BTG
June 2026
Pre-Read
Prepared for Boaz Shedletsky · A conversation, not a raise
A private credit thesis built from the inside: where manual workflows, fragmented legacy software, and unstructured lender data create the next operating layer for an increasingly non-bank market.
The team
BTG is built by three founders who came to this problem from inside the credit world. The thesis did not begin on a whiteboard. It began with seeing, first-hand, how much of a lender’s daily work still runs by hand, and asking why no one had closed the gap.
Yuval Rabina, CEO
Direct experience from the decision-making table of a private credit firm, focused on solving the industry’s operational bottlenecks from the inside.
Bar Ashkenazi, COO
An entrepreneur who runs end-to-end business operations, project development, and cross-border transactions.
Nevo Strauss, CTO
A software engineer specializing in platform architecture and scalable backend infrastructure.
We are clear-eyed about our one gap: none of us is a career credit professional. What we carry instead is the operational reality of our backgrounds, two combat officers from Sayeret Matkal and a Chief Engineer from the Navy’s Submarine Flotilla. We operate with a mindset built on hard execution, deep planning, and a real read on both people and systems.
The market, and why now
48% → 29%
Bank share of U.S. corporate lending over the last decade.
Federal Reserve, 2026
$1.3 to $1.4T
U.S. private credit, up about 5x since 2009.
Federal Reserve, 2026
Capital has migrated, permanently, from banks to non-bank and private lenders. In its place, U.S. private credit is now close to a third of the entire leveraged credit market, and in commercial real estate alternative lenders already close around 37% of deals. It is still early: private credit is only about 10% of total U.S. corporate borrowing today, against a middle market of nearly 200,000 companies.
This market looks nothing like the one traditional financial software was built for. Private credit is structurally decentralized: capital flows through hundreds of independent funds and dozens of business development companies, a sector that alone holds roughly $450 billion across some fifty public vehicles and many more private ones.
Who we are testing this with
This is not theory. We are running rigorous discovery and validation cycles across the credit ecosystem, lifting the hood on how lenders actually operate and pressure-testing our architecture against the realities of the market. This is unglamorous, ground-level work.
The operators
Every seat inside leading credit companies.
The executives
CEOs and board members of Israeli lenders managing billions.
Capital providers
Funds deploying capital, including one near $4 billion.
Tech ecosystem
Public AI networks facilitating about $40 billion in loans.
The borrowers
Mapping friction across the full loan lifecycle.
The pains we discovered
Across these cycles, four structural pains surfaced again and again. Each is a large standalone bottleneck in the private credit lifecycle, and each points to a direction on the roadmap.
Manual workflows in a high-stakes environment.
Portfolios worth hundreds of millions run on shared folders and manual process, with no infrastructure to scale or defend the book.
Validated by
Institutional capital providers
Where it points us
Agentic-ready infrastructure. AI agents that automate credit monitoring and back-office work, so lenders scale without linear headcount.
A fragmented legacy stack.
Systems do not talk to each other. The real work spans far beyond an LOS or CRM, leaving disconnected data and broken workflows.
Validated by
Fund CEOs and financial board members
Where it points us
An AI-native operating layer. One connected, queryable surface on top of the existing stack, unifying the loan lifecycle.
Zero secondary liquidity.
Once a loan is funded it is illiquid. There is no clean way to price it, move it to free up capacity, or step out of a souring deal.
Validated by
Fund CEOs
Where it points us
Secondary-market infrastructure. A way to price notes and trade loans between independent private lenders.
Institutional amnesia.
Lenders do not learn from their own history. They lean on generic outside models, so the same underwriting and risk mistakes compound.
Validated by
End-to-end market validation
Where it points us
A proprietary intelligence layer. A retrieval and prediction brain trained on the firm’s own book.
The root problem, and the convergence point
These are not four distinct problems. They are four symptoms of one breakdown in how a lender’s data flows and connects. The data needed to run a private credit business, let alone to cross-examine deals for secondary trading, is unstructured, trapped, and unformalized. It stays scattered across disconnected systems, where it cannot be queried, reused, or leveraged across the firm.
Unstructured, trapped, unformalized data
Deals
Notes
Risk
Memos
Spreads
The convergence point
Formalization & connectivity
Agentic infrastructure
AI-native layer
Secondary market
Intelligence brain
Four directions it unlocks
The ask
We are not raising. What we value most from this conversation is the chance to walk you through it in depth: the validation process we are running, the four directions we are weighing, and our immediate next steps. From there, your honest, unvarnished read on all of it.
The most useful thing you could give us is raw criticism, the kind that pokes real holes in the thesis and sends us back to the drawing board: which direction is sharpest, where the strongest institutional pull is, and what we should be trying hardest to disprove next.
BTG · Pre-Read
Boaz Shedletsky · June 2026