Healthtech CFO Says the Hard Part Isn’t AI — It’s What Lies Underneath
Discover how a healthtech CFO uses AI to automate billing and collections workflows, while building the data foundations needed for reliable finance operations.
Fernando Blumenkron
CFO at Eden
AI Workflow Transformation: AI is streamlining finance tasks, improving speed and proactivity in billing, collections, and expense management.
Human Oversight Required: Despite AI's efficiency, human judgment remains critical for budget approvals and other key financial decisions.
Data Quality Importance: Clean data infrastructure is essential for AI effectiveness; poor data quality can lead to false confidence.
Redesigning Cash Flow: CFOs should focus on automating the Order-to-Cash process to enhance cash flow management and controls.
Continuous Improvement: Embracing AI's potential means designing finance operations that self-improve and drive efficiency over time.
Fernando Blumenkron is CFO at the global healthtech company, Eden. He's currently implementing AI across finance workflows, with special emphasis on billing, collections, and expense categorization.
We sat down with Fernando to learn what he has successfully automated. He said success depends on the underlying infrastructure.
Scaling the finance organization beyond startup execution
I lead Finance at Eden, a US-based AI healthtech company operating across Latin America. The company is in a growth stage, with a US parent structure and complex commercial, legal, and tax operations across Mexico, Brazil, Colombia, and other regional markets.
Create a Free Account to Read More
You'll also get access to a growing community of modern CFOs and finance executives accessing proven frameworks, tools, and insights to navigate AI-driven finance.
Have an account? Log In
The finance function covers a multi-entity, multi-currency operation with SaaS and usage-based revenue, implementation services, distributor relationships, enterprise accounts, local invoicing requirements, FX exposure, and AI/infrastructure cost management.
The team is lean and built around the core finance operating pillars: FP&A, accounting, billing, collections, cash application, tax/compliance, and strategic finance. A major part of my role is scaling the finance organization from startup execution into a more institutional operating model, with stronger controls, better data, cleaner O2C processes, and closer partnership with Sales, Legal, Operations, and Product.
How AI Transforms Finance Operations
In the last year, we shifted parts of Finance from manual data compilation to AI-assisted process monitoring, especially across billing, collections, and expense categorization.
Before, the finance team spent too much time pulling data from different systems, cleaning reports, checking invoice status, reviewing aging, categorizing expenses, and manually identifying follow-up needs. The work was necessary, but highly operational and reactive.
We introduced AI-enabled workflows that monitor financial data, surface exceptions, classify transactions, identify collection priorities, and prepare the first layer of analysis for human review. The goal was not to replace Finance's judgment, but to remove the repetitive preparation work that prevented the team from focusing on decisions.
As a result, Finance became faster and more proactive. The team now spends less time assembling information and more time acting on it: prioritizing overdue accounts, identifying billing gaps earlier, improving expense visibility, and escalating issues with better context. The biggest improvement is that AI moved us from “reporting what happened” toward “detecting what needs attention.”
Finance became faster and more proactive. The team now spends less time assembling information and more time acting on it: prioritizing overdue accounts, identifying billing gaps earlier, improving expense visibility, and escalating issues with better context.
How AI Streamlines the Finance Dunning Process
As an example, we are building an AI-powered dunning workflow.
Finance aims to manage collections within one application, eliminating the need to switch between billing data, aging reports, email, Slack, payment records, and customer notes.
AI identifies overdue accounts, reads the payment and communication history, prioritizes risk, and recommends the next action: reminder, escalation, banner, payment plan, or legal review. The system then generates the message, tracks the response, updates the case, and escalates when needed.
It all happens within Claude's business tier, which protects sensitive information and also has token spend.
Humans still approve sensitive actions, but the workflow becomes more automated, centralized, and proactive.
Why Financial Judgment Still Needs a Human Touch
That brings me to another point.
AI will power forecasting inputs, budgeting support, variance analysis, expense categorization, collections prioritization, billing exception detection, cash flow monitoring, and risk flagging. But human judgment remains required for budget approval, capital allocation, credit decisions, payment plans, legal escalation, investor communication, and runway tradeoffs.
The principle is simple: AI should surface risks, patterns, and options faster. Finance still owns the decision, accountability, and business tradeoff.
How AI Creates False Confidence If Data Infrastructure Is Subpar
Results are still early, since this is part of our roadmap and not yet fully scaled.
The main benefit, of course, is speed: less time compiling and cleaning data, and faster identification of billing gaps, collections priorities, expense issues, and variance drivers.
The qualitative improvement is focus: Finance can spend more time on judgment and escalation, and less time on first-pass analysis.
The main downside is data quality. AI can create false confidence if the source data is incomplete or poorly structured.
It also has not eliminated the need for clean systems and controls. If billing data, customer status, contracts, or expense categories are inconsistent, AI may produce faster analysis, but not necessarily better analysis.
In the end, AI is not the hard part. The hard part is clean data, clear ownership, and a well-defined process.
So, spend more time up front defining the data model, exception rules, approval points, and human review layers before trying to automate. If we had done that, we could have avoided a lot of rework and false positives.
How AI Redefines Financial Control and Forecasting
AI shows that control does not have to mean adding more checkpoints, spreadsheets, or approvals. In many cases, stronger control comes from better real-time visibility: detecting exceptions earlier, flagging unusual patterns, and routing issues to the right person before month-end.
AI also changed how we think about forecasting. The forecast should not be a static monthly exercise. It should become a living model that continuously updates as new signals come in.
Why CFOs Should Redesign Order-to-Cash for AI
CFOs need to redesign Order-to-Cash.
The goal should be an integrated, automated, and auditable flow from order creation to billing, collections, and cash application, so the business can free cash flow as quickly as possible.
The target result is simple: faster cash conversion, fewer missed invoices, cleaner controls, better auditability, and more Finance time spent on judgment instead of chasing data.
That means AI should help detect billing gaps, trigger dunning, prioritize collections, reconcile payments, flag disputes, and escalate risks — all inside one connected workflow.
We are redesigning this area so that Finance moves from manual tracking and reactive collections to proactive revenue control.
The target result is simple: faster cash conversion, fewer missed invoices, cleaner controls, better auditability, and more Finance time spent on judgment instead of chasing data.
How CFOs Can Leverage AI for Continuous Improvement
My advice? CFOs should record and document everything. In an AI-driven operating model, if something is not captured, it did not happen.
The biggest opportunity is empowering teams to do more with less: higher output, faster analysis, and more time for judgment, not manual preparation.
But the real destination is not just individual productivity; it is building continuous improvement loops where AI constantly observes the process, identifies friction, recommends fixes, and makes work easier and more efficient over time.
Finance should head in this direction: not just using AI as a tool, but designing finance operations that improve themselves 24/7.
Follow Along
Follow along with Fernando Blumenkron's work on LinkedIn.
More expert interviews to come on The CFO Club!
