Overview

Akomi's mission is to automate administrative operations for healthcare providers using intelligent digital staff. Currently, our agents are able to navigate complex Prior Authorization workflows on behalf of providers. Ultimately, Akomi agents will be deployed at every step of the Revenue Cycle Management flow, transforming how healthcare operations are managed. We have raised $5 million in seed capital to launch Akomi.

We're currently hiring Machine Learning and Full-stack engineers to build Akomi. If you're based in Bangalore or San Fransisco, have expertise in either domain, and are excited by the opportunity described below, please reach out!

The Opportunity

At Akomi, you'll be tackling one of the most pressing and high-impact challenges at the frontier of Agentic AI Research.

I. The market is very large, and the problem is clear

In the US, $4.9 trillion was spent on Healthcare in 2023 alone. Fifteen to twenty-five percent of this goes toward administrative tasks. This high overhead increases the total cost of care, leads to clinician burnout, and ultimately makes care inaccessible to many.

II. The technology is exciting and it's only getting better

We believe the current state of technology offers an unprecedented opportunity to tackle these issues and redirect the 15-25% administrative tax to patient-centric activities. The “Agentic Era” is the beginning, and we are strong believers in the fact that a team of highly specialised agents will eventually be able to perform all administrative tasks part of the RCM workflow.

III. There's so much work to be done

Akomi's digital staff already executes full Prior Auth flows autonomously. Here's a high level visual of how our system currently works:

The crux of our system is the Orchestrator. The Orchestrator is built to handle complex, long-running web automation tasks. Here are the core design principles that make this system robust:

• Context Management: Uses contextual prompting and just-in-time prompt injection to ensure only relevant information is used at each step.

• Action Selection: Evaluates different potential actions and chooses the most effective path forward, ensuring continuous progress toward task completion.

• Sensing: Leverages both visual inputs and the web page DOM to accurately interpret the current state of the payer portal.

• Clinical grounding: All decisions taken by Akomi are grounded in the clinical reality of the patient and abide by the rules outlined in insurance policies.

By combining these elements—context handling, informed action selection, and accurate UI interpretation, and clinical grounding—the Orchestrator can reliably navigate and complete intricate, multi-step processes.

The Team

We are a small but skilled team with large amounts of experience in Healthcare, AI, and entrepreneurship.

• [Co-founder currently in stealth mode]. He previously built a healthcare AI startup that was backed by YC and Sequoia India, which is now a market leader in its domain.

• Rahul spent the last 5 years as an ML Engineer at Twitter (X) and then at Reddit, where he built large scale Ads Recommendation systems.

• Sid is a veteran entrepreneur and investor, who previously co-founded and scaled Aloha Technology to 1,000+ employees.

• Dr. Uday and Dr. Raju owners of Remix Medical—a multi-specialty medical center—are the clinical experts of the founding team.

Contact Us

If you're a full-stack or machine learning engineer based in Bangalore or SF and are interested in the problem statement described above, please reach out to us!

Write to us