We didn't start with AI. We started with people.

FirstPoint began in the Philippines in 2014, as a BPO contact centre. Not a tech startup. Not a product company. A contact centre, with agents on phones, handling real customer interactions for real businesses. Inbound calls, outbound campaigns, sales, support, escalations, disputes. The full stack of what customer service actually looks like at scale.

That decade of operation built something that most AI companies don't have and can't buy: a deep, ground-level understanding of how customer conversations work. The psychology of a caller who's frustrated. The moment a sale is about to be lost. The difference between a policy answer and a satisfying resolution. The patterns that repeat, industry to industry, call type to call type.

That knowledge became the foundation for everything that followed.

In 2019, the team started building internal AI tools. Not because the market told them to. Not because investors were asking for it. Because the people running the contact centre operations could see clearly where things were heading, and they understood the gap between what off-the-shelf automation could do and what contact centre work actually required.

The early builds were logic-based. Decision trees and conditional routing designed to help agents handle more calls, make faster decisions, and deliver more consistent outcomes. Simple by today's standards. But they were built on a specific operational understanding: what the calls looked like, where agents succeeded, where the system was letting them down.

The tools were built from the inside out. That distinction matters.

Then COVID hit. The staffing disruptions of 2020 broke contact centre operations globally. Businesses that ran on headcount scrambled. Competitors fell behind. Clients were left without coverage.

FirstPoint didn't scramble. The automation groundwork was already in place. The operation absorbed the pressure that others couldn't, and the team kept building. Not despite the disruption, but because of it. Real operational pressure, applied to real tools, made them sharper.

What started as internal efficiency tooling became something more significant: a platform that had been stress-tested under conditions nobody planned for, and had held.

In 2021, the platform went live in its first commercial deployments. Locksmith operators. Home services businesses. Real clients, taking real jobs, dispatching them automatically through an AI system handling the customer interaction end to end. No human in the loop for those calls unless the situation required it.

This matters, because it means FirstPoint AI has been running live commercial deployments since 2021. Not pilots. Not proofs of concept. Live operations, with real customer interactions, real outcomes, and real accountability.

That track record is the foundation everything since has been built on.

As open-source large language models became viable for fine-tuning, the team moved quickly. The difference between FirstPoint and the wave of companies that appeared in 2023 bolting a general-purpose LLM onto a phone number is what sits underneath. Years of live deployment data. Real contact centre operational knowledge. The understanding of what a good call outcome actually looks like, not just a completed interaction, but a satisfied customer and a measurable business result.

That knowledge was fine-tuned into proprietary model weights. The platform moved from logic-based automation into genuine conversational AI, capable of handling complex, context-sensitive customer interactions that general-purpose models weren't built for.

By the end of 2023, FirstPoint had something that couldn't be easily replicated: a Voice AI platform backed by years of real deployment history and trained on operational knowledge that only comes from having run the contact centre yourself.

In early 2024, the first casino clients came onboard. Casinos are demanding. High-volume inbound calls. Complex event and reservation management. VIP guest relationships that require consistency and care. Player activation and re-engagement campaigns that need to feel personal, not automated. It's an environment where a bad AI implementation doesn't just fail. It damages the guest relationship and, in a regulated industry, has real business consequences.

The platform was ready for it. Not because it was new technology, because it had three years of live commercial deployments behind it, running in environments that required it to work every time.

By the end of 2024, a meaningful roster of US casino operators was live on the platform.

As of 2026, FirstPoint AI operates a US-headquartered Voice AI platform serving more than 30 casino operators, hotel groups, and service businesses across the United States and internationally.

The product suite covers inbound AI, outbound AI, answering AI, live-managed web chat, and a self-serve platform for businesses that want to build their own. A dedicated contact centre operations team provides genuine human escalation. Live agents, trained in service psychology, available when the call needs a human rather than an AI.

The ecosystem has grown beyond the core Voice AI platform. RediBookings brings channel management and reservations technology to hospitality clients. RediRSVP handles event guest management for casino operators. InviteIQ brings smart invitation and event communications to a broader audience.

We are still building. The core belief hasn't changed: that AI built by people who understand the operational problem is fundamentally different from AI built by people who don't.

Twelve years in. Still building from the inside out.