Track:
Sourcing
▾
Key Takeaways
Private market coverage is genuinely ahead of where most corp dev teams think it is: 21M+ companies, intent-to-sell signals, live deal tracking from a banker network
Grata claims 96% accuracy on intent-to-sell signals, predicting sale activity within 6-12 months based on company research behavior
Differentiation is in proprietary data layers (seller intent, live deal network, conference intelligence) that Claude can't replicate from public sources
Acquired by Datasite in late 2024: $500M+ being deployed to consolidate the sourcing data market (acquired SourceScrub, Value8, MergerLink, Silk)
MCP connector to Claude demonstrated live: run searches directly from Claude, pull results back into Grata for full data set
Pricing: $6K–$12K/seat (tiered: Growth, Scale, Alpha). Alpha tier limits seller intent to ~10% of user base to maintain signal value
Pros & Cons
Strongest private market coverage for North America and Europe: not just AI-scraped, but human-annotated by a team of 600 researchers
Intent-to-sell signals are genuinely differentiated: no other sourcing tool showed this capability
Live deal network via MergerLink: bankers sharing real processes you can get into
Instant setup; CRM integration in 30 minutes
Not built for LatAm, Africa, East Asia: core geos are North America, Europe, Australia/NZ
Not suited for early-stage VC targets, real estate parcels, pharma patents: operating companies only
Alpha tier (seller intent) limited to 10% of users by design: not accessible at lower tiers
Datasite acquisition creates questions about product roadmap independence
"We have data on intent to sell. We know when companies are in market, when they're talking to advisors, signing up for data rooms. We calculate this at 96% accuracy: when we see a spike, very high chance that company is going to try to sell in the next 6-12 months." - Nevin Raj
Sourcing
▾
Key Takeaways
AI-first from day one: no manual data annotators, all verification done with models. Company raised Series A in 2024, 30-person dev team
Global coverage strength: works wherever companies have websites. Solid Middle East, LATAM, South Africa coverage: more global than most competitors
Claude MCP integration shown live: search HVAC companies in Texas with owner nearing retirement age. AI delivers results with ownership/contact data automatically
Most common failure mode is entity retool: two companies with same name in different countries getting linked incorrectly, not outright data errors
Pipeline management is lightweight by design: not a replacement CRM; connects to HubSpot, Salesforce, Affinity, DealCloud
Pricing: $5K–tens of thousands/year for individuals/teams; Big Four range into six figures. Free trial available
Pros & Cons
Best global coverage in the room: works in Middle East, LATAM, South Africa, anywhere with a web presence
Claude and ChatGPT integrations both work; built specifically for M&A workflows and deliverable templates (one-pager, timeline slides) built into the plugin
Honest about limits: explicitly says "use CapIQ for public company multiples, use Apollo for bulk outreach"
Instant setup, free trial offered without hesitation
No manual human data review layer: pure AI verification means some errors won't be caught by the human eye
Public company data is a known weak spot vs. FactSet, CapIQ
Pipeline management lite: teams needing full CRM functionality should budget for that separately
"Getting from 0 to 95% accuracy is pretty easy. The most common data mistake isn't that a data point is incorrect: it's a missed linkage between companies. The harder part is entity retool." - Niilo Pirttijarvi
Sourcing / CRM
▾
Key Takeaways
Relationship intelligence CRM built for private markets: not a sales tool skinned for M&A, built from scratch for deal teams
Auto-captures email and calendar interactions without manual data entry: relationship strength scored on reciprocity patterns, not raw interaction count
Slack/Teams integration shown live: ask "@4Degrees tell me about our relationship with [company]" like querying a teammate
Network Scout surfaces new companies entering your team's network via people connections: prior employers, portfolio companies of new contacts
Clear on what it doesn't do: no company discovery (Grata/Inven territory); focuses on relationships you already have or are building
Pricing: $4K–low six figures/year depending on team size. Migration from existing CRM typically 1-2 months, low-to-mid five figures one-time
Pros & Cons
Relationship strength scoring is sophisticated: goes beyond interaction count to reciprocity, meeting types, content of communications
Slack/Teams integration genuinely changes how teams interact with deal data: conversational rather than dashboard-bound
