inbound channels. The AI scores inbound enquiries, enriches them with company data, routes hot leads to sales immediately, and nurtures the rest with tailored content based on their industry and behaviour.
Operational efficiency. Objective: reduce order processing time by 50%. Key result: 95% of orders processed within 2 hours. The AI reads incoming orders, extracts data from any format, validates against inventory, flags exceptions and routes the clean ones straight through without human touch.
Technical debt. Objective: reduce system maintenance overhead by 30%. Key result: decommission 5 legacy systems by year end. The AI maps dependencies, identifies which systems are safe to retire, prioritises migration work by risk and impact, and tracks progress against the target.
Every one of these starts with the objective. The technology is the servant, not the master. And every key result is measurable, so you know at the end of the quarter whether it worked.
The AI Agent as an OKR Hunter
Here is where this gets interesting, and where I think the real opportunity is for businesses that get this right.
When you set OKRs and share them with your AI tools and agents, you give those agents a mission. Not a task. A mission. The difference matters.
A task is “summarise this document.” A mission is “our objective is to reduce churn to 3%. Find me the accounts most likely to leave this month and tell me why.”
An AI agent with a task does what you ask. An AI agent with a mission goes looking for ways to achieve it. It can proactively analyse data, spot patterns, draft recommendations and even execute actions, all in service of the objective you set.
This is not theoretical. We are building this now. The AI tools available today, the agentic frameworks, the data analysis capabilities, they are all good enough to do this. What they have been missing is direction. They have been given hammers and told to find nails. OKR-driven development gives them a blueprint and tells them what to build.
Outcome-driven development started the conversation. It said: stop counting features, start counting impact. That was right. OKR-driven development finishes the thought. It says: set the objective, give the AI a mission, and let it hunt.
What This Means for Your Business
If you are a UK SME owner reading this and thinking “we have been doing AI without objectives,” you are not alone. That describes most of the market. The good news is that the fix is not complicated, and it does not require new technology.
It requires you to sit down and decide what you are actually trying to achieve. Set three to five OKRs for the quarter. Make them specific. Make them measurable. Then look at every AI tool, every pilot, every technology investment and ask one question: does this serve one of our OKRs?
If it does not, stop doing it. I mean that. Stop. The 77% revenue failure rate is built on businesses that never asked that question.
If you are not sure how to start, that is exactly what we help with. Our AI Opportunity Audit is designed to identify where AI can actually move the needle on your business objectives, not just where it might be technically interesting. And if you want to talk through what OKR-driven development could look like for your specific business, book a free discovery call.
The businesses that figure this out first are going to pull ahead. The ones that keep building AI without objectives are going to keep showing up in the 77%.
Related Services
- AI Strategy and Roadmap - Define your objectives, then build the technology to match
- AI Opportunity Audit - Find where AI can actually move your numbers
- Contact us - Book a free discovery call to discuss OKR-driven development for your business