Today, Arkifi emerged from stealth and announced that Nyca Partners has co-led its Seed round alongside Khosla Ventures. There are many opportunities for AI to drive efficiencies and enhance the quality of product offerings across financial services, but we believe material value creation will be entirely dependent on specific domain context and controls. So we can’t wait to see the reaction of sophisticated investment professionals when they try Arkifi’s AI-driven workflow automation platform, and the impact it will have on their productivity.
First, some background. A core belief we’ve had is that virtually all profit pools in financial services firms result from competitive advantage of information. If you are underwriting someone’s life, issuing an installment loan, trading bonds, or deciding to approve or turn down a payment transaction because of fraud or AML risk, being better at capturing or analyzing data is core to competitive advantage. Obviously, competitive advantage of data gets harder and harder and–ironically–more expensive to achieve as data costs decline. In capital markets this is particularly stark: being physically in the pit at the CBOE was critical until the late 1990s, but pits don’t even exist anymore, and HFT algorithms now decide about two thirds of futures volume. In US equities bid/ask spreads have fallen from 65 bps to 5bps over the same period.
When Hans started on Wall Street in 1980, most investors did not have access to real-time market data. The first Bloomberg Terminal was introduced in December 1982, and within a few years you could not be a professional investor without one. Last year, Bloomberg earned $12bn in revenue. Some of the most notable recent acquisitions in financial services — including LSE Group’s $27bn acquisition of Refinitiv, Nasdaq’s announced $10bn acquisition of Adenza, and Deutsche Boerse’s $850mm acquisition of Nyca portfolio company Axioma — demonstrate how exchanges realize that pre-trade and post-trade data and workflow are where value is created. But yet: repetitive number crunching in Excel remains the core of every investment analyst’s job; it’s not really that different from 1995. And the numbers have to be 100% right, not directionally right.
Enter Arkifi. Arkifi is building a Financial Analyst copilot that will allow investment professionals to conversationally query and analyze data from a wide range of sources: SEC filings, Bloomberg, CapIQ, FactSet, and other trusted provenance. Whether a hedge fund analyst wants to build a new model or modify an existing one, pull trading comps or format output for slides, Arkifi will automate these tasks, allowing the analyst to focus on the subjective, more nuanced aspects of their work. Indeed, Arkifi’s promise lies precisely in how little financial analysts will need to change their workflow: analysts will primarily interact with Arkifi in spreadsheet software like Excel, and critically, all of Arkifi’s outputs will be intrinsically verifiable. Arkifi’s outputs are formula-linked to the underlying data sources creating immediate trust between the Arkifi product and its users.
Arkifi sits at the center of many exciting developments applying innovations in generative AI to financial services. These include BloombergGPT, a finance-specific LLM that interprets sentiments in financial news and generates Bloomberg queries, as well as FinGPT, a set of finance-oriented LLMs that have emerged from the open source community with envisioned use cases spanning everything from robo-advising to credit underwriting. Arkifi is unique in its focus on freeing up financial analysts’ time to do more critical thinking and analysis, its integration with a range of data sources, and its auditability. Hedge fund PMs for whom “there might be a mistake” is never an acceptable response will be comfortable with their analysts using Arkifi.
As we have gotten to know the Arkifi team over the past few months, we have been impressed by the experience and insights that they bring to the task of building AI for financial workflows. Hesam and Jeremy first envisioned a version of the Arkifi product when they met eight years ago at Johns Hopkins, where Hesam was a PhD student in biophysics and Jeremy was an undergrad. They then went their separate ways — Hesam wrote the textbook on statistical securities analysis, and Jeremy spent time in banking at JP Morgan, at Bridgewater, and received a Master’s in ML — but they reconnected as AI was advancing enough to make their long-simmering idea to automate investment analyst workflows feasible. Since January, Hesam and Jeremy have recruited an equally high caliber team and have impressed many others beyond us, including executives at some of the world’s most respected hedge funds, as well as banks, consulting firms, and others in the Nyca network.
Arkifi is just getting started in its journey to help investment professionals spend their time on the decisions that most drive investment returns. We look forward to supporting them in the years ahead as they launch and expand their product offering and win the trust of key decision makers across financial services.