Kscope · deep dive · 2026-10-22
Accern sold a score on a story. The score on public news is a model. The score on the Dow Jones wire is a license. Wand now sells the agents.
The row stays fast. Story sentiment, story volume, and the impact score are models on the public news and the filings. A score that needs the Dow Jones wire stays a license. Text a client uploads only to them stays theirs. Wand acquired Accern on 7 February 2025 and now sells an agent workforce. The pages read on 2026-10-22 print no Accern factor file and no seat price.
| Feature | Accern | Kscope | |
|---|---|---|---|
| Where the product went | Wand AI acquired Accern on 7 February 2025. The Wand post calls Accern a provider of real-time data for finance, compliance, asset management, and risk, and frames the deal as data for Wand's agents. Forbes the same day says Accern worked with hedge funds, asset managers, and equity firms on AI insights for US public companies. Forbes says Accern had raised roughly 40 million dollars, and that the last round was a 20 million dollar Series B in 2022. Ten of the tech and data team joined Wand, which then stood at roughly 70 people. Kumesh Aroomoogan was taking a role, and the title was not decided. Forbes says neither party disclosed the price. The Accern homepage did not answer on 2026-10-22. | The news archive and the filing sections are in house. A factor computed from that text is a model. The company that used to sell the factor now sits inside an agent platform. | Partial |
| What Wand sells now | Wand's 16 June 2025 post describes an operating system for an agent workforce: create an agent, have it run a task, and escalate the edge case to a person. The products named there are an Agent Control Panel and a collaboration platform. The commercial door on the pages read is a demo. Those pages print no Accern factor file and no factor price. | A cited answer over the news archive and the US filings is software on a corpus we hold. An agent workforce that runs inside the client's own systems is their platform. | Beside |
| Feature | Accern | Kscope | |
|---|---|---|---|
| Sentiment, volume, and the story | The docs page, still up, with metadata of 20 November 2017, defines story sentiment as the average sentiment of the related articles so far, and story volume as the article count on that story. A sample story on the page has sentiment 0.072 and volume 58. The same page ranks a source and an author from 1 to 10 on how early their stories are and how often they are republished. First mention is the flag for a story before it has spread. | A mention count and an average sentiment on a story are a query plus a model on the archive. A source rank is the same kind of model. None of those fields is shipped as a daily company factor. The earnings-sentiment note of 2026-09-19 says a sentiment number arrives with the live transcript and has not been stored as a company factor. | Partial |
| Saturation | The same docs page defines story saturation as how much exposure the story has, in three levels: low, mid, or high. The method note says the level is predicted from Alexa Rank web traffic of the related articles, plus historical data. | Article volume is a count on the archive. A level that needs a web-traffic panel is a different input. Alexa Rank is that panel in their own note. | Beside |
| Impact | The docs define two impact scores. The overall score is the historical chance that an event type moves stock prices in general by more than 1 percent by the end of the day. The entity score is the same chance for the company named in the story. The method note says both come from overlaying more than 3 years of events on prices. A sample on the page puts an earnings event at about 49 overall, and at 26 and 36 for two named companies. | The event side is the news we hold. The probability also lays a price history under those events. That join is a model once the prices are in hand. It is not shipped. | Partial |
| Feature | Accern | Kscope | |
|---|---|---|---|
| Public news | The platform guide, dated on the page 2 June 2021, says the public-news collection is crawlers over the web, hundreds of thousands of financial signals a day, in more than 100 languages, translated to English before the model runs. The older docs page, still up, says something wider and does not match that sentence: about 300 million-plus sources through data providers, and proprietary crawlers on about 0.5 million public sources. Both figures are left as printed. | The news note of 2026-09-19 counts 30,176,943 documents. 23,388,726 are RSS. GDELT is 139,355. A live newswire supplies the rest. The corpus holds articles in about 40 languages. The ingest monitors about 100,000 RSS feeds. Search over that archive is live. Their two source counts and our stored-document count are different objects. | Partial |
| Dow Jones | The 2021 guide lists Dow Jones Newswires as its own collection, branded Accern News Analytics, powered by Dow Jones. The wires named are Global Equities, FX, Fixed Income, and Energy and Commodities. The guide says those wires are curated, and that Accern quantifies them. | A Dow Jones story that never hits a public feed is their copy. A factor that requires that wire is a license. | Beside |
| Filings, and the client's own text | The same guide lists EDGAR 10-K and 10-Q sections as a third collection, delivered as a file or an API with the analytics applied. A fourth path is custom data: unstructured text the client uploads, scored alone or beside the Accern collections. | 10-K and 10-Q sections are in the section index: 17,871,301 sections across eight form families, 2016 through 2026, sections note of 2026-09-19. The 10-K and the 10-Q are two of those families. Text a client uploads only to Accern stays in their deployment. | Partial |
| Feature | Accern | Kscope | |
|---|---|---|---|
| The feed, and the price | The docs say the REST feed returns articles processed in the last 30 days, with equities and financial events extracted, and that older history is provided another way. The page metadata is 20 November 2017. No seat price is printed on the docs, on the 2021 guide, or on the Wand pages read 2026-10-22. The Wand door is a demo. | The archive we hold is the history. A 30-day window is their API shape. A daily factor on the archive is a model, and it has no price until we publish one. | Partial |
Ahead The feature is data we already hold, in a form an investor can use.
Match Both products do this.
Partial We hold the filing or the extract, and not their finished screen.
Gap They ship it. We do not.
Beside The Dow Jones wire, a web-traffic panel, and text a client uploads only to them stay theirs. A score computed from the news and the filings we hold is a model.