Voice Settings Back to Home

Building the Goldman Sachs for Small Businesses with Ori Eldarov

Generated Posts

26 posts
Legacy data vendors are flexing 100M+ records—but can’t tell you what 90% of those businesses even do. We built AI agents that actually visit company sites, analyze offerings, reason through relevance, and walk away only when the job’s done. Full-stack deal intelligence, not fluff.
Most investors are flying blind on deal flow. They think there are 10,000 targets. After our AI filters for size, product mix, and exclusion logic, it’s often 500 max. Sourcing is broken because legacy search is dumb. Ours reasons like an analyst. At scale.
Why do we need BDR teams buried in CRM setups and admin hell when AI agents can do 90% of the pipeline work—better, faster, and without a calendar invite gone wrong? M&A isn’t just being automated, it’s being rebuilt. From sourcing to SIMs to NDAs, AI runs the stack. This isn’t ‘productivity tooling’—this is ownership.
Business owners don’t need to care about AI. They care about results. That’s why OffDeal lets you drop your company’s URL and get matched with buyers in seconds. No fluff. Just names.
Most small businesses get the worst end of the M&A deal. Not because they’re bad companies—because the model is broken.    

In sub-$10M EBITDA deals, fewer than 20% go through a formal auction. That means most owners aren’t choosing from multiple offers. They’re hoping one buyer follows through. Spoiler: many don’t.    

While buy-side advisors flood the space with outreach and tech-enabled processes, sell-side remains hyper-fragmented—1 to 3 person teams, analog tools, zero scale. Selling a business should be a real process, not a hail mary.    

Our bet is simple: turn sell-side advisory into a tech-enabled, AI-native engine. Map workflows. Automate the busywork. Bring national scale to a market of local brokers.    

We’re not building a better data tool. We’re rearchitecting the investment bank from the ground up for owners—not deal hunters.
Most firms say they have the data. Few can actually prove it.

Legacy platforms boast 100 million+ companies indexed, but anyone serious in the middle market knows that math just doesn’t add up. There are only ~33 million businesses in the US. Half are sole proprietors. When you add deal constraints like revenue thresholds, staffing, specific services, or ownership models, target lists shrink from thousands to hundreds.

We built our own data engine—an AI agent that quality-checks itself, navigates websites like an analyst, and builds structured company profiles at scale. Then we layered on custom-built semantic search using embeddings, giving investors precision they’ve never had. You don’t just get ‘pest control,’ you get ice blasting, sand blasting, dry chemical applications—plus the filters to exclude everything that doesn’t matter.

This is not a sales pitch. We’re not selling the dataset. Because when everyone has access to the same tools, there is no edge left.

This is what targeted deal origination looks like when AI is actually built for the job.
Search is not solved. At least not when you're looking for buyable businesses.

Most investors assume they’re building a platform with thousands of potential add-ons. But the moment you get specific—industry, service mix, revenue profile, ownership type—that 10,000 turns into 200 real targets, fast. 

OffDeal's AI doesn’t just scrape the web. It recreates a reasoning analyst—cataloging products, customers, ownership details—and then layers conceptual search filters that actually understand what you're looking for. 

It’s how we surface not just ‘power washing’ companies, but the ones doing ice blasting, dry blasting, and serving commercial rather than residential clients. Precision like that doesn’t come from NAICS codes. 

That’s why we built our own embedding models. And that’s why our clients aren’t wasting time chasing ghosts.
Most firms still rely on a patchwork of software to run M&A ops at scale. Meanwhile, we’ve built a system of AI agents that do everything from prioritizing leads to scheduling meetings to chasing financials—and they never sleep. No virtual assistants. No bloated admin teams. No legacy tools duct-taped together. This isn’t just process optimization. It’s a redefinition of how you run the M&A pipeline from first touch to close. The right agent gets the right data to the right human at the right time. You don’t need Salesforce consultants. You need systems that actually work. That’s why we built our own.
AI isn’t the headline — outcomes are.

Most business owners couldn’t care less about how a deal is sourced. They just want to know: can you get results?

That’s exactly why OffDeal hides the tech behind a clean interface. You drop in your company website, and in under 30 seconds you get a list of buyers interested in your business. No pitch decks. No jargon. Just answers.

Behind the scenes, multiple AI systems are working in parallel: one reads your company’s website like a human would, while another crawls tens of thousands of private equity transactions to find matching investors.

The future isn’t about showing off the algorithm. It’s about who can deliver real value, now — without expecting the owner to become a technologist.

We think AI will change how M&A deals get done. But we’re already giving business owners a taste of that change today.

