Kela Blog

Deal Intelligence, Decoded

Perspectives on AI-powered deal workflows, virtual data rooms, and what it means to run a smarter process.

July 27, 2026 · 6 min read

Your books shouldn't wait for your administrator to tell you they're right.

Most funds only find out whether their books match their administrator's once a delivery arrives — monthly, quarterly, always after the fact. FundOS now computes its own trial balance daily and reconciles it against the administrator automatically, classifying every break by cause. The administrator stays the official book; the fund gets an independent, always-current second opinion.

July 24, 2026 · 5 min read

Who to call today — and what your companies did overnight.

A fund's edge was never more data — it's attention pointed at the right thing at the right moment. We shipped two features that turn signal you already own into a daily answer: the LP Intent Radar ranks which LP is warm this week (from your own data-room views, email opens, and meetings — not third-party tracking), and Portfolio & Pipeline News surfaces what your companies did overnight, relevance-filtered so a startup named “Sherpa” never buries you in mountain-guide headlines.

July 24, 2026 · 7 min read

The model can be wrong. The decision it feeds can't be.

Financial institutions want AI in their decision workflows — risk, credit, valuation, compliance, reconciliation. They can't deploy it, because a model that's usually right is unacceptable where decisions carry regulatory and fiduciary weight. The answer isn't a more deterministic model — that doesn't exist. It's a deterministic enforcement layer that validates every AI output against a structured policy graph before it reaches the person making the call. We built it. It's called the Governor.

June 10, 2026 · 5 min read

The next fund manager won't Google your software. They'll ask Claude.

AI answer engines — Claude, ChatGPT, Perplexity — have replaced Google for how fund managers evaluate software. Most fund tech vendors have no strategy for this. Here's what GEO (Generative Engine Optimization) is, why it matters more than SEO for B2B fintech, and what we built to be visible where it counts.

June 2, 2026 · 4 min read

Your fund needs a brain, not more connectors

Funds spent a decade connecting every system to every other system. The data is all in one place now — and funds are no closer to understanding what's happening inside them. Connectors move data. They don't resolve it. Here's the difference between access and understanding, and why your fund needs a brain.

May 14, 2026 · 4 min read

Your annual audit shouldn't take three weeks of email.

Your auditor asks for the same ten things every year. An auditor data room built for a private fund — trial balance, capital account roll-forward, fee reconciliation, waterfall — turns three weeks of email into a single secure link.

April 29, 2026 · 4 min read

Nobody writes a valuation policy until an examiner asks for it.

Your fund marks Level 3 positions every quarter, but the valuation policy explaining how usually doesn't exist until an examiner asks. Here's what a private-fund valuation policy has to cover — and how to generate a defensible draft from your own fund data.

April 9, 2026 · 4 min read

You already know what the SEC will ask for.

An SEC exam is a books-and-records document request, not a surprise — the list is largely knowable in advance. Here's the readiness checklist private fund advisers are graded on, and how to know your status on each item before the letter arrives.

March 24, 2026 · 4 min read

The red flags are already in the data room. Nobody reads 4,000 pages.

The red flags in a deal are almost always already in the data room — buried in thousands of pages nobody has time to read end to end. Here's the operational due diligence red-flag checklist, and how an AI agent surfaces them with an entity map and risk summary.

March 5, 2026 · 4 min read

The quarterly letter is the most expensive document in private equity.

The quarterly LP letter is the most expensive document a fund produces — and most of the cost is the per-LP capital account statement (PCAP) underneath it. Here's what LPs look for in a capital account statement, and why reconciling it by hand every quarter is the real bill.

February 17, 2026 · 4 min read

Your private credit LPs want loan-level detail. Your reporting stops at the fund.

Private credit LPs increasingly expect loan-level reporting — delinquencies, recoveries, position-by-position performance — not a fund-level summary built for equity. Here's why the bar moved, and what loan-level LP reporting actually requires.

January 30, 2026 · 4 min read

Your LPA is already out of date

The carry on Fund II isn't in the LPA. It's in the LPA plus Amendment 1 plus a side letter plus Amendment 2, read together in the right order — reassembled from memory every quarter. Here's why your fund's real terms decay the day the ink dries, and what it takes to actually know them.

January 13, 2026 · 4 min read

Agents don't replace your CFO. They replace your CFO's Monday.

The fear is that AI comes for the CFO's job. It doesn't. It comes for the hour of reassembly before the job starts — the reconciling, the threshold-checking, the what-happened-over-the-weekend — performed every Monday by the most expensive person in the building. Here's the difference between a chatbot that waits and an agent that already ran.

December 19, 2025 · 7 min read

How to connect Claude Code to your fund management system using FundOS MCP

FundOS ships a Model Context Protocol server with 24 tools covering every fund module. Connect Claude Code in one line, then ask it anything about your fund.

December 2, 2025 · 6 min read

Your Data Room Shouldn't Just Store Documents. It Should Work the Deal.

Traditional VDRs are filing cabinets with a permission layer. Kela is something different — an AI participant in your deal that reads every document, flags every risk, and drafts every email.

June 3, 2026 · 5 min read

Month-end close shouldn't take a week.

Every fund CFO knows the month-end fire drill: reconcile the prime broker file, gather journal entries, compute NAV, look up each LP's individual terms, run the waterfall in a spreadsheet that hasn't been properly versioned since Q3, then issue allocation statements. Here's why that week doesn't have to exist.

December 10, 2025 · 5 min read

The compliance moat: why enterprise AI agents built for security win the Fortune 500.