Browser extension gives relationship "x-ray vision" on any company website or LinkedIn profile
SIM/pitch deck upload → auto-populate CRM fields via AI extraction
Does not surface net-new targets: only manages relationships you already have. Must pair with a sourcing tool
Off-the-cuff phone calls (not on calendar) can't be tracked. Apple limitation
Migration from Salesforce/established CRM is real work: 1-2 months, services cost
"The piece that is proprietary to your organization is the internal context: the deals you're evaluating, the emails you're exchanging, the relationships your team has built. That's where 4Degrees and Claude become genuinely powerful together." - Ablorde Ashigbi
Financial Diligence
▾
Key Takeaways
Pronounced "Quicker": deterministic Excel-native analytics for customer data cubes, P&L, and financial statement analysis
Core value prop: financial analysis has a right answer. LLMs are probabilistic and degrade over time; QuikIRR is deterministic and fully auditable to source data
Live demo showed customer data cube analysis: raw file in → structured cohort/retention/segmentation analysis → Excel workbook in ~3 minutes
Every output is formula-based in Excel: traceable to the underlying dataset, same as if a skilled analyst built it by hand
Data cleanup module detects duplicates, negative values, anomalies before analysis runs: designed to catch dirty data before it corrupts results
Pricing: Several tens of thousands/year per firm. Enterprise (large funds) can reach six figures. Per-seat + base platform fee model
Pros & Cons
Deterministic outputs: no probabilistic drift, every formula auditable back to source data
Used alongside Claude/ChatGPT, not instead of them: handles the structured numerical work those tools aren't reliable for
Generalizes across thousands of messy real-world data sets: not trained on a handful of clean examples
Customizable output structure to match firm-specific presentation standards
Not freeform: won't answer "what are the tax implications of this transaction structure." Built for specific financial analysis workflows
Excel-only output: teams that want web-based dashboards will find this limiting
"Given a set of data and calculation rules, there is a right answer, and you have to get to that right answer every time. There is zero room for error. That's philosophically tough for an LLM to deliver repeatably." - Dan Callahan
Full Lifecycle
▾
Key Takeaways
M&A Operating System: pipeline → diligence → integration on one platform. Claims to cut total person-hours per deal in half
Two AI approaches: embedded AI within the SaaS app, and MCP connector for Claude/ChatGPT to trigger DealRoom workflows directly from those tools
LLM-agnostic by design: benchmarks Anthropic, OpenAI, Google on speed/cost/accuracy continuously, routes to best model per use case
Email integration shown live: AI scans incoming deal emails and attachments, surfaces updates for corp dev to approve/import into DealRoom pipeline with one click
Explicitly not for sell-side investment bankers (separate product: Firm Room) or granular integration project management (pairs with Smartsheet/Monday)
Pricing: Per deal room ("room") not per user. Works for 2+ deals/year: explicitly not recommended for teams doing 0-1 deals/year
Pros & Cons
Only tool in the room covering the full pipeline-to-integration lifecycle on one platform
Per-deal pricing aligns cost to value: not penalized for bringing more people onto a strategic deal
SOC 2 certified; data never used for model training: important for sensitive deal docs
Diligence-to-integration handoff is built into the platform: preserves context across phases
Not built for sell-side or for teams doing fewer than 2 deals/year
Integration module is high-level program management only: tactical execution needs a separate tool
"The ability to actually take action on documents: not just analyze them: and move the deal forward from within Claude or ChatGPT: that's where the real value comes from. It's not either/or. It's 1+1=3." - Greg Lord
Legal / Diligence
▾
Key Takeaways
End-to-end diligence management: questionnaire tracking, data room organization, and bulk contract review: all three in one platform
Key insight: data input quality determines AI output quality. Tower's value is ensuring the AI is operating on a complete, organized, reliable dataset before any analysis runs
Auto-organizes messy data room uploads against a diligence questionnaire in minutes: renaming, filing, flagging incomplete document sets (missing amendments, unsigned contracts)
Operating at 95th percentile accuracy on legal benchmarks. Every AI answer has source language citation for human verification
