I have a very clever niece, Roxanne Cwikowski, Product Specialist @ Zest for Freshservice ITSM. Recently she wrote a sharp and insightful newsletter article titled “The magic is in the work you’re avoiding.”
In this refreshing and blunt article, she captures a truth that many leaders are about to confront as they rush toward AI adoption. She writes:
“We’d quite like to look at twenty years of organisational chaos and somehow infer what we meant to be doing. And preferably have it ready by Tuesday. There’s just one problem. AI needs something to work with.”
This line stopped me in my tracks because it perfectly summarises a challenge I’ve seen across countless organisations in my change and project management work. AI is not the problem. The problem is the missing organisational foundations leaders have tolerated for years.
The uncomfortable truth: most organisations don’t actually know how they work
When organisations decide to change a process, upgrade a system, or restructure a team, the first step is always the same: analyse the current state.
But here’s the recurring issue: there is often no current state to analyse.
Yes, corporate policies may exist. But the translation of those policies into:
- procedures,
- workflows,
- agreed practices,
- and documented knowledge
is frequently missing, outdated, or contradictory.
I once worked with three separate agencies merged into a quasi‑entity to accelerate major infrastructure projects. Their biggest challenge wasn’t designing the future state, it was figuring out the present. Procedures were outdated or non-existent. Workflows differed between staff sitting right next to each other. Business analysts and I uncovered:
- conflicting practices within the same team,
- references to national frameworks long since replaced,
- legal citations that hadn’t been valid for years,
- and staff who were genuinely shocked to learn that colleagues had been doing things “wrong” for years.
This wasn’t incompetence. It was the natural consequence of undocumented work.
Now imagine that group trying to implement AI.
What AI would inherit from most organisations
As Roxanne points out, the reality in many workplaces looks like this:
- Poor‑quality data
- Out‑of‑date knowledge
- Processes that exist mostly in people’s heads
- Policies nobody has looked at since 2019
- SOPs that don’t exist
- Five different versions of the truth
- No clearly defined service ownership
- Dozens of undocumented exceptions that “everyone just knows about”
And then we place AI on top of it and say:
Go on then. Be intelligent.
AI cannot infer clarity from chaos. It cannot automate what isn’t documented. It cannot improve what isn’t defined. It cannot optimise what isn’t agreed.
AI will expose what change managers have been saying for decades
AI promises automation, productivity, and efficiency but it cannot compensate for organisational neglect.
For years, leaders have failed to ensure the basic governance that makes change possible:
- policies reviewed regularly,
- procedures documented and followed,
- processes mapped and maintained,
- workflows agreed and updated,
- knowledge stored somewhere other than people’s memories.
Instead, staff create their own methods. New employees don’t learn from a system; they learn from Sally the Supervisor and her way of doing things.
Ask any business analyst who has tried to reverse‑engineer an organisation’s operations. The consultation meetings are always the same: staff stunned to discover that a colleague has been doing something differently, sometimes for years, simply because nothing was ever documented, reviewed, or governed.
Yet leaders still expect rapid change. They want new systems, new processes, new behaviours without acknowledging that the foundation they failed to maintain is the reason change is slow, painful, and inconsistent.
Leaders aren’t steering the organisation — they’re flying blind
In many organisations:
- governance systems have no review dates,
- staff operate with near‑total autonomy,
- reporting doesn’t baseline or track process performance,
- issues are only discovered when something breaks,
- and change is only initiated when failure becomes visible.
When failure occurs, consultants are brought in to find the culprits. But the failures didn’t start with staff. They started with leadership. They started with the absence of expectations, structure, and oversight.
Leaders often want change to be fast but they haven’t built the conditions that make fast change possible.
AI is the wake‑up call organisations needed
AI will challenge most organisations — but it’s a healthy challenge.
It forces leaders to confront the organisational deficiencies they’ve tolerated for years. AI requires:
- clean data,
- current knowledge,
- documented processes,
- clear ownership,
- consistent workflows,
- governance that is real, not theoretical.
AI is not a magic wand. It is a mirror.
It will reflect back the quality of your organisational foundations or the lack of them.
But if organisations finally do the work they’ve been avoiding, AI can deliver what leaders have always wanted:
- accurate reporting,
- real performance insights,
- predictable workflows,
- and incremental improvements staff can manage when they are proactively and positively led.
AI won’t fix organisational chaos. But it can help leaders finally clean it up. And AI can make faster change possible, particularly when everyone knows what everyone else is doing!