AI can write your policy. Good.
AI and automation are changing the landscape of professional work, and quickly. Policies, procedures, job descriptions, strategies, frameworks, and reports can already be produced faster and more cheaply than they could a few years ago, and that is only going to continue.
That’s a good thing, if you ask us.
If AI can produce a useful first draft of a policy in minutes, that’s a win. Take it. If it can summarise information, structure a procedure, compare documents, or take some of the administrative load out of running a business, even better.
Using AI can take some of the grunt work out of the process. However, that’s not the end of it. The high-value part of creating better business is not in writing policies; it’s in translating them into reality.
The document was never the hard part
Anyone who has worked inside organisations for long enough knows that having a policy and having a policy that works are two very different things. The same goes for strategies, procedures, risk frameworks, organisational structures, and most of the other systems that businesses rely on.
The difficult part comes in creating a policy or process that looks good on paper AND fits the organisation and works in practice. Before turning to AI for help with the wording, there are some important questions to answer:
- Does this reflect how work actually happens?
- Is it appropriate for the size, complexity and risk of this organisation?
- Who needs to do something differently as a result?
- Do they have the capability, authority, time and resources to do it?
- What other processes, roles, systems or documents does it affect?
- How will we know whether it is actually working?
A new policy rarely exists in isolation. It might affect responsibilities, delegations, training, systems, forms, reporting, leadership expectations, budgets, communications or measures of performance. Change one part of an organisation and there are often ripples elsewhere.
That translation and integration is where much of the real work sits.
Risk lies in the gap between polish and practice
As AI makes it easier to create polished organisational documents, there is a risk that organisations accumulate more apparent sophistication without backing that up by becoming more capable.
To put that more simply: documents can make you look like you know what you’re doing, but it’s not good to be all talk and no action. If you have posh documents that aren’t making a difference to what’s happening on the shop floor, you're no better off.
Policy doesn’t always change practice, training might not build capability, strategy might not be influencing decisions. A procedure that’s not properly thought through, by a real human with skin in the game, can easily become one that people routinely work around because it doesn't reflect the job.
Over time, the gap between what is written and what actually happens can create duplication, confusion, and extra administrative burden. People might develop workarounds, and when these are left unofficial, different parts of the organisation start doing the same thing differently. Leaders can believe an issue has been addressed because a document exists, while the people doing the work know very little has actually changed. This misalignment is a very real safety risk.
It can also create legal liability. Organisations make commitments through their policies and systems, and use them to demonstrate adherence to the law. If you say something “will”, “always” or “must” happen, there needs to be a reasonable basis for expecting that it does. When a Board, customer, auditor, regulator, or court asks what actually happened following an incident, the difference between what an organisation says it does and what it can demonstrate will be very important.
Making it fit is the tricky bit
Using AI for more of the drafting work is great. Writing every word for another 20-page document isn't necessarily the best use of our expertise or our client's budget.
What it can do is free us up to spend that time at the pointy end. We can put our efforts towards understanding what an organisation is trying to achieve, working through what good looks like, testing assumptions, considering the evidence and different perspectives, and translating all of that into something that will actually work.
A big part of our work happens in investing time outside individual client engagements. We participate in standards development, industry and professional groups, research, and ongoing professional development. We stay connected with changes in regulation, organisational practice, and the way work itself is evolving. While AI drafts up fairly generic policies, we make sure we have our finger on the pulse.
That broader engagement matters because organisations rarely experience issues that are easily predictable or that fit into predetermined categories. Something that starts as a health and safety requirement might have implications for workforce capability, operational processes, leadership, technology, or governance. A problem of compliance will likely become an implementation concern.
It’s important to be across all of it!
Our job is to bring together the right knowledge, evidence, and perspectives at the point they are most useful, and help translate them into the organisation concerned. It’s critical thinking.
Depending on the issue, that might mean considering:
- legal, regulatory, contractual and standards requirements;
- current evidence, research and good practice;
- how work actually happens, rather than how it is assumed to happen;
- what already exists and works well;
- the perspectives of the people who use or are affected by the system;
risks, dependencies and unintended consequences; - and what implementation, communication, capability, and evidence will be needed afterwards.
There is rarely one perfect policy, process, or system that can simply be lifted from another place - or from a generative AI tool - and dropped into an organisation. Good practice still has to be drawn up and guided through the lens of the organisation expected to use it.
The information and even the drafting work is, increasingly, just a click of a button away. Knowing what matters, how it applies, and how to make it work is a different capability.
Judgement is needed when the script runs out
As more routine and repeatable work can be supported or performed by technology, the most valuable human capability sits in the moments that aren't easily scripted. Anyone who has ever been in a workplace knows that “not easily scripted” covers a lot of what goes on.
Policies, procedures, systems, and AI can provide guidance for situations we can anticipate. But circumstances change. Information is incomplete, priorities conflict, and then something will happen that nobody thought to include in the procedure.
To find a resolution in those moments requires people who understand what the organisation is trying to achieve, recognise when an AI-generated or scripted approach doesn't fit, exercise sound judgement within appropriate boundaries, and know when to act, escalate, or challenge.
Organisations have to think about more than whether people can simply follow a process. Do they understand the purpose behind it? Do they have the capability and authority to make decisions? Are accountabilities and boundaries clear? Can they question something that doesn't make sense? And will the organisation learn from what happens next?
As technology becomes better at supporting the scripted parts of work, organisational capability will increasingly be revealed in what happens when the script runs out.
We’ve seen it, and we know how to help
Our work gives us an interesting view of the gap between intention and reality from both directions (before and after something has gone wrong).
We see policies that look good but haven't translated into practice, and controls that made sense when they were designed but don't fit the way work is actually done. We’ve witnessed training that didn't build the necessary capabilities, and systems that work well independently but make no sense when connected to each other.
We also see the opposite: organisations with good practices, capable people, and practical ways of getting things done that aren't particularly visible in their formal systems. In these cases, improvement isn't about introducing something new but about understanding what already works, strengthening it, and making it clear, visible, and able to be referenced.
Our approach to supporting clients, and our value to them as consultants, is not in providing documents, but in understanding what is happening, where issues are cropping up, and how to make things work.
As is so often the case, building strong systems within a business is about capability uplift. We don’t want to write documents for you. AI can do that, to a draft level at least. Our intention is to help you build up real capability and oversight within your organisation, adding that high-value human judgement to the equation.