Bloomberg's ASKB now reads your chats, spreadsheets and files

Private context added to market data

According to Bloomberg, ASKB is built on the reliable, proprietary data the company has held for decades. The latest step lets clients bring in content that previously sat outside their AI tools: messages from Instant Bloomberg, their own spreadsheets and their files.

Wayne Barlow, Bloomberg's global head of core terminal products, described the gap this way: some of professionals' most valuable information, such as a trader's conversations, a portfolio manager's watchlist or an analyst's research, was outside the AI tools they use to make decisions. In the release he says ASKB's strength is combining an investor's professional context with Bloomberg's market information in a single conversation.

Searching the noise of Instant Bloomberg

Bloomberg says its clients send about 40 million instant messages a day in Instant Bloomberg. The company argues that this volume holds some of the most recent and relevant market views and trade ideas available to a finance professional, but that the information is scattered across conversations and disappears by the afternoon.

Once ASKB is activated, clients can ask questions in plain language about their own Instant Bloomberg interactions and receive an answer that links back to the source conversation. Bloomberg gives several examples:

• a summary of what contacts said today about oil;
• a catch-up on everything that came in overnight;
• surfacing the dissenting view of a bearish contact before making a decision.

ASKB is optimized to recognize the abbreviations, initialisms and acronyms that make search difficult in financial messaging, and it can use three months of a client's chat history.

Spreadsheets and the Bloomberg local drive

Clients can use spreadsheets in ASKB to ask about the securities already on their coverage lists and watchlists. Instead of typing tickers one by one or rebuilding filters by hand, an analyst can ask for the top performers in a spreadsheet, the biggest movers, or which companies report earnings this week. ASKB then returns an analysis based on its existing data, news and research tools. According to Bloomberg, the client's own lists become the starting point of each question rather than a separate preliminary step.

The Bloomberg local drive in ASKB gives clients one place to upload, organize, browse and share files, from internal research such as articles, investment theses and earnings recaps to other working documents. Bloomberg says a file attached in ASKB used to vanish once the thread ended. Now, once ASKB is activated, everything on a client's local drive, including files and folders shared by colleagues, is available for questions and remains a source for future conversations.

Users can ask about a single file, a whole folder or their entire library and get one answer drawing on their own content as well as Bloomberg data, news and research. Each answer cites the file and passage it comes from. Uploaded files are visible to others only if the uploader chooses to share them with colleagues at their firm, and client documents stay permissioned and separate from Bloomberg's own sources.

What it means for finance professionals

Bloomberg's pitch is less data gathering and more time for analysis. The company says finding an answer usually means knowing which sources to consult, collecting the data and assembling it by hand, and that ASKB does this in one conversation, drawing on Bloomberg's experience rather than a general-purpose AI model.

Shawn Edwards, Bloomberg's chief technology officer, calls ASKB the most advanced tool the company has built and the one adopted most quickly. He says the feedback Bloomberg hears most often concerns trust, because answers rest on familiar sources and each one clearly shows where it came from. In his words: "Ce n'est pas Bloomberg qui s'appuie sur l'IA. C'est l'IA qui s'appuie sur Bloomberg."

The release adds that Bloomberg has developed and used AI in finance since 2009.