Valsoft Corp case study: reduced external legal spend by hundreds of thousands in one year by insourcing confirmatory diligence reporting
Pricing: Flexible: user-based or usage-based depending on team needs. Free one-month trial (full deal). Operational within an hour of setup
Pros & Cons
Strongest audit trail: every AI output has hoverable source citation, one-click to source document. Verification is fast
Free trial for a full deal: real evaluation, not a sandbox
Works for both buy-side and sell-side; law firms, serial acquirers, brokerages
Auto-flags missing signatures, missing amendments, expired documents on upload: immediate catch before review begins
Not suited for low-document-volume deals: teams with ~50 documents don't need the infrastructure
Y Combinator stage company: less enterprise track record than some alternatives
Bulk review is a strength, but the biggest value is in the organization layer: which may require internal change management to adopt
"An AI efficiency gain is only real if it exceeds the verification cost. We make verification fast: every AI answer is cited to source language you can check with one click." - Noah Walters
Practitioner Session
▾
Key Takeaways
Real corp dev practitioner (170+ location dermatology group, 5-8 deals/year) showing his live Claude setup: built on nights and weekends over 1-2 months
Core message: AI is for everyone. You don't need to be technical to get started; you can get significant value from tools you already have
Built custom Claude skills for M&A diligence: P&L analysis, EMR review, quality of earnings prep: built specifically for for healthcare physician services deals
Running on HIPAA-compliant enterprise Claude setup: not a personal account. Security was a foundational requirement before any workflow was built
Demonstrated skill orchestration: multiple skills working together across a deal workflow, not just isolated prompts
The M&A Science MCP is the practitioner intelligence connector powering DealPilot's AI search capabilities
Built by Wyecliff AI, the team behind M&A Science's AI infrastructure layer
Key Lessons for Practitioners
Start with the workflows where you're already doing the work manually: that's where the leverage is highest
Skills are the differentiator: the quality of your outputs is a function of how well you've defined the task, not just the model
HIPAA-compliant enterprise setups exist and aren't difficult to get approved: don't let security concerns be a reason not to start
Small deals (solo practices, $1-5M revenue) are where AI creates the most leverage: the work that still has to happen regardless of deal size
Setup takes real time: this was a 1-2 month nights/weekends project, not a plug-and-play deployment
Skill quality depends heavily on domain expertise: generic prompts produce generic outputs
"AI is for everyone. What I wanted to show is that it's so easy to get started: and you can get pretty far with the tools you probably already have on your desktop." - John Palusci
Sourcing / CRM
▾
Key Takeaways
Relationship intelligence CRM for private capital: the value prop is knowing "who do we know at [company]" without hunting through Slack
Open platform strategy: meets customers where they are. If you want to update the CRM from Claude, do it. If you want to use ChatGPT for meeting prep, Affinity connects to it
SIM/pitch deck upload → auto-populate CRM fields is a newer feature generating strong customer response
Relationship strength ranking on warm intros is the primary use case: predicts value of a relationship, not just who you know
Roadmap is moving toward agents: the CRM evolving from system of record to a system that takes action autonomously on your behalf
Pros & Cons
3,000+ investment teams as customers: proven at scale across private capital
Vertical SaaS advantage: meeting prep, relationship scoring, pipeline updates all optimized for private capital workflows specifically
Open platform doesn't lock you into Affinity's AI: you keep using the tools you prefer
Agent roadmap is directionally right for where deal teams are heading
Overlaps with 4Degrees in several core use cases: procurement teams should run both against their specific workflow needs
Not a sourcing tool: still need Grata/Inven for company discovery
"Generic LLMs don't know your firm's 10 years of relationship history. We don't want to compete with ChatGPT or Claude: we want to meet you where you're at and make sure your proprietary data is always available to whatever tool you're using." - Rebecca Campbell
Financial Diligence
▾
Key Takeaways
Deterministic quantitative analysis platform for PE diligence: co-founders from Goldman and Vista Equity; technical co-founder from Tesla and AI startups
