Hazel is on a mission to collapse the tech stack for financial advisors

Financial advisory firms are drowning in software.

In the day-to-day business of helping their clients plan, manage, and build their wealth, the typical advisory firm juggles an average of 12 tools, many of which tend to be fragmented and poorly connected.

Their tax software may know a client’s tax liability but doesn’t understand how it fits into their financial plan. Their financial planning software has a snapshot of their financial situation but it might not have the most up-to-date information from their last meeting, which is stuck in their notetaker. 

With no unified data model, each program sees only a sliver of the full picture and requires constant manual work to keep up to date. That means advisors spend an overwhelming amount of time context switching, doing data entry, learning different softwares, and troubleshooting it.   One recent study found that advisors spend 45% of their time doing behind-the-scenes work

Many industries struggle to unify their tech stack, but financial advisory firms face particular challenges.

According to the 2026 Investment Adviser Industry Snapshot, 92.8% of advisors employ 100 or fewer employees. Firms that size often don’t have an IT leader to make sure systems talk to each other and the data outputs stay clean. It often falls on the advisor, who usually doesn’t have the technical background or appetite to do this kind of work.

“Advisors already wear a lot of hats: They’re expected to be a financial planner, a tax strategist, a money therapist, and an operations team. Then we hand them 12 pieces of software and ask them to be a systems administrator too,” said Gokul Ramanathan, Product Lead at Hazel.

Ramanathan and the team at Hazel are using OpenAI to build an AI platform for financial advisors that replaces many of these fragmented systems and creates a unified data layer to seamlessly stitch together the rest.

How Hazel helps financial advisors resolve their tech stack

The core of Hazel’s platform lies in the context layer it builds for financial advisors’ clients. It’s designed to let advisors run their firms the way they normally would, while Hazel builds deep client context in the background.

Each time an advisor uploads a financial document to Hazel, it automatically parses and synthesizes the information within it, organizing it and storing it to be put to use. And whenever an advisor has a client meeting, email exchange, phone call, or updated CRM entry, Hazel captures the most important information to get a complete view of their financial situation and goals.

Hazel uses all the information it knows about clients to automatically build financial and tax plan drafts that are highly personalized to their clients’ goals. Hazel takes just minutes to build these plans for advisors to review, a task that would have taken hours, and several pieces of software, to complete in the past.

Hazel also simplifies and automates several of the day-to-day tasks that occupy the bottom of financial advisors’ to-do lists. It can use its client knowledge to generate meeting prep briefings so advisors are always prepared and automatically draft follow-up emails, task lists, and other client communications for review.

“Hazel isn’t one more tool in a stack of many. We want to make most of that stack unnecessary and give advisors back hours they’d rather spend with clients,” said Ramanathan.

How Hazel builds with OpenAI

Hazel uses OpenAI GPT-Sol models to run complex financial planning analyses at a cost point that allows advisors to build detailed financial plans for every client, not just the ones with the highest net worth.

Sol is at the heart of Hazel’s financial planning capabilities, acting as the central orchestrator of its work throughout the entire process.

First, it identifies the intent of the user and the type of plan they want to build, sorting through all the context Hazel has gathered about the client and fetching the appropriate data it needs in order to fulfill their query.

Next, it breaks down each of the client’s goals into smaller tasks and delegates each of the tasks to built-in financial calculation programs for computation. Sol then aggregates the data produced by the calculation programs and serves it to the advisor as a clean financial plan that’s highly accurate, ready to review, and custom-tailored to fit their client’s unique goals and needs.

Ramanathan said Sol scores very highly on reasoning when faced with advanced topics like retirement projections and cash flow modeling.

“Sol is excellent at following complex, multi-step instructions and when faced with an open question, it’s much better at asking clarifying questions rather than assuming on behalf of the user,” he said.

Ramanathan worked closely with the OpenAI team to ensure its models perform optimally on financial planning use cases and collaborated on building the deterministic calculation programs into the platform to ensure the accuracy of each financial plan.

“When an advisor puts a number in front of a client, it has to be right, and it has to be right for the same reason every time,” said Ramanathan. “So the model reads, synthesizes data, and reasons and dedicated calculators handle the math. The thing that makes this architecture work is having a model good enough to know where the boundary is.”