When Elise AI started building on-prem, self-trained LLMs, most investors saw a constraint. It turned out to be the product. Here's why the compliance-first approach to enterprise AI agents is the moat that wins the Fortune 500.

Founder, Eight Capital

January 7, 2026 · 5 min read

Why a profitable consumer privacy company is a venture bet worth making.

Optery removes your personal data from data brokers. It has $12M ARR, is profitable, and raised only $9M. Most venture investors passed because it looked like a feature. Here's why the thesis was right and what it tells us about regulatory tailwinds.

Founder, Eight Capital

January 21, 2026 · 4 min read

422% growth in a year: what Taiv's numbers tell us about AI in physical spaces.

Taiv installs AI-driven screens in venues and replaces generic cable content with targeted advertising. 422% revenue growth in 2024, $7.9M run rate, 264% net dollar retention on advertisers. Here's why the unit economics work and why physical spaces are the next AI frontier.

Founder, Eight Capital

Tax & Investing

February 4, 2026 · 4 min read

When does the QSBS 5-year clock start on a SAFE?

Investors in YC companies via SAFE often assume the Section 1202 QSBS 5-year clock starts when they signed the SAFE. It doesn't — it starts when the SAFE converts to equity. If your SAFE converted in 2022-2023, your 5-year clock may expire years later than you think.

Founder, Eight Capital

February 18, 2026 · 4 min read

What a YC LP co-invest actually looks like.

Most LP co-invest programs involve late-stage deals. YC co-investing is different — it's about access to founders who won't return cold calls. Here's how it works in practice alongside an early YC fund.

Founder, Eight Capital

March 5, 2026 · 4 min read

The $10M ARR in 36 months filter: why it separates real YC companies from the rest.

Can you reach $10M ARR within 36 months of launch? It's the question Eight Capital asks every YC founder. Not because every company will hit it — but because the answer reveals whether the market, the motion, and the founder's grasp of the business are real.

Founder, Eight Capital

March 18, 2026 · 4 min read

Why we invest in 8 consecutive YC batches instead of cherry-picking our best years.

Most investors try to pick YC vintages — more exposure to 'hot' batches, less to quiet ones. The problem is you can't know which batch produces a unicorn until a decade later. Covering 8+ consecutive batches smooths the power law and statistically guarantees you're in the high-output years.

Founder, Eight Capital

April 1, 2026 · 4 min read

The consensus YC pick rarely becomes the biggest winner.

The companies every investor talks about at Demo Day are already overpriced and over-allocated. The breakout YC companies — Airbnb, Stripe, DoorDash — all had detractors at seed. Here's why contrarian bets outperform consensus in early-stage YC investing.

Founder, Eight Capital

April 15, 2026 · 4 min read

Founder market fit is not founder-product fit. Here's the difference.

Founder-product fit means the founder can build the product. Founder-market fit means the founder is the right person to win in that market. The second is harder to evaluate and more important to returns. Here's how Eight Capital assesses it.

Founder, Eight Capital

April 22, 2026 · 5 min read

Pro-rata rights are the whole game in YC investing.

Pro-rata rights give you the right to maintain your ownership in future rounds. In YC investing, where the biggest outcomes come from a handful of breakouts, the ability to double down on your winners is the entire return driver. Here's why Fund II increased check size to secure them.

Founder, Eight Capital

May 6, 2026 · 5 min read

YC's 6.25% unicorn rate vs. 2.5% in broader venture — what the gap actually means for returns.

Y Combinator produces unicorns at 6.25% — roughly 2.5x the rate of broader VC. In top batches, the rate reaches 12%. Here's why the gap exists, what it means for expected returns in a 20-company-per-batch portfolio, and why the statistical advantage compounds.

Founder, Eight Capital

Tax & Investing

May 19, 2026 · 5 min read

QSBS stacking: write 10 checks, potentially exclude $100M from federal taxes.

Section 1202 provides up to $10M in federal capital gains exclusion per qualifying investment — not per year, not per investor. Investors who make 10 qualifying YC investments can potentially exclude $100M in federal capital gains. Here's how stacking works and what limits it.

Founder, Eight Capital

May 27, 2026 · 5 min read

Why we invest in Week 2, not Week 10.

By Demo Day, the best YC companies are already allocated. The founders chose their early investors in weeks 4-8 — the investors who met them early, gave useful feedback, and committed when there was no FOMO. Here's what the eight weeks before Demo Day look like from inside the process.

Founder, Eight Capital

Tax & Investing

June 3, 2026 · 5 min read

The best tax break in venture is hiding in plain sight.

Section 1202 of the tax code lets you exclude up to 100% of capital gains — up to $10M per investment — from federal taxes if you hold qualifying startup stock for five years. YC Demo Day companies check every single box. Here's why, and what could still disqualify you.

Founder, Eight Capital

June 9, 2026 · 4 min read

Your fund was supposed to go live six weeks ago.

The gap between signing a fund software contract and going live is usually three to six months. It's not because the work is hard. It's because nobody asked the right questions early enough. The AI FDE interviews your fund in 30 questions and configures FundOS automatically — no consultants, no kickoff calls, no requirements documents.

June 9, 2026 · 5 min read

The questions that never get asked.

The questions that matter most in fund operations are the ones nobody thinks to ask. The capital call that sat overdue for two weeks. The KYC that expired quietly. The compliance filing that slipped because everyone assumed someone else was watching. Fund Intelligence is the AI analyst that asks those questions every morning — and only tells you something when it actually matters.

Ready to run a smarter deal?

Join teams using Kela to close faster with AI-powered due diligence.

Book a walkthrough