Core argument: investors could only diligence 10-15% of deals that came across their desk. Keye enables another deal per year: a meaningful North Star for PE funds
Strong position on LLM limits: "anything below 100% accuracy on financial data is 100% inaccurate: you don't know which 10 numbers are wrong in a 300-number model"
Builds industry-standard and custom data packs: cohort analysis, retention, revenue segmentation, price-volume mix: from customer cubes and pipeline files
Odin AI agent (conversational interface) layered on top for questions and insights: but deterministic engine runs underneath
Pros & Cons
Clearest articulation of the determinism argument of any vendor: the 98% accuracy framing resonated strongly in the room
PE-specific: built specifically for for the actual quantitative workflows PE associates run, not a general tool adapted for finance
Can say no to a deal earlier with more confidence: the value of speed and clarity on bad deals is as high as the value on good ones
PE-first focus means corp dev teams may find it less directly applicable to strategic acquirer diligence workflows
Doesn't address qualitative diligence (management quality, customer interviews, commercial risk)
Overlaps significantly with QuikIRR: procurement teams should compare these two directly
"Getting from 0 to 95% accuracy is easy. Getting to 95-100% is impossible because of the probabilistic nature. And you don't know where the 1% inaccuracy lies. That non-deterministic output: that reliability lapse: is the problem." - Rohan Parikh
Legal / Diligence
▾
Key Takeaways
Noah Waisberg co-founded Kira Systems in 2011 (bootstrapped to 100+ people, $50M raised, sold to Litera in 2021). Zuva is the spinout: he kept the core AI and 30 of 180 people
Two primary use cases: (1) disclosure schedule prep and pre-diligence at big law firm quality for 10-20% of the cost; (2) bidder question management for sell-side processes
Sell-side focus is a genuine differentiator: no other vendor in the room addressed the seller's pain of running a process
Uses AI plus experienced big law lawyers to supervise output: hybrid AI-native services model, not pure software
Noah is author of "AI for Lawyers" (WSJ bestseller) and "Robbie the Robot Learns to Read" (2015): possibly the most experienced AI-in-law practitioner in the room
Pros & Cons
Deepest legal AI pedigree in the room by far: 15 years building AI for M&A legal work
Sell-side use case is underserved: disclosure schedule prep and bidder question management solve real pain nobody else addressed
Human-in-the-loop supervision model builds in quality assurance that pure AI tools don't have
API-first approach means it can power other products and workflows
Primarily a services product wrapped around AI: less of a self-serve platform than other vendors
Buy-side contract review is increasingly commoditized by general LLMs. Zuva's moat is strongest on the sell-side and in the data quality layer
Pricing not discussed in detail during the session
"18 of the world's top 25 M&A lead table law firms were Kira customers. People who used our software said they could review contracts in about half the time with the same or better accuracy." - Noah Waisberg
Legal / Compliance
▾
Key Takeaways
450-person company serving ~1,500 private funds, asset managers, and investment banks: more established than most vendors in the room
Product suite across negotiation (NDA/contract automation), compliance (fund obligation tracking), and DDQ (due diligence questionnaire management)
AI-native services model: AI plus human-in-the-loop, not pure software. Relevant for high-volume, high-stakes workflows where 100% accuracy matters
DDQ product is newest: helps asset managers and PE funds respond to the growing volume of investor/LP due diligence questionnaire requests
Atlas entity management product is built specifically for for high-volume M&A transaction teams: not a horizontal tool adapted for PE
Pros & Cons
Most mature company in the room: 12-year track record, 1,500 fund clients, 450 employees
Human-in-the-loop model provides quality assurance that pure AI tools can't match for mission-critical legal workflows
Breadth of product suite means less vendor fragmentation for PE back-office operations
DDQ product addresses a genuinely painful and growing problem for fund managers
Less focused on M&A transaction diligence specifically: strength is in back-office PE operations (fund docs, KYC, compliance)
Services-heavy model means some workflows aren't fully self-serve
Not the right fit for corp dev teams focused on transaction diligence rather than fund operations