Why the Future of Investment Banking Is Sell-Side First and AI Native

The vast majority of M&A advice, software, and data tools have been built to serve the same audience: buyers. Private equity firms and their armies of buy-side advisors have had the pick of the toolbox, leaving sellers with little more than cold emails and a desperate hope someone finds them. But what if the biggest opportunity in lower middle market M&A isn’t helping buyers find deals — it’s fixing the broken seller experience?

At OffDeal, we started with the obvious: legacy data vendors like Grata and SourceScrub hadn’t evolved. But we quickly realized the real problem wasn’t just data quality — it was asymmetry. Buyers benefit from every refresh of the pipeline, but sellers burn out after one failed deal. They don’t want to hear from one more PE associate who might kill the deal at the altar.

Every failed deal chips away at a business owner's willingness to ever try again. That’s the true cost of poor experience — and the origin story behind OffDeal reshaping the sell-side market.

Unlike buyer-focused tools that create anti-network effects — where value decreases as more people join — our sell-side infrastructure gets stronger the more we scale. By restructuring the workflows of traditional investment banking with AI-native automation, we’re not just streamlining processes. We’re rebuilding the model entirely.

Here’s the data point that cracked it open for us: 80 percent of sub-$3M EBITDA businesses don’t use advisors. And when they don’t, they end up selling at 15 to 20 percent lower valuations. That’s not just market inefficiency — that’s lost generational wealth on a national scale.

Our vision is simple: build the Goldman Sachs for small business. A truly national, tech-enabled, sell-side-first firm that doesn't rely on outdated brokerage networks or cold email spam. We’re putting power back in the hands of founders, one automated workflow at a time.

Why Your Deal Sourcing Funnel Is a Lie — And How AI Fixes It

Private equity dealmakers are operating under a false premise: that there are tens of thousands of viable targets across fragmented industries. But once you cut through generic classifications and apply real-world filters like revenue bands, customer types, and service mix, most verticals shrink down to a few hundred relevant opportunities at best.

The problem lies in the data. Legacy platforms rely on outdated proxies like NAICS codes and clunky keyword indexing. They can’t resolve nuanced distinctions between, say, a janitorial business specializing in industrial contracts vs. a residential maid service — and they certainly can’t process a search like 'multi-site ketamine clinics with no talk therapy and at least five nurse practitioners.'

Our approach flips the script. We built an AI agent that navigates, reads, and reasons through millions of company websites like a trained analyst. It extracts structured insights on products, services, ownership characteristics, and end markets. Then, using embeddings-trained recommendation systems, we map your intent — not just your keywords — to surface companies that actually match the thesis.

This dual-layered system transforms precision sourcing from a slog into a solved problem. And the market impact is non-trivial. With accurate scope at the outset, investors can focus resources, calibrate platform theses, and stop wasting cycles on bloated target lists that never had signal to begin with.

Why Your TAM Is Wrong: The Real Deal Universe Behind Every Roll-Up

How many companies are actually available for your roll-up strategy? OffDeal’s AI shows it’s almost always fewer than you think. Most private equity buyers rely on outdated databases or raw keyword filters. What they’re missing is the nuance: product mix, end markets, ownership traits, and differential tagging that can’t be captured in spreadsheets.

Our system starts by deploying autonomous AI agents to navigate millions of company websites. They collect and verify context-rich data assets — not just a line of NAICS-code metadata, but structured intel on what the company actually does, who they serve, and how they operate.

Then the hard part. Built from embedding models — not generic search — our discovery engine interprets concepts, not just labels. That means if you type “power washing,” we understand you may also want companies that offer dry blasting, chemical treatments, or specialized industrial cleaning. Add in constraints like region, revenue, ownership status, and customer segment, and you stop guessing.

The result? Instead of seeing 10,000 maybe-companies, you get 500 precision-matched ones. This is more than better search. It’s redefining what “sourcing” means in private markets.

How OffDeal Replaced the M&A Back Office with AI Agents

Most firms in the lower middle market still rely on generic CRM tools and armies of admin assistants to manage the grind of sourcing, diligence, and close. What if all that could be eliminated?

The traditional M&A pipeline is bloated with busywork: qualifying inaccurate leads, chasing NDAs, scheduling with founders who still use landlines, updating spreadsheets, scrambling through Dropbox folders, and waiting on signature platforms built before streaming existed. It’s exhausting—and it’s the norm.

At OffDeal, we've rebuilt the entire workflow. AI agents now manage every phase of the pipeline, end to end. From pointing outbound toward high-propensity sellers to automating call scripts and CRM updates, to handling onboarding, NDA dispatch and signature, and even generating buyer lists and SIMs—we’ve reduced the human operational load to near zero.

