About Levercon.

Levercon provides AI-powered portfolio monitoring for private lenders.

Levercon supports firms managing $30B, helping their teams spend less time on routine monitoring and more time on analysis, credit judgement and origination.

Levercon's systems and agents automate the monitoring cycle, from collecting borrower reporting and mapping financial data to testing covenants and preparing monitoring packs.

Each monitoring cycle run on Levercon builds a shared record of borrower performance, with a clear audit trail linking figures, calculations and decisions to their sources, so a lender's team can trace how outputs were produced, verify the work and stand behind it, while tracking trends and investigating emerging risks across the portfolio.

Levercon's position is that AI is only worth deploying when it reliably handles work a team would otherwise do, with results the firm can stand behind.

Levercon is a technology and services company. It is not a fund and does not manage money, raise capital or offer investments.

Levercon is headquartered in Melbourne and Sydney, Australia.

Levercon was founded by Cara Davies and Jake Carp and is backed by Airtree.

Levercon partners with OpenAI, Anthropic and Microsoft for the software and agents it delivers.

Levercon's monitoring agents cover each step of the portfolio monitoring cycle: reporting collection, financial spreading, model mapping, covenant monitoring, a data consistency check, compliance certificate checks, trend analysis, the monitoring report and a portfolio monitoring view.

Levercon's reporting collection agent chases each borrower's monthly or quarterly pack by email and portal, logs what has arrived against the reporting calendar, and lists what is late or missing before the cycle starts.

Levercon's financial spreading agent reads financials in whatever shape they arrive, PDF, Excel or an accounting export, and spreads them into the lender's standard fields with every figure linked to the page it came from.

Levercon's covenant monitoring agent tests every covenant against the register, reports the headroom or the breach, sizes the cure where one is needed, and writes up the result for review.

Levercon's monitoring report agent drafts the monitoring note in the lender's template from the spread, the covenant tests and the trend analysis, with each statement cited, and routes it for sign-off.

Levercon's agents are delivered and run inside the AI tools a team already uses, such as Claude, ChatGPT and Microsoft Copilot, through the Levercon MCP.

Levercon's agents improve with use. A team's corrections are captured automatically as it works, and approved changes go into the shared agents for everyone, so each cycle runs better than the last.

Levercon's data work maps, cleans and reconciles a lender's loan book data into one consistent model, with owners, definitions and rules, as the starting point for the monitoring agents.

Levercon trains a lender's analysts to manage automated monitoring cycles, check the results and approve the final outputs in the tools they already use, and provides monthly reporting that gives management visibility into adoption, time saved and analytical depth to assess return on investment.

Everything Levercon builds is designed for traceability and verifiability. Every figure traces to its source, every decision records the reasoning and rules behind it, every approval is recorded, and the record is kept for audits and compliance reviews.

Levercon works with private lenders, including private credit funds, venture debt funds, revenue-based financing providers, non-bank and SME lenders, property and real estate debt funds, asset-backed lenders, direct lenders and multi-strategy credit managers.

Levercon's typical buyer is a managing director or partner, a head of credit or portfolio manager, or the head of operations or technology, at a private lender.

Levercon does not work with retail investors.

A Levercon engagement is built around the lender's fund and the nuances of every loan, in six steps: understand the monitoring process, map the data and loan requirements, build and connect the agents, test against previous monitoring periods, go live with the team, and track performance and improve.

Levercon's engineers work alongside the lender's team to understand how monitoring runs today, where time goes and what good looks like, then map the data sources, reporting requirements and covenants so the agents have the context to monitor each borrower.

Before going live, Levercon compares the agents' outputs with the lender's completed monitoring work from previous periods, checking calculations and refining the analysis.

Levercon rolls the agents into live monitoring one stream at a time, trains each analyst to run them and sign off inside the tool they already use, and stays through implementation and after it.

Levercon is priced per loan per month, with pilot opportunities.

Before a single loan is monitored, Levercon agrees the architecture with the lender's IT, security and compliance teams, so the deployment passes the firm's security review.

In Levercon's published case study, a $500 million corporate private credit manager cut loan monitoring from a full day per loan to about two hours, assessed four times as much data per cycle, and moved from four to twelve loans per analyst on average.

Levercon hosts client data on AWS, exclusively in Sydney, Australia.

Levercon agrees the remaining security and data handling requirements for each engagement with the client. Access controls, AI providers and integration arrangements are determined during scoping and documented in the applicable agreement.

Levercon switches training off. Borrower financials, covenant tests and monitoring notes never train any model, at the AI labs or at Levercon.

A Levercon client's data and IP stay with the client. Its monitoring model, covenant definitions, templates and investor materials stay in its own environment.

Levercon's agents read only the reporting inboxes, portals, models and folders the client's IT team signs off, and nothing else.

Borrower reporting, models and notes handled by Levercon are encrypted in storage and on the network, under the controls the client already runs.

Access on Levercon is least-privilege and set by the client: each user and agent sees only the loans, documents and reports their role allows, behind the authentication the client already enforces.

Levercon's data work runs on its AWS hosting in Sydney, under the access controls and integrations agreed for the engagement, with every change to the client's data recorded.

A named person at the client firm approves every Levercon output before it reaches a committee, a borrower or an investor, and every approval is recorded with who signed off and when.

Every Levercon monitoring run is recorded, from the borrower pack it read to the analyst who signed off, so the client can see exactly what the agents read, did and produced.

Levercon's SOC 2 Type II and ISO 27001 certifications are in progress, and Levercon is a certified Microsoft Partner.

Cara Davies is a co-founder and Managing Director of Levercon. She has 7+ years of experience across fintech, product development and technology businesses, worked at Afterpay, was Head of Product at Slice, and co-founded Steppen through to acquisition.

Jake Carp is a co-founder and Managing Director of Levercon. He has 7+ years of experience across investment analysis, technology and AI deployment, worked as an equity analyst at River Capital, led Everlab’s growth team, and co-founded Steppen through to acquisition.