The most common ask I get from operations leads is "we know we should be automating things — where do we start?" The honest answer is "smaller than you think, sooner than you think." Here's the version of the answer that fits in an essay.
The first move isn't picking a tool. It's listing every task that someone on your team does more than ten times a week. That list is usually 30–50 items and writing it down takes about an afternoon. You will recognise most of them as boring; that's the point.
From that list, score each one on three columns: how much time it costs per week, how much variance it has (does it always look the same?), and how much judgement it requires. The candidates worth automating are high-time, low-variance, low-judgement. Email forwarding and expense categorisation rank top; legal review never makes it.
Start with the boring ones. The first automation I build for a new client is almost always something nobody is excited about — usually invoice routing or a calendar-to-CRM sync. The reason: boring automations have clean success metrics. Either the email gets routed to the right person or it doesn't. You build trust before you spend trust.
Use existing tools first. Zapier, Make and n8n cover most of the high-volume, low-judgement workflows for a fraction of what custom code costs. The threshold for moving to custom is when an off-the-shelf tool stops fitting — usually because the workflow has too many branches, or it touches data the SaaS provider can't see.
Where to layer in AI: the moment in any workflow where someone has to read something and decide. "Read this email and decide which team it goes to." "Read this invoice and decide what category it is." "Read this support ticket and decide if it's urgent." Models are good enough now that a small custom build on top of a workflow tool can collapse minutes into seconds.
Where I'd urge caution: anything that affects money, contracts, or customers without a human in the loop. Even when the model is right 99% of the time, the 1% creates work that erases the savings. Keep approval gates on the edges; let automation handle the volume in the middle.
The metric to watch: not "tasks automated" but "hours given back." A team that's automated forty things and saved zero hours has automated the wrong things. A team that's automated three things and recovered half a day a week has done it right.
The hardest part is stopping. Once a team starts automating, the temptation is to automate everything. Most processes have a layer that's worth touching, and a layer underneath it that isn't. Knowing where to stop is the senior judgement, and it's most of what I bring to a build engagement.
If you want the version of this conversation that's specific to your business, the scoping call is the right place for it. We'll talk through your highest-volume workflows and decide whether one of them is worth building.