There are no virtual assistants. No back-office drag. No missed reminders or manual calendar syncs. Everything communicates, learns, and acts in real-time, through independent agent systems. You won’t find third-party tools duct-taped together—we built our own.

This isn't a dream of the future. It's real, it's here, and it's reshaping how M&A gets done.

What's wild is one of our co-founders, ex-Meta, said our stack today is more sophisticated than enterprise-grade Salesforce deployments for name-brand investment banks. Why? Because everything we built is designed for one goal—finding real sellers and closing real deals, faster. And we don’t charge $25,000 to tell an unsellable business what it wants to hear.

The future of private deals doesn’t live in slide decks or cold calls. It lives in systems built to find truth, automate friction, and scale seller-first outcomes.
The moment it all flipped: we thought we were building a better data platform. But when we spoke to real business owners who’d been burned by failed deals, we saw the real problem. Not data. Not access. It was deal fatigue. Burnout from broken promises. That’s when we shifted everything—building OffDeal as a true sell-side engine built for outcomes, not just activity. Because owners deserve better deals, not just more bidders.
Most firms are still throwing bodies at M&A bottlenecks. OffDeal isn’t. Everything from targeting to outreach to NDA tracking flows through AI agents we custom built. That’s why we don’t need PAs, CRM admins, or even Zoom links. The entire M&A pipeline hums on its own. Sellers feel it. Buyers feel it. Legacy firms have no idea what's coming.
A business owner plugs in their website. Thirty seconds later, they’re staring at a list of real, matched buyers. No pitch decks. No cold emails. Just names. That’s what OffDeal built. The AI does the work in the background. All founders see is clarity.
Almost no one tells you this about M&A: most small business deals don’t even touch an advisor. That’s a massive crack in the market. When we realized sellers were ghosting buyers after failed deals, we ditched data for deal execution. The opportunity isn’t on the buy side. It’s helping founders get what they’ve earned—from the first call to closed. That’s how we’re using AI to build the Goldman Sachs for small business.
300 million data points? From who? Here's why legacy data vendors are bluffing—and why OffDeal built their own AI to replace them. They trained agents to surf millions of websites and extract real company info like an analyst would. No buzzwords. No recycled databases. Real answers. It's deal sourcing precision at a level the old platforms can't touch.
Most investors think there are thousands of roll-up targets. They’re wrong. OffDeal’s AI proves it’s often just a few hundred once you get specific. Real filters. Real data. Reality check.
Your CRM can't do this. OffDeal built AI agents that automate every admin task in the M&A flow — from calendar invites and owner follow-ups to teasers and data room setup. No virtual assistants. No third-party tools. Just clean, fast execution.
Put your company website in. Get matched with serious buyers in 30 seconds. No forms. No sales calls. Just names. The best part? Business owners don’t even care that it’s AI. They care that it works.
Most small business owners never make it to a real deal. Not because their companies aren’t good, but because the process is broken. This clip breaks down the messy truth: buyers dominate the outreach, sellers get fatigued, and most 'deals' die before closing. The insight? It’s not about finding buyers—it’s about overhauling the whole system with AI workflows. OffDeal is flipping the investment banking model to serve the sellers, not just chase deals.
Most people think there are thousands of companies to buy for every rollup idea. Here's the truth: once you filter by revenue, team size, product mix, customer type, and exclude what doesn’t fit, you’re often left with just a few hundred real targets. We use AI not to find 'more' but to find exactly who matters. It's not quantity, it's clarity.
Most investors think there are thousands of businesses ready for rollups. There aren’t. OffDeal’s AI tears that myth apart by showing how rare it actually is to find companies that check every box you care about: size, services, staffing, and more. It’s not about dumping keywords into a search bar. It’s about teaching a machine to think like a dealmaker. From analyzing millions of websites to translating vague categories like 'facility maintenance' into actual matches, this system filters noise into precision. The wildcard? Most deal teams are still guessing.
Most people think M&A deals fall apart because a company isn’t sellable. But what actually kills them? Admin bloat. Calendars. CRMs. Missed reminders. Manual follow-ups. OffDeal built AI agents that handle all of that—every little detail between first contact and close. No VAs. No Salesforce. Just end-to-end automation. This is what deal flow looks like when the humans focus on the human part, and the code handles everything else.
Most business owners don’t care about AI. They just want to know: who’s buying my company? So we built a tool. Drop your company’s website and 30 seconds later, you’re staring at a real list of buyers.

No sign-ups. No pitch decks. Just answers powered by an AI agent that learns your business and matches it to thousands of private equity deals. All hidden under the hood.

This is the only way software eats M&A—by earning trust with real outcomes, not tech jargon.