Imagine a support manager at 3 AM, staring at a Slack channel where the “Urgent” tag is being applied to 400 simultaneous WhatsApp threads. They have a team of ten, but the bottleneck isn’t the number of humans available to type—it’s the sheer volume of noise. They are paying for ten expensive SaaS licenses, yet the work is piling up because the software is just a window, not a worker.
This is the friction Respond.io is betting on. The Malaysian startup just raised $62.5M to scale its AI agent-powered messaging app, and while the funding is the headline, the pricing model is the actual story.
Most CX software is still stuck in the 2010s, charging per seat. It is a model that makes perfect sense when you are selling a tool for a human to use. But when you introduce AI agents that can handle a thousand conversations at once, the “per seat” logic collapses. Why are we still pretending that a headcount is a proxy for value?
Respond.io charges per convo, not per seat. It is the difference between paying for a monthly gym membership you never use and paying a taxi meter for the actual distance traveled. If the AI does the work of ten people, the company doesn’t need to buy ten more licenses; they just pay for the volume of work completed.
The per-seat model is a relic.
(I suspect some VCs hate this because it’s harder to project linear growth, but that’s a “them” problem). In an AI-first world, the traditional SaaS seat is a liability. It creates a perverse incentive where vendors are rewarded for more people using the tool, rather than the tool actually solving the problem. If an agent can resolve 80% of tickets without a human ever touching the keyboard, the vendor shouldn’t be penalized with a drop in MRR.
Of course, the reality of this is messy. You still have to deal with the underlying API costs and the inevitable latency spikes that happen when a model decides to hallucinate a very long poem instead of answering a shipping question (and the token costs are still a headache). But moving the cost to the outcome—the conversation—is the only way this math works.
The $62.5M isn’t just for polishing the UI. The company explicitly eyes acquisitions in North America and Europe. This is a bold move for a Malaysian firm, but it makes sense if you look at the fragmentation of the messaging market.
Between WhatsApp, WeChat, Line, and the endless sprawl of SMS, the “unified inbox” is a graveyard of failed promises. Most platforms claim to unify the experience, but they usually just slap a few APIs together and call it a day. To actually win in the West, Respond.io needs more than just a good agent; they need the local infrastructure and the existing client lists that only come with buying other companies.
They are effectively trying to build a global layer for AI-driven commerce. If they can successfully integrate a few North American players who have the distribution but lack the agentic intelligence, they skip five years of organic growth.
By Q4, we’ll see at least two other major CX platforms switch to usage-based pricing to stop the churn. The market is shifting away from “tools for humans” toward “outcomes provided by agents.” Respond.io is just one of the first to put a price tag on the outcome rather than the user. Whether the Western market is ready to ditch the seat-based invoice remains to be seen, but the money is already moving in that direction.
