Instabase now valued at over $2 billion
Instabase has announced the completion of a Series C round for $45 million. This values the company at $2 billion, double the valuation from its Series B round of 2019. Instabase is now the most valuable software company whose main business is to provide solutions for unstructured document or data processing.
How did the company get here? I see two reasons. First, Instabase built a solid AI platform and the resultant growth metrics must be impressive. Although revenues are confidential, the company now has 360 employees (up from only 42 at Series B). I’ll go out on the proverbial limb and estimate annual sales at $50-75 million.
The second – and in my opinion most important – reason for this valuation boost is that Instabase hitched a ride on the generative AI rocket. The new investment coincides with the company’s introduction of new generative AI applications built on large language models (LLMs). Instabase’s AI Hub is now generally available to customers and includes the Converse app where one can carry on a natural conversation with documents and receive precise answers.
As I mentioned last week in the 4th Wave of IDP blog, foundational LLMs are disrupting the product plans of every IDP company and altering expectations of every IDP end user. What has Instabase done with generative AI that would super-stoke the new valuation?
Kick the Tires
You know me. Former product manager who geeks out on product demos and testing new stuff. Instabase’s AI Hub is free to try with 1,000 tokens, so I enrolled in 30 seconds and opened the Converse app to – well – have a conversation with my documents.
I uploaded a five-page promissory note, a PDF file with text. This note is not a standard banking or lending form; it is a highly modified version created by the lawyers for a private company specifically for loans to executives. I think we can agree this is a good example of an unstructured document with unknown variability.
I carried on a conversation with the file, requesting information such as:
- Summarize this document.
- What is principal amount of the loan, and what is the interest?
- When is repayment due?
- If the Maker’s employment is terminated, what happens?
- Does the file contain a signature?
- Show me the signatures.
In each case, Converse gave accurate and complete answers and it also took me to the page where the first answer was extracted. Behind the scenes, the engine is reading the data and extracting data for each reply. The next step would be to automate this for hundreds or thousands of documents a day, extracting the data requests and entering that data into a business application, or triggering a workflow based on the data.
Converse never saw this document before. I did not have to train a model or annotate and label samples. The app applied zero shot learning for document classification, data extraction, and summarization, and did it without error or correction. This is the 4th Wave magic and it upends many preconceived notions about training and deploying IDP.
Who are those guys?
Instabase is not normally covered as an IDP company because its marketing has focused on building AI apps for the enterprise. Very wise positioning which probably led to the first unicorn status in 2019. But go to the website and you will learn that “Instabase is a platform that…applies the latest AI innovations to understand…unstructured data and build applications.” With Instabase, you can “get answers, summarize content, and analyze data instantly” and “leverage…generative AI to converse with your documents.” The featured case studies are good old, classic IDP use cases: invoice processing, know-your-customer (KYC), contract analysis, etc.
So by its own messaging, Instabase self-identifies as an IDP company and deserves to take the top spot for highest ever enterprise valuation. Deep Analysis has tracked Instabase for years and we included the company in our market research for the 2022 IDP Market Report. We will continue to track their progress and report on innovations.