6 signals that separate real AI potential from wishful thinking - before you choose any technology.
AI is everywhere. Or at least, the conversation about AI is. But in most organisations we work with, the same question keeps surfacing: where do we actually start?
The honest answer: not with technology. Not with tools. With a rigorous look at your business processes. Because automating a broken process doesn't fix it - it just makes it break faster.
At ngage, we've built a practical checklist of 6 signals that help assess the AI potential of a business process. The key word: 'process'. Not solution. No technology is assumed, no platform is implied. This is a diagnostic of the work itself.
If you want to understand why clean process foundations matter before any AI deployment, our article What comes before AI: the case for Process Management covers that in depth. And once you've mapped your processes, The Power of Process Mapping explains how to turn that mapping into a real competitive edge.
The 6 signals of a high-potential AI process
These criteria don't presuppose any technology. They question the process itself - its nature, its weaknesses, its context. That's deliberate. The right diagnosis always comes before the right remedy.
1. Strategic importance
Does this process directly contribute or is linked to a key organisational priority - customer satisfaction, profitability, compliance, speed to market? A high AI-potential process must first be a process that matters. Automating the peripheral is investing for marginal gain.
2. Signs of (poor) performance
Are there concrete indicators (audits, observations, customer surveys, internal reports, etc.) of underperformance: abnormal delays, high error rates, recurring complaints, frequent rework? Poor performance is both a warning sign and an opportunity. Where things grind, AI can flow.
3. Data availability
Does the process generate structured, accessible, good-quality data? No data = no AI. But quality beats quantity every time. A process with few clean data points beats one drowning in noise.
4. Volume of occurrence
How often is this process performed within the organisation? Volume is a key indicator of automation value. High-frequency processes are typically the strongest candidates: repetitive manual effort adds up quickly - and so do the efficiency gains once it's automated.
5. Repeatability and stability
Do the process steps follow a predictable pattern? Are the business rules relatively stable over time? A highly variable or context-dependent process will resist automation - or will require a much more sophisticated approach.
6. Decision complexity
What is the nature of the human decisions involved? Are they genuinely irreplaceable expert judgements - or implicit rules, rules of thumb, and data-driven calls that a machine could learn? The biggest AI opportunities lie in processes where decisions follow recognisable patterns and can be traced back to data, clear criteria, and past outcomes.
This sixth signal is the most discriminating. It separates what belongs to human expertise from what is actually a hidden rule that AI can surface and apply. This assessment can be made for an entire process or for specific steps within it.
How to use this checklist
The point isn't to tick boxes. It's to create a conversation. The real value of this framework is what it provokes: a structured dialogue between operational teams, business stakeholders, and decision-makers - in the same room, looking at the same processes.
In practice, we suggest scoring each candidate process across the 6 dimensions - say, from 1 to 3 - and building a prioritisation matrix. Processes that combine high scores on strategic importance, volume, and moderate decision complexity are typically the strongest quick wins for a first AI initiative.
And here's what consistently surprises teams when they go through this exercise: some processes they were convinced were obvious automation targets turn out not to be. And others, ignored until now, reveal unexpected potential.
That's the diagnostic at work.
What this approach changes
Too many organisations approach AI from the solution side: they pick a tool, then search for a use case, then wonder why results disappoint. Our conviction runs the other way: start from the process, not the technology.
This checklist is a co-creation tool. It involves the people who live the process every day - not just IT or transformation teams. That cross-functional lens is what surfaces the real friction, the real data availability, the real decision complexity.
And sometimes, the conclusion of this exercise isn't 'let's deploy AI'. It's 'let's simplify this process first'. That's not a failure. That's exactly the right diagnosis.
Because AI doesn't fix broken processes. It accelerates them. The groundwork is what makes the difference.
Ready to evaluate your processes?
We offer a free 2-hour process evaluation workshop. Together, we map your AI potential - no commitment, no jargon, no vendor agenda. Just an honest look at where AI can genuinely move the needle for you.
Reach out to our Operational Excellence team to book your session:
info@ngageconsulting.com | Contact us
Learn more
● Process foundations: What comes before AI: the case for Process Management
● Mapping methodology: The Power of Process Mapping
● Cost lens: How to optimize costs without cutting people