“Do You Invest in AI Companies?”
Very often I get asked whether we at Genesia invest in AI companies, or if it’s an area of focus for us. It would be strange to say no considering what AI can help us achieve, but my response is more nuanced than that.
Building a successful business starts with building a solution that provides value for the target users, not whether the business uses a specific technology. What I look for is whether the founder can articulate specifically what AI improves for the business and the customer — such as speed, cost, accuracy, accessibility — rather than describe what the AI does technically.
Most companies implementing AI fall into one of three categories. The signs below are not exhaustive, but they are the ones I notice most often.
1. Commodity
This is AI that any competitor can replicate in days, or even hours, and that doesn’t meaningfully change the unit economics, the operations, or what customers pay for. Value may technically exist, but it is commodity value.
When the pitch revolves around the technology rather than the solution, and the business would function largely the same without AI, there’s a good chance the AI is only providing commodity value.
Other signs:
- The AI features don’t connect to the customer’s stated pain
- The product is a thin layer over an API call
- The roadmap is “add more AI features” instead of “solve the next customer problem”
For example, a retailer adds an AI chatbot to handle customer queries. It works, customers use it occasionally, and every competitor ships the same thing within a quarter, but the customer still chooses based on price, delivery speed, and product range — nothing has changed.
2. Enhancement
This is AI that meaningfully improves the core solution or the business operations — faster, cheaper, more accurate — in a way the team or customer notices, and that shifts the economics.
Competitors can arguably offer the same thing, which is where other factors matter. Proprietary data feeding a feedback loop is one way an enhancement becomes durable rather than replicable.
Signs:
- The customer experience or internal operations are meaningfully worse without AI
- Margins are meaningfully worse without AI
For example, a logistics company that used to route deliveries manually now does it algorithmically. It offers the same service to the same customers, but with a higher fulfillment rate, a lower cost per delivery, and the ability to predict workforce availability. The business existed before and works the same way, but now it works better.
3. Unlock
These are solutions that were economically or technically impossible before AI.
Signs:
- Serving the lower end of the market when it never made economic sense before
- The cost structure that made it irrational has collapsed
- The solution was previously possible but only for those who could afford human labor
Take per-client customization. Building bespoke features for individual customers used to be economically irrational because of the development hours involved, so software companies standardized and asked clients to adapt. Now that building a new feature takes hours, serving one client’s specific needs is worth doing, and a company can offer something closer to bespoke at close to standard pricing.
Using AI to unlock new opportunities has potential for higher returns, but it carries the risk of doing something that has not been done before: spending resources to experiment, getting things right, and educating the market. Unlocking something new does not guarantee success. It might give you first-mover advantage, but you still have to make the business model work.
What this means in Indonesia
Most of the companies I meet in Southeast Asia and Indonesia fall under Commodity and Enhancement. We are unlikely to invest in the Commodity category given how easily those businesses can be displaced, so Enhancement and Unlock are where we look for opportunities.
That said, in a market like Indonesia where there are still many unsolved inefficiencies, solving those problems is the priority, and AI should enhance the solution when it makes sense rather than because the technology is available. Chasing Unlock for its own sake is the same error as adding AI for its own sake. This is also where the AI Adoption Gap applies: an Unlock built for a market that isn’t ready for it will not get adopted.
So the question worth asking is how technical your users need to be before they get value from what you’ve built.
Instead of asking “should I use AI when everyone else seems to be using it?” ask “how can I make my solution better?” and “what does AI unlock for my customer that wasn’t possible before?” If the answer is AI, then use AI.
Using AI doesn’t make a company investable. What matters is what it makes possible for the company and customer.