Falcon Investment Advisors turns 70-page CIMs into signed-off Financial Summaries with Claude
Centrilogic enables Falcon Investments to turn a complex and manual CIM review into a 5-minute, fully automated process with a custom Claude skill.
The Problem:
Falcon Investment Advisors, LLC (“Falcon”) is a private investment firm. Investment Committee (“IC”) memos include several financial output slides that have historically been prepared manually, creating the potential for human error. In each memo, a member of the deal team builds a one-page financial summary from the company’s Confidential Information Memorandum (“CIM”), a deck of roughly 70 pages, much of it images, in which the financial overview occupies a handful of pages whose position varies from deck to deck. The analyst identifies the relevant pages, manually re-keys the underlying financial figures into Excel, and performs the required calculations and analysis, an inherently time-consuming process that is susceptible to user error.
The work recurs for every memo, and is exposed to exactly the mistakes that matter in this domain: a figure in the wrong units, mis-types, gross revenue where net was wanted.

The Solution:
CentriLogic built Falcon a Claude skill that outputs the historical financial profile of the business from the CIM into a templated, formatted, and consistent output for use in all memos. The analyst attaches the CIM in Claude Cowork, names the company and asks for the summary. Claude finds the financial pages and proposes the line items for every reported period. The analyst reviews the proposals, in the chat itself or in an Excel review workbook if preferred, approving or correcting each figure. Deterministic code then writes the approved figures into Falcon’s own template, creating a consistent, clean output for each memo.
The principle behind the design is simple to state: Claude proposes; the analyst approves; deterministic code writes. The model never performs arithmetic on a financial figure. Growth, margin and adjusted EBITDA remain the template’s own formulas, and nothing reaches the template that an analyst has not signed off.
Every figure the skill records carries the page it was read from, the row as it was printed, the unit as printed, a confidence score and the name and date of the analyst who approved it. Anything below Falcon’s 90 percent confidence threshold is flagged for a human decision rather than written quietly. When a company’s next CIM arrives, the earlier approvals are kept, and a new extraction that disagrees is shown to the analyst as a pending re-proposal instead of being applied over their judgement.
The skill runs inside Claude’s own execution environment against a folder on the analyst’s machine, so Falcon’s records stay where Falcon keeps them. Everything specific to Falcon, from the template to the confidence threshold, lives in configuration, which let CentriLogic bring the same skill to a private-equity client without changing the code.

The Outcome:
The skill went live with named Falcon analysts on September 24, 2026, with the first live CIM processed end to end on that date and 8 analysts or associates running it against real deals inside the first week.
Building a Financial Summary by hand took Falcon’s analysts about 20 minutes of manual effort per deal. Through the skill, from attaching the deck to a populated template, it takes less than 5 minutes, while the analyst can work on other tasks while Claude is populating the template.
Against a reference set of CIMs with known answers, the skill’s extraction was highly accurate, and every figure that fell short of the confidence threshold was surfaced for review rather than written.
Most importantly for a financial team, every number in every summary now answers “which page, who approved it, and when” without a conversation.
“Centrilogic helped us turn a time-consuming, manual process into an efficient and controlled workflow. The skill preserves oversight while improving consistency, traceability, and confidence in the financial information supporting our investment decisions.”
— Ryan Hewes, Associate.
Centrilogic delivered the skill under a fixed-fee engagement of roughly four weeks, from first workshop to production users, and provided Falcon knowledge-transfer allowing for ongoing use of the skill.

Explore our AI Solutions
AI Discovery Assessment
Identify opportunities for AI-powered innovation to drive operational efficiency, revenue growth, and competitive advantage.
Centrilogic’s AI Discovery Assessment will empower you to realize the full potential of AI and uncover opportunities to improve scalability, reduce operational costs, and accelerate decision-making. In a focused, executive-level engagement, our experts work alongside your leadership teams to identify high-impact AI opportunities, prioritize use cases, and deliver a clear, actionable roadmap to pilot AI solutions.
Build & Deploy Custom AI Agents
Transform your business with custom AI Agents developed and managed by Centrilogic’s team of AI experts. We will work together to build and deploy AI Agents designed to automate tasks, enhance productivity, and address a wide range of unique use cases while integrating seamlessly with existing applications, data, and infrastructure.
AI Agents can transform countless business processes across customer success, business development, predictive maintenance, financial analysis, and many more.
Are you ready to realize your digital potential?
Partner with Centrilogic to turn your AI ambitions into reality.