A startup known for its API for summarizing massive amounts of data is coming out with a new app of its own that proves there’s a profitable market in rendering texts in a digestible fashion. Calling it sdan.io/summary, the startup explains that it has been helping startups with large quantities of customer feedback and applications by “anonymously hashing and preserving all the company’s web interactions”. So it launched sdan.io/summary as an optional user-designed snippet of text with some artificial intelligence smarts to reveal an insight or describe data you’re processing.
That’s not big news to many, but sdan.io/summary took our test run and distilled it so long-form paragraphs aren’t necessary. So the startups that use it might not have to wait for the next marketing analyst to get back and polish their story.
Unlike Summarizer or Summarkey, sdan.io/summary is just a predictive snippet. You can browse by term or entire fields. Once you click one, a snippet appears with two or three sentences that summarize, describe, and identify data in context. Each snippet comes with a host of predefined phrases that emphasize different possibilities and themes, like “The Daily Pulse” and “The Startup Story”.
If you want to break the sentences into sub-items, like “Inbox iscoding project”, it can quickly create up to 100 sub-items. That’s useful if you want to assign different weights to the version of the same thing that’s been excerpted. Or it can provide detailed descriptions like “Perform turnaround value”, “2 HR assets”, or “Internal strife” so you can look at the data more easily.
Sdan.io/summary starts with your data, like your customer feedback or internal documents. Then your programmers connect it to existing web tools and test what words may be useful. After a lot of brainstorming, they’re left with up to 150 words describing the data.
Sdan then sits back and explains what it sees. It can break the data down to hyperlinks, PDF files, tables, slices, graphs, or images to get deeper analysis. Then it “writes” the words in the summary to a fixed length. That’s where the AI comes in. Wordflow uses machine learning and natural language processing to see if words on the summary are likely the ones to describe the data, but leaves out redundancies and sarcasm. Meanwhile, it gets your data down to a precise size, in a columnal format. Then it edits the summary, gets your data to fit, and writes the word count in the same way it would a PR person talking to investors, partners, or the press.
Wordflow is co-founded by Classified Intelligence’s CEO Rafael Huschka and former CNN reporter Anne T. Donahue, as well as Amazon and Hadoop veteran John Fitzpatrick and two other ex-coder buddies. With more and more data, now is the time to come up with quick and concise forms that summarize it and explain it in a way more accessible to researchers or anyone who wants to read a report of a PhD seminar.
The free version of sdan.io/summary lets you annotate data, for free. The paid plan gives you up to 100 free authorship tags, which allow you to give each snippet an individual title, and unlimited full-text annotations. For $89 per month you can upgrade to Full Text, which provides cloud storage of full text annotations and the ability to export on popular API callbacks like XlsK, and $149 gets you a paid plan that gives you access to Wordflow’s Sage Insight engine.
I’m excited to see how startups are using sdan.io/summary. Right now it’s only available on some emersion APIs, but Wordflow is building a link back into the rest of the data they’re pulling in for automatic moderation. That may let those startups eliminate the prospect of anything getting labeled out of context or redacted unnecessarily, freeing them to truly highlight insight.