Skip navigation
estate-planning-keyboard.jpg vaeenma/iStock/Getty Images Plus Debuts ‘Ester,’ an AI-Powered Estate Planning Legal Assistant for Advisors

The proprietary model can read and restate documents and hold both short- and long-term memories of previous client actions.

Estate planning platform debuted its new artificial intelligence–powered legal assistant, 'Ester,' on Tuesday.

With Ester, advisors can upload an existing estate planning document and receive a summary in seconds, allowing them and their clients to understand their existing plans.

Danny Lohrfink, co-founder and chief product officer at, said the proprietary model seeks to enable financial advisors to make “high-quality estate planning more accessible, cost-effective and scalable.”

Unlike other off-the-shelf large language models, what sets Ester apart is its specific training on the data, Lohrfink said.

“If you have an existing trust document, for instance, and you upload the document into the Ester ecosystem, the AI legal assistant will analyze and understand the document in a matter of seconds,” said Lohrfink.

Ester will then facilitate a “frictionless” restatement that extracts all the key information from the existing trust and allows the client to make changes.

In addition, Lohrfink said Ester is conditioned with both short-term and long-term memory, enabling it to incorporate a client's previous actions into its responses and assist in updating existing documents such as wills, trusts and guardianships.

The potential pitfalls of using AI in legal documents have been a hot source of controversy as of late. Earlier this month, lawyer Steven A. Schwartz appeared in a Manhattan courtroom to answer about a legal brief for a case in Federal District Court that was filled with fake judicial opinions and legal citations, all generated by ChatGPT.

Lohrfink said they had considered the implications of these so-called hallucinations generated by AI when they were developing Ester.

READ MORE: Advisors and AI

“When properly supervised, these LLMs, these technologies can be incredibly powerful, but if left unsupervised or untrained, they would be very misleading and even dangerous,” said Lohrfink.

F1 scores are the measure by which machine learning tools are used to gauge accuracy. A recent study showed that OpenAI’s ChatGPT scored at 49% on this scale. Lohrfink claims Ester’s F1 score is above 93%. Also, unlike ChatGPT, Ester is trained to give specific instructions to users to validate the output.

“I think that really speaks to how we have trained this model to be highly accurate but also conditioned with our internal knowledge, experience and data,” said Lohrfink.

In addition, is SOC 2 Type 2 and Health Insurance Portability and Accountability Act of 1996 compliant. Other security measures include 256-bit data encryption, trusted site certification and multifactor authentication.

Lohrfink said they have been working on Ester since October 2022 when they hired Eduardo Fontes, who is now senior vice president of data science. Fontes, who holds a master’s degree in data science from Massachusetts Institute of Technology, served as the managing director and head of data intelligence at TIAA from August 2021 to August 2022 and was head of data science and machine learning at Vanguard from July 2017 to August 2021. In addition, Fontes previously held data science and machine learning roles at major financial institutions such as Amex, Visa and American Family Insurance.

“He’s been doing machine learning for 28 years well, well, well before it was cool,” said Lohrfink.

William Trout, director of wealth management for Javelin Strategy & Research, said AI has increased advisor workflow enablement across the industry. However, Trout said, it remains to be seen which of these truly stand the test of time.

“I think long term, a lot of these tools are first-movers, but the core hypothesis that you build an LLM and the tools that can support it (is) interesting,” said Trout. “I think a lot of them are green shoots that may or may not survive. But I think the place of AI is pretty fundamental.” is used by over 1,000 advisors in the United States, including firms like Keebeck Wealth Management, MarketCounsel and Farther.

“We work with independent RIAs, broker/dealers and custodians, but all in the financial advisory space,” said Lohrfink.

Lohrfink said they charge advisors on a per-seat model and not a per-client or a per-document basis. Ester is being embedded in that existing per-seat cost. was listed as a finalist in the ninth annual Industry Awards in the Technology Providers category under Specialized Planning Applications. Winners will be announced Sept. 7 at a black-tie gala at the Ziegfeld Ballroom in New York City.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.