Bloomberg is rebuilding the Terminal around a chatbot. ASKB, a natural-language interface stitched on top of a basket of language models, is open in beta to roughly a third of the Terminal's 375,000 users — about 125,000 seats — and chief technology officer Shawn Edwards told WIRED it will become 'the new Terminal' and the default starting point for most workflows. No general-availability date has been set.
The pitch is that the Terminal's data trove has outgrown its keyboard-shortcut interface. Bloomberg has spent years piling weather forecasts, shipping logs, factory locations, consumer spending data, and private loan information on top of earnings and asset prices. 'It has become more and more untenable,' Edwards said. 'You miss things, or it takes too long.'
ASKB is designed to let users pose investment theses in plain English instead of hunting for specific tickers and functions. Edwards framed the core problem as helping finance professionals 'find key insights and synthesize a view of the world around a particular idea,' with the wager that 'untapped alpha lurks somewhere in the data.' The example he offered: asking how a war in Iran and a shift in oil prices might affect a portfolio, and getting a synthesized answer in minutes.
Key facts
- 01Bloomberg's ASKB beta is live for roughly a third of the Terminal's 375,000 users, or about 125,000 seats.
- 02CTO Shawn Edwards told WIRED ASKB will become 'the new Terminal' and the default starting point for most workflows.
- 03The chatbot is built atop a basket of language models with validation checks at each step to limit hallucinations.
- 04Edwards pitches ASKB as agentic, letting analysts schedule earnings-prep workflows and trigger them on market conditions.
- 05Bloomberg has not announced a general-availability date for ASKB beyond the current beta.
The agentic piece is workflow templates tied to the earnings cycle. An analyst preparing for a quarterly call can write a single long query that pulls peer pricing, fundamentals, Street estimates, and prior guidance, then schedule it or trigger it on a market condition. Edwards described earnings season as a stretch when analysts 'are not sleeping,' and ASKB is positioned to absorb the legwork that typically falls to junior staff.
“ASKB is open to roughly a third of the Terminal's 375,000 users, and Edwards says it will become the primary way most interactions on the platform start.”— Jaeden Schafer
That framing has obvious labor-market consequences, and Edwards did not dodge them. He said the industry has not solved how to 'educate, train, and mentor junior analysts' once the work that taught them the craft gets automated. He also pushed back on the idea that the tools level the field: 'These tools are not magical. They don't make an average employee all of a sudden great. The difference will be your ideas.'
His sharper line on that point: in the hands of experts, ASKB lets analysts 'sift through 10 great ideas when they might have only had time for one.' For weaker analysts, the math runs the other way. 'If you're a mediocre analyst, they'll be 10 mediocre ideas.'
Hallucinations are the obvious risk in a product where customers move money on the output. Edwards said Bloomberg builds validations 'at every single step,' including a check that compares facts in a summary back to the underlying paragraphs, a semantic check to catch inverted statements, and a check on citations. He stressed that Bloomberg 'never gives a buy or sell signal' and that the more common failure mode is incomplete answers rather than fabricated ones, with transparency to underlying sources as the guardrail.
On whether ASKB is a defensive move against startups trying to vibe-code a cheaper Terminal clone, Edwards said no. Bloomberg has tried versions of this for years using earlier machine-learning techniques and brittle natural-language command lines that demanded specific phrasings. 'But finally, we have the technology to do it,' he said, adding that mission-critical decision systems are not something to vibe-code.
The Terminal's most loyal users — the ones who pride themselves on memorizing function codes and keeping sticky notes on their monitors — are the constituency most likely to resist. Edwards said the redesign is aimed at them anyway. He wants to retire the sticky-note workflow, not the power user, and he conceded that GUIs are not going away. The mouse stays; the muscle memory of arcane commands is what ASKB is trying to make optional.
The unresolved question is what happens to the analyst pipeline when the work that produced senior analysts is the first thing automated. Edwards acknowledged the gap without claiming to have an answer, and in finance the cost of acting on a confidently wrong synthesis is higher than in most domains the chatbot wave has touched so far. Bloomberg's validation stack is more elaborate than what most consumer chatbots ship with, but no vendor can promise zero errors on open-ended prompts.
For Bloomberg, ASKB is a bet that the Terminal's moat is the data and the workflows around it, not the cryptic command syntax that defined the product for four decades. If the chatbot becomes the front door for 375,000 of the world's most demanding software users, it sets a high-water mark for how generative AI gets embedded into mission-critical professional tools — and a template the rest of the enterprise software stack will be measured against.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
