The market commentary hidden inside trader chat

Most messages arriving in trader chat carry context on the market being watched: its ebbs and flows. A morning update blasted to every client on a distribution list. A salesperson flagging an unusual flow or a data surprise. A long, detailed answer to a question a portfolio manager asked 20 minutes earlier. Taken together, this traffic carries much of what the street actually knows about a market on any given day.

It also arrives in the least usable form possible. Unstructured prose, spread across dozens of rooms, read once by whoever happened to be at the screen, then lost. None of it connects to the position it affects, the axe a dealer is showing, a related RFQ or trade. When a PM asks why a market moved, the answer usually sits in one of those chat windows. The desk's own analytics track quoting activity, maybe win rates, response times, where the flow moved. All useful, but none of it says why.

Quote Hub already takes the axes, RFQs, quotes and trades moving through chat and turns them into structured, usable data today: instrument, size, counterparty, time, pulled out in real time and put in front of the desk. Building these core features made us and our customers look at the goldmine of what we weren't capturing. A broker or trader typing 'quiet out there today, EM CB chatter keeping people on the sidelines' hands a client or colleague something useful and instant, but nothing captures and structures it to be shared more broadly or referenced later.

What we're building next

That's the gap News Hub is being built to close. It's still in development, so this describes the direction it's heading in. The plan is to read the same messages already flowing through chat: the broadcasts, the unprompted colour, a research note someone's shared, an invite to a call or event. Each one gets tagged against the same product hierarchy that already organises the axes and RFQs and trades. Look up a product and you'd see what you'd expect: the axes shown in it, the RFQs and responses, the trades that came out of them. Sitting right alongside that, everything the desk has actually said about that market recently, filterable however you need it: asset class, region, counterparty, sentiment.

None of the tagging happens without a person watching over it. If a trader corrects a bad tag, the model learns from it, so the same fix doesn't need making again next week. Every message, tag and change is logged, under the same permissioning Control Hub already runs, so nothing about this becomes a black box just because it's faster.

Matt, our CEO, puts it better than I can: 'These messages help traders combine all the strands of the market together and understand the context behind the numbers. Now that we have AI, and ipushpull has solved the qualitative data problem, we can combine it with the market commentary and news to make an easy-to-understand view of the market.'

Get both halves on one screen and a trader checking their own read of the market finally has something solid to check it against. It gives that judgement a stronger foundation and additional reference point.

A desk that's tracking the quantitative chat data behind its trades, but not the qualitative, is missing half of the information.

It's still in development with prototypes targeted to hit the market in early 2027. If you're interested to get onboard early please reach out.