Building a Scalable Finance Function for a Fusion Energy Startup

The Client

A fusion energy startup developing compact, modular fusion technology designed to make fusion more accessible and commercially viable. Founded in 2018, the company is developing desk-sized fusion machines using magneto-electrostatic confinement, enabling rapid engineering cycles and the ability to combine high-energy-density units for a range of power applications.

Beyond clean energy generation, the company is developing fusion technology for applications including space propulsion, defense, distributed energy grids, neutron sources, and advanced materials testing. With approximately $69 million in funding from notable investors, the company is working to advance the commercialization of small-scale fusion systems.

As the company scaled, its finance function needed to support a capital-intensive, highly technical business while remaining lean. The company partnered with Echo Park Consulting (EPC) to build a more structured, automated finance operation capable of scaling without requiring a significant increase in headcount.

Legacy System: QuickBooks

The Challenge

The company had no structured month-end close process in place. With a finance function consisting of just one person and growing to two, the business needed stronger financial processes and controls that could scale with complexity—not simply with headcount.

Several challenges were creating significant operational strain:

• Multi-currency vendor bills were flowing through a spend-management tool without standardized checks.

• Foreign-currency synchronization issues were creating additional reconciliation work.

• Expenses were not consistently tagged by department or project.

• US government grant reporting required detailed cost breakdowns by department and project milestone, previously requiring separate spreadsheet rebuilds.

• There was no standardized month-end close process or checklist.

As a result, the month-end close could take two weeks for a single finance lead, and up to a month when audits required additional reporting and reconciliations.

The company needed a finance infrastructure that could introduce structure, automate repetitive work, improve data quality, and support increasingly complex reporting requirements, all without significantly expanding the finance team.

The Solution

EPC built a structured month-end close process in Rillet from the ground up, establishing a 43-task close checklist covering pre-close, close, and review.

EPC also introduced an AI-enabled layer across the finance function, combining Rillet Aura AI with a custom EPC-built Claude skill. This allowed a lean finance team to automate routine analysis, identify anomalies, and prepare work for human review.

Six automated Aura workflows were implemented to support the close process:

• Staff cost allocation

• Missing department alerts

• Project code completeness checks

• Automatic accruals for bills over $2,000

• Post-close flux analysis for movements greater than $10K or 25%

• Monthly P&L reporting by R&D and G&A

EPC also developed a custom Claude skill that combines Aura output with general ledger, journal entry, and integration data to generate US government grant reporting directly from the ledger. This eliminated the need to rebuild grant reports manually in spreadsheets and established the general ledger as the single source of truth for both financial reporting and grant reporting.

EPC also addressed payroll posting schedules at the source, creating cleaner and more consistent compensation postings.

EPC Services

• Month-end close design and implementation

• ERP implementation and close checklist development

• AI workflow implementation

• Custom AI skill development

• Financial process optimization

• Government grant reporting automation

Technology Stack

• Rillet — ERP / General Ledger

• Rillet Aura AI — Automated workflows and financial analysis

• Spend-management platform — Vendor and expense management

• Claude — Custom EPC-built AI skill

• QuickBooks — Legacy system

The Results

EPC transformed a largely manual finance operation into a structured, highly automated four-day month-end close—giving a lean finance team the infrastructure to manage significantly greater complexity.

Month-end close reduced to 4 days
The close timeline dropped from approximately two weeks to four days, with audit periods that previously could extend the process to a month also significantly streamlined.

60–80% reduction in close time
The new process reduced the close timeline from approximately 14–30 days to just four days.

One finance lead operating like a larger team
AI-powered workflows enabled the finance lead to shift from manual execution to reviewing and posting AI-prepared work.

Automated financial controls
Pre-close data-quality checks now identify missing department tags, missing project codes, accrual requirements, and significant financial variances before they become larger issues.

Grant reporting directly from the ledger
US government grant reporting now derives from the same general ledger used for the company's financial books, eliminating manual spreadsheet rebuilds and creating a single source of truth.

Improved financial data quality
Staff cost allocation, payroll posting schedules, project coding, and foreign-currency synchronization are now monitored through structured processes and automated checks.

Key Takeaway

By combining a structured finance process with Rillet and an AI-powered automation layer, EPC helped a lean finance team build the infrastructure of a much larger finance organization—reducing the month-end close from weeks to days while eliminating manual reporting processes and creating a scalable foundation for continued growth.