Empty Spaces
Rebase (rebranded as Empty Spaces) deploys AI agents that automate month-end close workflows—reconciling accounts, investigating exceptions, and preparing journal entries continuously throughout the month. The platform runs single-tenant in the customer's own AWS account, with role-based permissions, human approval for ERP writes, and full traceability of every agent run. Finance teams review only judgment-based decisions, with evidence and source records attached to each conclusion.
- AI Agents
- Artificial Intelligence
- Enterprise Software
- Financial Technology
- Software Only
Funding
Founders
Founder details are not available yet.
Product
Problem
Traditional month-end close processes require finance teams to manually reconcile accounts, investigate discrepancies, and prepare journal entries, often leaving the bulk of analysis to be done under time pressure at period end. This creates bottlenecks, increases risk of errors, and delays the availability of timely financial insights for business decisions.
Solution
Rebase deploys AI agents that continuously reconcile accounts, investigate exceptions, and prepare journal entries throughout the month, keeping the close moving in real time. The agents surface material changes across revenue, margin, and cost, tracing drivers and attaching source records and calculations to each conclusion. Finance teams review only what requires judgment—explanations, evidence, and next actions already assembled—and can approve and post entries with a full audit trail. The platform learns from each run, capturing company definitions, policies, and approval rules to improve future handling. Rebase runs single-tenant in the customer's own AWS account or a dedicated one operated for them, with all data and compute isolated per customer.
Target Audience
Finance and accounting teams at mid-to-large enterprises seeking to automate month-end close processes, reduce manual reconciliation effort, and accelerate financial reporting cycles.
Features
- AI agents automate account reconciliation, exception investigation, and journal entry preparation across the entire month
- Material change detection across revenue, margin, and cost with driver tracing back to source records
- Evidence-linked conclusions showing period, entity, and calculation for every number and explanation
- Skills engine that encodes finance procedures (e.g., bank reconciliation, journal review, P&L variance analysis) with assigned-only usage and load-on-demand execution
- AI Gateway supporting any model provider with bring-your-own keys, per-access-group daily and monthly spend caps, and approved-model lists
- Role-based permissions with separation of duties—preparers cannot approve their own entries, and ERP writes require human approval bound to the exact payload
- Single-tenant deployment in the customer's AWS account with region choice, private subnets, and AES-256-GCM credential encryption that never exposes credentials to models
- Full traceability with logs of agent runs, model calls, tool calls, ERP writes, approvals, and all human configuration changes
- Read-only connection settings that prevent writes regardless of agent requests
- Isolated code execution in temporary containers with no credentials and no route into the private network
- Versioned agent instructions so teams can see what an agent was told when it performed work
- SOC 2 Type I and Type II compliance