S45’s Aman Singh said the startup hires only when a role cannot be done by AI, with most of its workforce split between engineers building agents and AI agent bankers operating them.
On the other hand, INDmoney said it is still too early to set a fixed AI budget, though AI spend is unlikely to exceed 10% of tech expenditure due to rapidly changing model economics.
Fellow panellist Open CEO Anish Achuthan said AI usage will keep rising, but falling token prices are reducing costs and improving the overall economics of adoption.
Artificial Intelligence has changed the way companies manage their workforce. Apart from shaving off the headcount, companies are also reassessing how they hire.
Case in point is AI-native investment platform S45
, whose cofounder Aman Singh said that they hire only if a piece of work can not be done by AI.
Speaking at Inc42’s AI Summit 2026, Singh revealed that the startup’s workforce is three-tiered, where the first layer is engineers who build the agents. The second is AI agent bankers, people whose role is to work alongside those agents, run them, and take ownership of their outputs. While these two layers constitute 80% of the workforce, the remaining share comes from the distribution and sales teams.
Singh was speaking in a panel discussion also featuring Anish Achuthan, cofounder and CEO, Open Financial Technologies, and Kausal Malladi, CTO investments, INDmoney, in a session moderated by Kshitij Shah, entrepreneur in residence at Digio.
Even with the team becoming more AI-native, Singh said that the total AI spend at S45 is still largely overshadowed by people costs. He noted that engineering makes up nearly half of the company’s technology expenditure, and the focus is on giving teams unrestricted access to the best tools available. According to him, engineers are free to use any model, framework, or developer tool that helps them move faster.
Weighing in on the topic of AI budgetary allocation, INDmoney’s Malladi said that while it is very early for his company to allocate a fixed AI budget, it is unlikely that it would exceed 10% of the technology spend, preferring to stay flexible rather than lock into a fixed number. He owed it to the volatility factor that a model his team used a few months ago worked extremely well and was cost-effective at the time, but usage patterns changed quickly, token consumption rose, and costs increased sharply.
Open Financial Technologies’ Achuthan added to the conversation by saying technology costs are actually declining over time. He pointed to falling token prices and said the same AI budget now goes much further than before. “AI usage will continue to increase,” he said, “but the cost per unit is coming down, which is improving the overall value equation.”
Source: Inc42 - Startups
