Powder, a fintech startup that launched at the annual T3 conference in January, has raised $5 million in seed funding from 40 Silicon Valley-based venture capital firms and technology leaders. These investors include YCombinator, General Catalyst, Funder's Club, Elefund, Litquidity Ventures and Script Capital, as well as Jon Xu, a co-founder at FutureAdvisor, and Bryant Chou, co-founder of Webflow.
Powder CEO Kanishk Parashar said the tech now has 20 mid-sized RIA firms signed up, ranging from $1 billion to $100 billion in assets under management. Clients include Catalytic Wealth, which is the wealth management firm of General Catalyst, EPIQ Capital and IEQ Capital. Over $13 billion in proposals have been generated since the launch of Powder.
Powder uses generative artificial intelligence “agents” to read brokerage/banking and wealth documents that come through a wealth management firm's proposal process. The technology enriches those documents with missing information, fixes the problems, and then provides an extract for analysis.
“When advisors meet a new prospect, they get a bunch of documents at one point, and then the advisor tries to demonstrate their value proposition,” Parashar said. “As they do that, they have someone on their team reading all these documents. They can be hundreds of pages of information.”
Dust is trying to solve that manual pain point. Its workflow starts when someone uploads a series of documents from a client.
“When it's uploaded, we use generative AI to read the document, figure out what part of the document is valid, and extract that information from the document in a format that can be used to do analysis,” Parashar said. “Firms are telling us they're saving 90% of the time they used to spend reading those documents.”
Parashar said he and his team got the idea for Powder after selling their last company, Navigator Addepar in January 2021.
“While we were building Navigator and while we were at Addepar, what we realized is that advisors and firms in general—the value of the relationship is the No. 1 differentiator—the trust they build.”
He said that trust is built through interactions and information, but that information is unstructured, such as in documents and conversations.
“Generative AI is able to understand unstructured data and turn that unstructured data into usable information,” he said.
They launched the technology with AI agents for brokerage/bank statements and property documents, but the funding will be used to invest in other workflows and use cases.
The company is also launching a chatbot that allows advisors to ask further questions about the same document, such as “What are the total unrealized gains?” or “What are the five main properties?”
The technology also has hallucination detection to double-check the accuracy of AI-generated responses.
“We built a ton of scaffolding around the basic generative part of the AI to make sure it works.”
Parashar said Powder is different from some of the similar AI tools out there in that it starts very specific, with AI agents doing a task. Then, over time it expands the field.
“Instead of building an AI bot that answers everything, we're taking a more precise approach,” he said. “Over time, we want to build trust for those use cases and add more use cases over time. Eventually, it will do almost all the manual work that firms are doing.”
Some fintechs have tried their hand at providing AI Assistant-type tools. Take Benjamin, for example, which launched in 2019 as a tool to streamline back-office workflows for financial advisors. starting reversed at the end of 2020 to become an AI-driven business support system for RIA firms — or, as the company described it on its website: “The world's first AI assistant built for advisors by advisors.”
company it closed last yeardue to the challenges of landing enough customers and additional capital.