Automate, assist or own: where AI belongs in a growing business
I use AI every day. It helps me stay organised by structuring my day, summarising lengthy documents, compiling research, and nudging me when something slips through the net. That usually surprises people, as I’m typically now known as the ‘human-first guy’. My book, Human-First Marketing, makes it clear that being human-first isn’t an anti-AI position, and I’m not about to change my mind.
These last few years have been exceptionally challenging for business owners not only to make conscious decisions about how to (or not to) utilise artificial intelligence within their businesses, but to just keep up with the pace of its development. What seemed an impossible use case is old news the following month. To help myself make consistent decisions about its use, I have a mental model I call: automate, assist or own.
I categorise tasks into three buckets:
- Automate The admin that’s dull, repetitive and still important: scheduling, sorting, transcribing, pulling reports together. If a task is the same every time and nobody’s judgement makes it better, I would lean towards automating it.
- Assist The work that becomes a joint effort. I know where I want to get to, and the tool helps me get there a bit quicker. Drafts, research, structuring an argument. My judgement, taste and creativity stay in it. The tool just shortens the route.
- Own Anything that runs on judgement, taste or human connection. I don’t automate LinkedIn messages, because a message from me should actually be from me. I get on the phone whenever I can. Some things are worth protecting precisely because they’re human.
Not having a clear plan on when and when not to use AI can lead to issues.
Faking expertise
We wrongly assume that AI is very good at producing content that looks like it came from an expert. But it isn’t. By its very mechanics, what it’s good at is producing the average. As feeds fill with the same AI-produced posts, what still gets noticed is real people saying things only they could say.
We’ve seen it ourselves: one of Giraffe’s most successful posts was an unpolished TikTok meme about a moment every social media manager knows. It reached 3.5 million views, drawing comments from Sky, Shopify and Zoom. It worked because it came from people who do the job daily. AI can write relatable content, but only guess how that moment feels.
In this next era of marketing, demonstrating expertise won’t be the goal; anyone can do that. It will be about displaying taste. Expertise has been democratised, first by the internet and now to new levels by AI tools. To stand out, you need a unique opinion on the expertise and the ability to
communicate it in a way that cuts through the mountains of slop.
Skipping the challenge
Ask an AI tool whether your idea is good, and it will usually find a way to agree. That isn’t just a hunch; it’s how they are designed. Researchers at Anthropic found that leading AI assistants consistently lean towards matching a user’s views over the truth, partly because agreeable answers get rated more highly in training. Don’t get me wrong, I like being told I’m a genius as much as the next person, but it’s not actually very helpful. Growing a business depends on surrounding yourself with people who will tell you when you’re wrong.
This year, we changed how we sell at my agency. We decided to stop emailing proposals and started presenting them live to the decision-makers. It’s slower, and it takes more of our time. That’s rather the point. A proposal sent by email is a document. A proposal presented is a conversation. You hear the hesitation, answer questions in real time, and continue to build the relationship.
Even if AI can take something off your plate, it doesn’t mean it should. If removing the challenge sacrifices human connection or stunts your creative and intellectual growth, it may deserve to be in the assist or own category.
Blurring who decided
One overlooked risk of automating everything in sight is losing track of who made decisions. When strategies, posts, or recommendations emerge from chains of prompts and edits, responsibility can become unclear. If something goes wrong, nobody owns it. The UK government’s AI white paper calls for “clear lines of accountability” wherever AI is used. Losing accountability and decision-making in a business is a disaster waiting to happen. Our thought leadership work for executives shows how we keep that line clear. There’s a monthly call where we talk through what’s happening in their world and what they really think about it. We write. They review, change and sign off. A tool can help with the writing, but it can’t provide the opinion or stand behind it. The leader does both, and everyone understands that.
That’s the test I’d give any founder. For anything in the “own” box, you should be able to point at the person who made the call. If the honest answer is “well, a bit of me, and also a bit of Claude”, it’s in the wrong box.
The better question
Most conversations about AI in business focus on efficiency: getting things done faster, improving margins, and needing fewer people. Don’t get me wrong, I’m all for efficiency; I just think it’s the least interesting question.
The one I keep asking is different. If AI takes the legwork off my team, what do they do with the time? The answer I want is more of the work that made them want the job in the first place: the ideas, the conversations, the creative calls. And for clients, more at the same price. Better thinking and more time with the people behind their brand, not the same service delivered slightly cheaper by fewer people.
That’s where the three boxes earn their keep: automate the legwork without guilt, and let AI help with work you already know. Put the saved hours where tools can’t follow: on the phone, with clients and candidates, and on calls where someone has to say, “this one’s mine.”
Take a look at your current workload – are you owning the right stuff?
For more startup news, check out the other articles on the website, and subscribe to the magazine for free. Listen to The Cereal Entrepreneur podcast for more interviews with entrepreneurs and big-hitters in the startup ecosystem.
