
I am a big 70% planning, 30% implementation individual. Always getting things in order and feeling really organised before something goes ahead. It feels built into my DNA, or just a core part of who I am.
On Sundays I have a cook-up to help the week feel prepared and smooth. Before the baking begins I set up the bench with the ingredients, measuring cups, oven heating up and all things ready to go.
My kids think it’s fussy, but to me it’s a creative process and usually about 3 recipes at once buzzing about the kitchen so I want it with some structure.
I’ve always known the prep isn’t the baking. A perfectly set-up bench has never once made a good cake on its own. It just means that when I get to the bit that matters (the feel of the mixture, the judgement about whether it needs another five minutes) I’ve got the headspace for it. I’m in my baking zone!
At work I’ve been building a different kind of bench.
As a Change Manager, the prep can eat into your week
Every change I’ve kicked off in the last fifteen-odd years has started roughly the same way.
Who’s affected?
What’s actually changing for them?
Why now?
What happens if we don’t?
Who needs to say yes?
Then the common templates: a change overview on a page, an impact assessment, a first stab at the stakeholder picture.
It’s important work. It’s also, if I’m truthful, mostly the same questions, asked of different people, typed into the same templates. And by the time it tends to be done, the energy you have for the actual change, the people, the conversations, the reading of the room, can be noticeably lower.
So I built a skill
I wrote recently about my confusion with the word “Skills” in AI, so a quick recap.
A skill is a reusable set of instructions that shapes how AI handles a particular task. Not a clever prompt you save somewhere. More like writing down how you do something, properly, so it can be done that way every time.
The first one I built was for change kick-off.
It runs the intake interview the way I would. It asks the questions in an order that makes sense and follows up when an answer is vague. Then it drafts a change overview and a lean impact assessment straight into my templates. Not perfect but generates a solid first draft that I then make mine.
Once I tested this out and built something I could work with, I built one for learning.
It takes a course idea, a script or a transcript and works through our design brief template with learning outcomes and designing a blended learning plan. I think built a second Skill that outlines the course structure and lesson breakdown from the initial ideation. I then go through the detail, review how it took our ideas and high level goals and suggested we break it down further. We also look thoroughly through scripts to tailor so the language suits our team or the learners.
These skills have become my bench. Ready for use at any stage across my week when I want a helping hand.
What worked
The first draft stopped me being the bottleneck.
The bigger surprise was consistency. When the method is written down, it’s the same method every time. As a human, some days I can be tired or distracted. It never it. When I am having an ‘off’ day, it still asks the question I’d have forgotten. I feel like I have cloned myself into a bot in my computer that completes some of my work with the same level of quality and detail every time.
And writing the skill made me better at the job. To explain to AI how I run a kick-off, I had to articulate things I’d been doing on instinct for years. That alone was worth it.
This work wasn’t to substitute me. It wasn’t to outsource hard thinking about change or avoid work I don’t want to do. It was to follow my steps and details to generate content for me to then review and enhance. It is supplementing me, and saving time so I can do other things and avoid hours on the initial change thinking or generating slides.
What didn’t work
This is the bit I want to be upfront about.
- It’s confident when it shouldn’t be. If an answer was thin, the draft could still sound authoritative. A well-formatted guess is still a guess. But there is often initial ambiguity or gaps in a program of work, and these shouldn’t be filled in by guesses, something time is needed to stakeholders to come to the table or additional detail gathered.
- It doesn’t know the history. It can’t know that this team has been through three restructures, or that the sponsor says yes in meetings and no in corridors. It is an initial take, without context and shouldn’t be the final version.
- The first versions were too long. Thorough is not the same as useful. I had to teach it what “lean” means to me. Once I asked it to generate some learning structure for a single transcript, from a single lesson, as a test. It created a 9 page blending learning outline. Yikes!
- People wanted to skip the interview. Which defeats the point. The questions are the value, not just the document. Again, it is to supplement, not the solution. I like having it to help general detail to kick my brain into gear and then progress.
What I still do myself
The bench handles the prep. I still do the baking.
It is always important to read the room. At times you notice who goes quiet when the change is mentioned. I decide things like how fast to go, and when to stop and have the harder conversation. I sit with the sponsor who is worried and doesn’t want to say so. I workshop with the team what is the technology, what is the learning challenge and how do we want our output to feel.
That’s the part I got into change management for, and later fell in love with learning design for. And it turns out I now have a bit more room for it.
This is why I don’t really buy the “AI is coming for your job” framing, at least not for this work. It’s coming for the bits of the job that were quietly crowding out the job.
If you want to build your own bench
You don’t need to be technical. You need to know your craft well enough to write it down.
- Pick something you repeat. A task you do the same way, for different people, often.
- Write down how you actually do it. The questions, the order, the follow-ups. The things you check for.
- Give it your templates and your quality bar. What does good look like? What would make you wince?
- Test it on a real one. Then fix what you’d have done differently. Then do it again.
- Share it. A skill your whole team uses is worth far more than one sitting in your own folder.
The bench
I’ll always set up the bench before I bake, it’s part of who I am and my baking day process. I just don’t confuse it with baking.
The butter’s out. The tins are lined. Now I get to do the bit I’m actually good at.