Nanonets, a startup leveraging AI to automate back-office operations, has successfully raised $29 million in a funding round led by Accel. This investment aims to enhance the accuracy of automation processes that handle significant amounts of unstructured data.
Managing unstructured data from documents such as invoices, receipts, and purchase orders often involves repetitive tasks and considerable human resources. Nanonets, primarily focused on the financial services sector, has developed an AI platform designed to increase the efficiency and cost-effectiveness of these processes.
As a graduate of Y Combinator, Nanonets has created a no-code AI platform that helps businesses extract vital information from documents, emails, tickets, and databases, transforming them into actionable insights. By utilizing advanced machine learning architectures, the platform analyzes unstructured data from uploaded documents, efficiently extracting valuable information. Its no-code AI agents can integrate seamlessly with ERP systems like QuickBooks, Xero, Sage, and NetSuite, optimizing accounts payable, enhancing supply chain management using historical data from platforms like Square and Tableau, and summarizing health reports from patient management systems.
Nanonets claims that while processing an invoice manually often takes around 15 minutes, their automation solutions can cut this time down to under a minute, facilitating processes such as accounts payable, reconciliation, accounts receivable, and expense management.
The startup plans to allocate the new funding towards research and development to enhance system accuracy and invest further in sales and marketing efforts. Currently employing around 100 people, predominantly in engineering roles based in India, Nanonets also intends to expand its workforce with this fresh investment.
The all-equity Series B round saw participation from existing investors Elevation Capital and Y Combinator, bringing Nanonets’ total funding to $42 million, which includes a $10 million Series A round in 2022.
Co-founder and CTO Prathamesh Juvatkar shared that Nanonets initially utilized convolutional neural networks for image analysis to identify key objects. The startup later considered implementing graph neural networks but ultimately shifted to transformer architectures, finding them to be more precise than existing machine learning technologies. “At our backend, we deploy multiple architectures. For each new customer, we train these models on their data to determine which one delivers the best accuracy,” he noted in an interview.
IIT Gandhinagar alumni Juvatkar and Sarthak Jain (CEO) co-founded Nanonets following the sale of their machine-learning platform, Cubeit, to fashion portal Myntra in 2016.
Unlike many AI startups that depend on large language models (LLMs) and GPTs, Nanonets opts for transformers to avoid the common issue of hallucinations—when AI generates information not contained in the provided documents, relying instead on LLM knowledge.
Even though Nanonets employs document-agnostic machine learning architectures, the startup primarily targets the financial services sector since about 50%-55% of its customers hail from this industry. Nanonets offers various integrations that streamline finance operations while gradually expanding into adjacent sectors, including healthcare and manufacturing, Juvatkar explained.
In the competitive landscape of AI-driven workflow automation, Nanonets distinguishes itself against traditional optical character recognition (OCR) platforms and rivals like Rossum AI and Hyperscience. Major companies like UiPath also provide structured data workflow automation. However, Juvatkar asserts that Nanonets stands out by achieving a remarkable 90% straight-through processing rate—the percentage of data processed without manual intervention.
“Our competitive edge lies in our accuracy, user experience, and the quality of our integrations,” he stated.
Nanonets provides its solutions across three pricing tiers: Starter, Pro, and Enterprise. Juvatkar highlighted that the Pro and Enterprise tiers are the most significant contributors to the startup's annual recurring revenue, each providing equal value. The company also offers tools to convert PDFs to various formats, including Excel, CSV, JSON, XML, and text, as well as image-to-text and image-to-Excel conversions. These offerings have attracted businesses seeking automation solutions, allowing Nanonets to reach over 34% of the global Fortune 500 companies in the past two years. Furthermore, the startup's user base has quadrupled in the last year, currently comprising over 10,000 customers worldwide.
While Nanonets serves a global market, the U.S. accounts for about 40% of its revenue, and Europe contributes between 30% and 35%, Juvatkar noted.
Since the 2022 funding round, Nanonets has seen a threefold annual revenue increase, with aspirations to double or triple its top-line performance this year. This consistent revenue growth is a key reason investors continue to support AI startups, despite a general market slowdown. According to Tracxn, funding for AI startups surged to $21 billion in 2023, up from $10 billion in 2022, even though the number of deals decreased by 61 to 399 last year. The U.S. remains the leading investment destination for AI startups, followed by China, the U.K., Israel, and India.
“We are excited to collaborate with Nanonets in their mission to transform back-office operations through AI. Sarthak and his team are committed to addressing customer pain points and have developed a robust solution that fully automates business processes from start to finish. Nanonets stood out due to its comprehensive platform and effective Straight Through Processing (STP) capabilities, which have already demonstrated significant positive impacts for customers,” said Abhinav Chaturvedi, partner at Accel, in a statement.
Note: This article was updated following a request from Accel for clarification regarding the firm's name.
Ranking
