About a-team Marketing Services
The knowledge platform for the financial technology industry
The knowledge platform for the financial technology industry

A-Team Insight Blogs

ipushpull and Koch Automate OTC Gas Trade Booking with Gen AI Agent

Subscribe to our newsletter

Data platform ipushpull has partnered with Koch Energy Services to launch an LLM-enabled agent designed to automate the entire lifecycle of physical natural gas trades, from chat to booking. The move tackles the long-standing challenge of manual, error-prone workflows in over-the-counter (OTC) markets and signals a major step forward in the application of Generative AI to complex trading environments.

The collaboration targets the US natural gas market, a vast, bilateral space where trades are negotiated and agreed upon almost exclusively within chat platforms like ICE Chat. This environment has traditionally forced traders into a cumbersome process of manually copying and pasting trade details into an Energy Trading Risk Management (ETRM) system, a workflow ripe for inefficiency and costly ‘fat-finger’ errors.

The new solution, which integrates Koch’s DealStream platform into ipushpull’s AI framework, directly addresses this pain point by deploying an intelligent agent that automates the trade capture and booking process. Matthew Cheung, CEO of ipushpull, explains to TradingTech Insight that the partnership combines deep domain expertise with cutting-edge technology. “Essentially, our solution is using our trade capture and booking technology and bringing in Koch’s domain expertise for the US natural gas market, which is a disparate, non-exchange-traded market that has traditionally relied on manual chats and manual keying data into forms.”

At the heart of the solution is an AI-powered bot that can interpret the unique, abbreviated language used by traders. It extracts key details from a conversation and structures the unstructured data into a validated, bookable trade.

Cheung elaborates on the technical mechanics. “We deploy a bot inside ICE Chat that extracts trade details from traders’ conversations,” he says. “Our agentic AI framework then ingests this unstructured chat data, structuring it using an LLM agent that is powered by retrieval-augmented generation (RAG) and vector databases. A key feature is its ability to self-learn; with every edit the system builds a unique profile for each trader, adapting to their individual terminology for instruments, delivery points, and refineries. The final, structured data is presented as a clear deal ticket. The trader simply reviews it and clicks a button to book the trade seamlessly into their ETRM system.”

While the efficiency gains are substantial – Koch reports a 75% reduction in trade input time across more than 35,000 trades – a critical component of the design is ensuring accuracy and maintaining regulatory oversight. In a sector where booking errors can lead to significant risk exposure, the platform is explicitly designed to keep the trader in control.

Cheung stresses that this ‘human-in-the-loop’ approach is fundamental. He says: “A core principle of our design is that while an LLM-enabled agent drives the transformation, a human is always in the loop,” he points out. “This is a necessity from a regulatory standpoint, as the trader is a regulated professional who is ultimately responsible, and our system is built to support their oversight. In practice, this means the system extracts the data, but always presents it to the trader for validation. The trader can then approve or amend the data. This intervention serves two purposes: it guarantees the accuracy of the trade booking and, crucially, provides a feedback loop that continuously refines and improves the agent’s performance.”

For Koch Energy Services, the benefits have been immediate. Darnell Bortz, Director of Natural Gas Trading, noted in the announcement that the tool has “significantly reduced errors and enhanced overall employee experience.”

The partnership not only solves a persistent challenge in the energy markets but also serves as a powerful proof-of-concept for applying advanced AI to other OTC asset classes. As ipushpull expands its presence in North America, the success of its collaboration with a major industry player like Koch demonstrates a clear path toward eliminating manual workflows and unlocking new efficiencies across the trading floor.

Subscribe to our newsletter

Related content

WEBINAR

Recorded Webinar: Navigating the Build vs Buy Dilemma: Cloud Strategies for Accelerating Quantitative Research

For many quantitative trading firms and asset managers, building a self-provisioned historical market data environment remains one of the most time-consuming and resource-intensive steps in establishing a new research capability. Sourcing data, normalising symbologies, handling corporate actions and maintaining infrastructure can take months and absorb significant budget before a single model is tested. At the...

BLOG

Building for the Next Big Event: What Prediction Market Operators Need from Exchange Technology

By Ian Salmon, Head of Product Marketing, Adaptive. Prediction markets have moved from the edges of the financial ecosystem into a space that increasingly resembles regulated market infrastructure. What began as a retail phenomenon around political events and sports outcomes has evolved into a sector attracting institutional capital, established exchanges and serious regulatory attention. The...

EVENT

ExchangeTech Summit London

A-Team Group, organisers of the TradingTech Summits, are pleased to announce the inaugural ExchangeTech Summit London on May 14th 2026. This dedicated forum brings together operators of exchanges, alternative execution venues and digital asset platforms with the ecosystem of vendors driving the future of matching engines, surveillance and market access.

GUIDE

AI in Capital Markets Handbook 2026

AI adoption in capital markets has moved into a more disciplined phase. The priority is now controlled deployment: where AI can be used safely, where it can deliver measurable value, and how outputs can be governed, monitored and evidenced. The 2026 edition of the AI in Capital Markets Handbook examines how AI is being applied...