Book a demo

ReconoSense: The AI native platform for private credit,
from underwriting to payoff

ReconoSense unifies your firm's siloed data and deploys AI to work across the full credit lifecycle: underwriting new deals, monitoring covenants, reconciling positions and managing liquidity.

Special SituationsAsset-Based LendingMerchant Cash AdvancesDirect LendingLeveraged FinanceReal Estate Lending

See how it works

0:00 / 4:02Capital Deployment
Read the Transcript

The film has no sound. This is its on-screen text, step by step.

Home

  1. 0:09
    Home

    The day starts here. The AI flags a new amendment for Harbor Logistics. Your day starts from what needs attention.

Capital Deployment

Use it when a signed amendment arrives and the deal team proposes an add-on to the loan.

  1. 0:14
    Step 1: Upload

    Drop the signed amendment into Document Center. The AI links it to its loan. Nothing to file or re-key.

  2. 0:21
    Step 2: Extract

    The AI reads four terms, each with its section and a confidence score. A reviewer approves them. Each term traces back to its page.

  3. 0:30
    Step 3: Reconcile

    The ledger still shows 8.00%. The amendment governs, so the Controller approves 8.25%. One fix, and every team reads the new rate.

  4. 0:40
    Step 4: Approve

    A $10M add-on passes five credit tests on checked figures, then goes to committee. The committee decides on terms that match the documents.

  5. 0:52
    Step 5: Settle

    Treasury checks bank details, cash and sanctions. A second person releases the wire. Operations books it and risk updates. The log keeps each step.

Covenant Breach

Use it when a borrower's quarterly compliance certificate comes in.

  1. 1:09
    Step 1: Ingest

    The Q3 certificate arrives by email. The AI reads it and ties it to the monthly financials. No analyst re-keys the certificate.

  2. 1:17
    Step 2: Test

    Covenants are tested against the amended agreement. Leverage is 4.62x against 4.50x: a breach. Caught the morning the certificate lands.

  3. 1:25
    Step 3: Simulate

    Scenario Lab runs 10,000 Monte Carlo paths and tests a covenant reset. See the range of outcomes before you decide.

  4. 1:35
    Step 4: Coordinate

    The AI drafts an action plan with one task for each team the breach touches. Every team acts on the same facts.

  5. 1:46
    Step 5: Act

    The Head of Credit approves the plan, and each task goes to its owner. One recorded approval sets six teams moving.

Voluntary Prepayment

Use it when cash lands in the bank that nobody has explained yet.

  1. 2:05
    Step 1: Identify

    A $7.5M wire hits the bank feed. The AI matches it to the prepayment notice. Unexplained cash is explained the day it lands.

  2. 2:13
    Step 2: Correct

    The administrator and the ledger still show $42.5M. The Controller approves $35.0M. Interest is recomputed before anyone bills the old balance.

  3. 2:22
    Step 3: Book

    The journal entry is drafted from the approved fix and posted. The administrator gets it too. The next invoice uses the new balance.

  4. 2:29
    Step 4: Deploy

    Distribute the cash to the holding company, or redeploy it? The AI recommends a new loan. The capital goes back to work within days.

  5. 2:39
    Step 5: Inform

    The deal team sees the breach cured and $7.5M ready to redeploy. Each team's view updated from one correction.

Vintage Analysis

Use it when the committee asks how past deals actually performed.

  1. 2:55
    Step 1: Connect

    ReconoSense pulls every loan, payment and fee from the loan servicing system through its API. No exports to clean up by hand.

  2. 3:04
    Step 2: Reconcile

    The servicing system's totals are tied to the ledger first. One amendment fee sat in the wrong account. The analysis starts from numbers that tie out.

  3. 3:12
    Step 3: Analyze

    2,148 deals, grouped by the year of first funding, with total funded, total interest and total fees. See which years earned the most per dollar lent.

  4. 3:23
    Step 4: Drill

    Open a year for its full breakdown: paid off against active, by sector, deal size and structure. Find what drove a year's results.

  5. 3:32
    Step 4: Drill

    All 334 deals in the vintage, each with its own totals. Switch to 2019 to compare. Check any figure down to the deal.

  6. 3:42
    Step 5: Share

    The analysis goes into the committee pack, each figure cited to the servicing system and the ledger. The committee prices new deals on real history.

Founders' Previous Experience

  • Morgan Stanley
  • Deutsche Bank
  • Lockheed Martin
  • ARK Invest
  • TWG Global
  • Cresset
  • CCaventus
  • CGI
  • Cybermonic
  • Apprentice
  • Gromelski & Associates

Prior employers of the ReconoSense founders. These firms are not ReconoSense customers, partners or endorsers.

Meet the founders

No firm data goes to outside AI companies

The models run in your own cloud, on your premises or in a single-tenant environment we host for you. Access follows each person's role.

From underwriting to payoff.

ReconoSense helps private credit teams close deals faster and catch problems earlier, with AI working from one shared record instead of manual workflows across disconnected systems.

  1. 1 of 6

    Underwrite

    Underwrite with the whole picture.

    From originations to investment committee, credit agreement terms flow straight into the record your portfolio team monitors.

    • Document Extraction

      AI reads credit agreements and deal documents, links every extracted term to its source clause, and a person reviews each one.