"We combine artificial intelligence, workflow software, and in some cases human in the loop to basically solve the whole problem out of the box: so that firms don't have to deal with it and can focus on other things." - Troy Pospisil
Full Lifecycle
▾
Key Takeaways
Vertical AI platform for M&A integration built on 20+ years of practitioner experience at Google, Microsoft, eBay: not a consultant's framework wrapped in software
Core product is M&AOP: a living playbook platform that encodes the team's M&A framework and adapts to each serial acquirer's specific deal patterns, risks, and capabilities
Force multiplication is the primary value prop: enables lean teams and non-dedicated M&A participants to apply senior practitioner judgment consistently across concurrent deals
Strong on the "it depends" problem: framework handles the 80% of repeatable patterns; the platform surfaces where the 20% of genuine snowflake decisions need judgment
Explicitly not another checklist tool: the framework is opinionated and dynamic, not a static template library
Pros & Cons
Only integration-focused vendor in the room: addresses the stage where most M&A value is won or lost
Built by practitioners who ran integration at Google-scale: operational credibility is real
Vertical AI framing is right: built specifically for beats general-purpose for integration workflows
Addresses the earlier-than-you-think problem: integration planning that starts during diligence, not after close
No pricing discussed during the session
Demo time was limited: harder to evaluate the product depth versus the framework thinking
"Each deal's a snowflake: but that's also not true. It's dangerous to say we have a magic pattern, but it's equally dangerous to treat every deal as entirely unique. The framework handles the repeating 80%; you focus your judgment on the other 20%." - Gwen Pope
Legal / Diligence
▾
Key Takeaways
Former M&A attorney (Connor) + MIT PhD in AI/ML (CTO) + Big Law M&A partner as Director of Product: the legal domain knowledge in the founding team is unusually deep
Works both buy-side and sell-side: sell-side prepares companies for diligence (document organization, disclosure schedules, issue flagging before disclosure); buy-side receives full day-one diligence reports
Document relationship graphing is the standout feature: the system maps how documents reference each other, links related files side-by-side, and flags missing referenced documents automatically
Automated disclosure schedule generation is the newest tool: upload the transaction agreement and the data room, wait up to 12 hours, receive complete, citable disclosure schedules ready for attorney review
Ensemble approach to accuracy: multiple AI models cross-check results; system flags uncertainty and errs toward over-inclusion rather than missing something important
Pricing: Flexible: seat-based for law firms; project-based for corporates; bulk document packages available (few dollars per document). One onboarding call, plug-and-play setup
Pros & Cons
Best automated disclosure schedule capability shown in the room: nobody else demoed this end-to-end
Document relationship graph is genuinely differentiated: catches missing exhibits, cross-referenced documents, and broken doc chains that humans routinely miss
Designed for both sides of the transaction: sell-side prep is an underserved use case most vendors ignore
All outputs citable to source document; easy to verify and edit; exports to Excel and Word
Best suited for smaller and mid-market transactions: max ~5,000 documents. Very large data rooms (tens of thousands of docs) are outside the current sweet spot
Disclosure schedule generation takes up to 12 hours: not a real-time tool for that use case
"Getting 80% of the way there isn't as difficult as that last 20%. Ask your competitors: can they really close that gap? That's where we focus: end-to-end work, not just a co-pilot giving you a starting point." - Connor Acle
Lightning Round
▾
Wokelo
AI research and diligence platform for deal teams: shifting to an API/data model as customers want to control their own UX and workflow orchestration
wokelo.com ↗
Watch Recording
DealSage
AI diligence tool focused on accelerating research and synthesis for deal teams
dealsage.io ↗
Watch Recording
Sc0red
Risk scoring platform for M&A: quantifies deal risk across multiple dimensions to help prioritize diligence focus and make faster go/no-go decisions
sc0red.com ↗
Watch Recording
PinpointAI
AI platform for M&A workflows focused on deal team efficiency and workflow automation
pinpointai.com ↗
Watch Recording
Lightning round format: 15 minutes per vendor, straight to demo. No slides. The best moment of the round was watching tools run live on real queries without prep time.