Customer Story
5 min read

How Roundtable automated its back-office processes?

Published on
September 23, 2025
Contributors
About the customer

Roundtable is a collaborative investment platform specializing in the creation and management of SPVs (special purpose vehicles). The company supports its clients in structuring, legal compliance, and the administrative management of these vehicles.

With complex operations and a high volume of documents, Roundtable’s teams deal with legal, administrative, and back-office tasks: reviewing contracts, extracting information, ensuring regulatory compliance, and tracking investors.

Watch the customer story video

Before Phacet: manual and scattered workflows

Julien, co-founder and CPO of Roundtable, sums up the team’s daily challenges before Phacet:

“Reading a 60-page document to extract information easily took an hour.”

The back-office relied on Forest Admin, Google Drive, Notion, and Excel, but no solution could centralize and automate data extraction and review.

Julien had tested general-purpose LLMs, but the chat interface quickly showed its limits:

“Chat interfaces are great for brainstorming, but not for reviewing or processing hundreds of documents.”

Why Phacet clicked: the “ExcelGPT” moment

What convinced Roundtable was Phacet’s spreadsheet interface:

“Each row represents a record, each column an attribute. It’s intuitive, everyone understands it. That alone drastically reduces onboarding time.”

Julien adds:

“Phacet is like having ChatGPT in Excel, with the same experience as Notion.”

The promise: using AI agents not as a one-off tool, but as a real work infrastructure.

AI agents in action

Document extraction

Roundtable began by automating the extraction of data from annual accounts and investment contracts.

“What used to take an hour per document now takes 30 seconds.”

Regulatory compliance

The legal team then trained a Phacet with ESMA documentation (European regulator). The result: an automatic report, page by page, highlighting discrepancies to fix.

“It saved our legal team an incredible amount of time.”

Ad hoc data processing

Phacet also handled more specific needs, such as catching up on thousands of missing data points.

“One of our interns updated a database of thousands of records in just a couple of days.”

The impact: speed, adoption, and value

The immediate effect was a drastic reduction in processing time:

  • Extraction: from 30–60 minutes to 30 seconds per document.
  • Compliance: hours of manual review eliminated.
  • Bulk data updates: several days of work reduced to a few hours.

But the impact goes beyond productivity:

“We start by automating repetitive tasks, which frees up time for more meaningful work.”

Roundtable’s vision: AI as a new colleague

Julien sees Phacet as part of a long-term transformation of back-office roles:

“We’re not eliminating jobs, we’re making people more productive. Their contribution to the company’s value creation increases significantly once they’re freed from repetitive tasks.”

He identifies three major areas where AI will become essential: customer support, KYC/KYB, and ad hoc tasks where a generalist tool like Phacet excels.

Julien’s advice to other leaders

“It has to come from the top. Just saying ‘we need AI’ isn’t enough. You need to show a concrete use case, how it was implemented, and what gains it delivered. That’s internal marketing.”

His closing thought:

“The promise of Phacet is to automate extremely time-consuming workflows, so teams can focus on their core business.”
“Since working with Phacet, I’ve saved precious time on purchase analysis. I no longer get lost in Excel sheets or PDFs—I save up to two days per month and catch mistakes I would never have spotted on my own.”