AppliedAI expands into APAC, launches Opus 2.0

The expansion includes new offices in Singapore, Kuala Lumpur and Hong Kong follows a pre-Series B funding round backed by Mubadala and Arbor Ventures.

Applied AI has expanded in the APAC region with new offices in Singapore, Malaysia, and Hong Kong. The “world’s most boring AI company” has also launched Opus 2.0, its platform for deterministic orchestration of frontier AI agents in regulated industries.

With data sovereignty and regulatory requirements overlooking the adoption of AI agents in the region, Applied AI believes this puts a constraint on AI projects with a majority of pilots failing to reach production. Specifically, general-purpose agents operating on open-ended tasks cannot satisfy the auditability, reproducibility, and separation-of-duty requirements that regulators demand.

This is where Open 2.0 can help enterprises in their AI journey as it decomposes every enterprise process into a directed acyclic graph (DAG), where each node is a schema-enforced harness binding a frontier agent to hard-typed input and output contracts.

According to AppliedAI, every step is provably bounded, auditable, and reproducible. There is a built-in Maker-Checker-Arbitrator governance model enforces separation of duty at every agent boundary, while a governed fine-tuning pipeline allows individual harnesses to improve in deployment without compromising process integrity.

"We think we can really increase productivity in these industries. So what we've seen in other markets like the UK and the US is we can increase productivity per person by at least 10x — so it's an order of magnitude change. And what's most interesting is that is not at the expense of quality. So even though we see the increase of that 10x, the accuracy also increases by around 20%, so you have less errors,” commented Arya H. Bolurfrushan, Founder and CEO of AppliedAI

He also recommended that organizations start with the back office on things that are operationally complex and regulated as AI is very good at that, because it can read all the laws and make sure an organization is fully compliant.

"The future is already here. It's just not evenly distributed. So it's an issue of adoption leadership. The tech is very, very powerful. It works, and it's just about helping get it into these different industries. If you have the confidence as a leader to say we're going to adopt AI, we can actually have it... we have the advantage of being able to move very quickly,” Bolurfrushan added.

Addressing data sovereignty requirements

Given the increasing demand for data sovereignty in AI development and deployments, Bolurfrushan shared that AppliedAI is deployed locally on the instance of the customer.

“So, whether it's their cloud — we also have deployments where we're air-gapped on premises where it's fully, fully secure. The problem with LLMs is you give them a lot of your personal data, and you don't know where it goes. Opus is trained to not do that. It stays locally,” Bolurfrushan said.

Bolurfrushan also pointed out that the whole Opus cluster is deployable on premises or even in a gap environment with one big cluster, which is really differentiated, to be able to take it offline and take it into a very secure environment

"If you're already on Azure or AWS or Google or Alibaba — all these clouds — then you can have Opus deployed in your instance so it never leaves the furniture of your company. If it's air-gapped and on-prem, then you can actually deploy it in your existing server, because it's in Kubernetes containers, it's very easy to deploy. But the idea is that nothing leaves and it's fully secure,” he added.

AppliedAI charges per AI man-hour, treating harnessed agent execution as a scalable labor outcome rather than fixed software. Where legacy SaaS vendors charge for seats and tokens, AppliedAI charges for work completed.

Opus 2.0 will be generally available to all Opus regions on 30th April 2026, including the US, EU and UAE. The expansion into the region follows a pre-Series B funding round backed by Mubadala and Arbor Ventures.