Last week I sat on a panel at Bali Tech Week. The longer it went on, the more I noticed that everybody talked about AI as a big concept coming and that companies should prepare for it. Very little of it was something an enterprise could actually do next week. Even the consultants, people who do what I do and get paid to make this practical, struggled to explain how a large company gets from the idea to real impact.
After that I spent a few days in Jakarta with several of our enterprise clients. They described the same gap, but from their side. They have given their people Copilot, and here is vibe coding going on. Some people are enthusiastic, and others don’t use it at all. What they asked me for is help to spread it across the company, and help to set up governance around it, which is a very different conversation from the one on stage.
I fully understand the worry about governance. You cannot leave a couple of thousand people in the wild with these tools, because data will leak, security issues will appear, and a lot of tokens will get burned on nothing. If you’re a bank, it’s even more serious, because you answer to the regulator and to your own compliance and risk teams. So the question I keep asking myself is: how do you put AI in everybody’s hands without hurting the company?
AI changes how people work
My impression is that many leaders here see AI as the next step after 30 or 40 years of computers and digitalization, like the ERP, the website, the mobile app, and the data warehouse. All of that was about building things. So the natural next step is to ask which AI app the company should build. Some of that is worth doing. But this wave touches the work people do with their heads, like drafting, checking, and summarizing, which fills most of an office day. So it becomes a question about how a credit analyst or a procurement officer spends her week. I don’t think a new app solves that.
A lot of that work is repetitive work that until recently only a human could do. Much of it can now be automated in a simple way, inside the tools people already use. I don’t see that as a reason to cut jobs. The same person can serve more clients and pick up work they never had time for, and the company grows. I call it putting your people on steroids. It’s a crude phrase, but it’s closer to what I mean than “productivity gains”.
Most companies will do what HR has always done: put everyone in a room for a one-day training and let them figure it out. I’ve been deep into AI for the last few years. It’s a huge field, it changes all the time, and I still keep finding possibilities I hadn’t thought of. Expecting someone in procurement to find the best uses for her job after one workshop is optimistic. To get beyond using Copilot as a better Google, you need some creativity and some imagination about what your work could look like. Most people need help with that.
I learned this over ten years of training teams in Indonesia on agile, mostly through Ekipa and our work with large enterprises in the region. A concept only landed when someone translated it to this team and this piece of work. With AI, you have to do that translation for every role, and the tools keep changing while you do it.
This is the part the speed framing misses, and it is the same ground Agile was built on. Agile came out of software teams that could not freeze their requirements for a year and hope the market waited, so the practice was designed to re-check assumptions often enough to change direction without starting over. That is the position an enterprise adopting AI is in now. The tools change month to month, and what a model can do today is not what it could do two quarters ago, so the teams that keep delivering are the ones built to adjust as conditions change.
What companies here need
When I’m in Europe or the US, the conversation has already moved to agents. People ask what a manager is responsible for when she oversees a hundred agents instead of a hundred people, and how you stop those agents from going rogue. Underneath, it’s the same repetitive office work, which they want agents to take over/automate. In Indonesia, I haven’t heard a single person talk about agents. At a conference in Singapore last year, the tech people were excited, while the business people were asking how to use ChatGPT, and that product is almost four years old now.
IT teams are the exception, and that’s no surprise because technologists love this stuff. They use AI for coding every day, they think hard about keeping the code good when AI wrote part of it, and some already test with agents instead of a testing team. They’ll find their way, as long as someone gives them rules to work productively and safely.
Outside IT, you need leaders who dare to give people room to experiment. Give them Claude, ChatGPT or Copilot, and accept that they will burn tokens while they learn. Some companies will give everyone access at once. Some will start with a small group that goes deep and comes back with examples. Some will set up an enablement team to spread what works. Whichever route you take, even a director with a good rollout plan still needs that plan carried out person by person. Most companies I meet don’t have the people inside to do that.
So I’ve been playing with propositions around exactly this: how do the things that work in Europe and the US translate to companies here? I could be wrong about the best way to help. And maybe I’m overestimating how far ahead the West is, since my view is based on Indonesia and one event in Singapore. But I’m sure this work needs to happen, and in a big company it has to happen with every single person.