      Interest Rate8.25%§1
      Commitment$42,500,000§2.1
      Max Total Leverage4.50x§7.2
    • Find Policy Rules That Misfire

      Show which rules drive overrides, declines or uneven outcomes so lenders can refine policy based on actual performance.

      Rule 12 · overridden in 9 of 14 approvals

    • Test Repayment Under Realistic Shocks

      Estimate how a borrower could manage a downturn, delayed customer payments, higher rates or the loss of a major contract, using its actual cash-flow patterns.

      DSCR 1.82x to 1.31x under three shocks

    • Compare Proposed Loan Terms

      Show how changing the amount, term, collateral or repayment schedule affects borrower resilience and lender risk.

      Proposal AProposal B
      Amortization1% a year5% a year
      Stressed DSCR1.31x1.44x
      Expected loss1.9%1.2%
      Value, intrinsic lens99.199.6
    • Underwrite Connected Businesses Together

      Detect shared owners, cash flows, guarantees and obligations across related entities, so each company is underwritten with the group it belongs to.

      Group exposure $54.5M · 9.1% of NAV

    • Turn Approval Conditions Into a Tracked Plan

      Convert conditions such as reducing debt or delivering updated financials into clear milestones, and flag when each is met or at risk.

      3 conditions · each with an owner and a date

  2. 2 of 6

    Close and Fund

    Fund on the terms you approved.

    Approvals follow your delegated authority, and the approved terms become the position without anyone retyping them.

    • Delegated Approvals

      Each approval goes to the person your authority matrix names, and conditions are tracked to closing.

      Approved with 3 conditions · Head of Credit

    • Funding Wire

      Treasury checks bank details, cash and sanctions, and a second person releases the wire.

      $42,500,000.00 · released by a second approver

  3. 3 of 6

    Monitor

    See a breach the morning it arrives.

    Certificates, borrower reports and agent notices are read on arrival, tested against the terms set at underwriting and reflected in each risk rating and the watchlist.

    • Covenant Compliance

      Each covenant is tested against the agreement as amended, with a warning before a breach.

      Leverage 4.62x against 4.50x · flagged 08:14

    • Borrowing Base

      Borrower reports update eligible collateral and availability, with your haircuts applied.

      Eligibility haircuts applied on arrival

    • Position Reconciliation

      Agent notices, the administrator and your ledger are tied out field by field.

      Break found and corrected the same day

    • Risk Management

      Each borrower's risk rating and probability of default update as its reports arrive, with every change kept.

      Harbor Logistics · rating 4 to 5 after Q3

    • Watchlist

      A credit that trips an early warning signal or a covenant test joins the watchlist with an owner and a review date.

      3 loans · each with an owner and a review date

  4. 4 of 6

    Work Outs & Recoveries

    Move every team on one plan.

    When a credit moves, the breach is sized, provisions are calculated, one plan sets a task for each team, and the agreed terms update the record.

    • Size the Breach

      Scenario Lab runs thousands of paths and tests a covenant reset before you negotiate.

      91% of 10,000 paths within a 4.75x reset

    • One Action Plan

      AI drafts a task for each team the breach touches, and the Head of Credit approves it.

      6 teams · 1 approval

    • Impairment and Provision Calculations

      Expected losses and specific provisions are calculated from the borrower's rating, collateral and recovery assumptions, each input traced to its source.

      Provision $1.2M · 2.8% of exposure

    • Recoveries Traced

      Every amendment, draw, paydown and fee stays traced to the document it came from.

      Amendment No. 3 · covenant reset to 4.75x

  5. 5 of 6

    Portfolio Management

    Report one set of numbers.

    Exposure, liquidity, vintage performance and returns are read from the same record, so every team reports the same figures.

    • Exposure and Limits

      Exposure by borrower, sector and rating, with each limit tested.

      Harbor group 9.1% of NAV · limit 10.0%

    • Liquidity

      A 14-day cash projection, with facility headroom and capital free to deploy.

      $18.4M free to deploy

    • Vintage Analysis

      Deals grouped by year of funding, with funded amounts, income and fees for the committee pack.

      2,148 deals by vintage

    • IRR & MOIC Analysis

      Realized and projected IRR and MOIC by loan, vintage and fund, from the cash flows on the record.

      TL-0142 · IRR 9.4% · MOIC 1.21x

  6. 6 of 6

    Pay Off

    Close the loan and keep its history.

    The payoff is matched to its notice, interest is tied out to the rate on the record, and the position closes.

    • Payoff Matched

      The incoming wire is matched to the payoff notice, principal and interest.

      $42,792,187.50 · matched to its notice

    • History Kept

      Every term, test, approval and payment stays on the record for the audit and the next vintage analysis.

      One export for the auditors

Figures are sample data.

Connect Any Source

ReconoSense connects to virtually any data source through its multi-modal ingestion layer: APIs, cloud storage, file feeds, document repositories and direct uploads inside the platform. Data arrives in many formats. ReconoSense cleans, standardizes and translates it into one common data language, so every piece of information fits into a single master picture.

  • Idempotency

    Loading the same file twice never creates duplicates.

  • Strict Schema Validation

    Every record is checked against the expected structure before it enters. Bad records are stopped and flagged.

  • Unified Security Master

    Each instrument has one identity across every system that reports it.

The platform runs in your firm's environment and pulls from connected sources automatically. Your team can also upload files inside the platform. The public website is never where your data goes. See the sources it connects.

The problems we built it for.

See it with your own numbers.