Our latest research

Poke

Poke is an early mover on iMessage and the only agent Apple approved for Messages for Business -- your 'Gen Z bestie' that texts first. But NPS is 0 and most users sit on the free tier. The deal: convert light users to daily habit before Siri AI ships free this fall and Claude Tag reaches iMessage.

Town

Town came out of beta in June 2026 and reached ~10,000 users. Users split: one founder calls it '100 times faster' than a $45K assistant, while some users plan to churn because Claude does the same thing. The bet: Town converts hype-stage trials into embedded retention, and can hold a preference-based edge as competitors close in.

Dust

French-born, now setting its sights on Silicon Valley. The edge: productized RAG, which Claude and other AI coworkers have yet to ship. The bet: converting adoption into operational dependency before Claude ships persistent memory, and expanding TAM from Europe to the US

Tasklet

Tasklet hit $7M ARR in 7 months on almost no marketing. The bull case: Tasklet successfully defines a defensible ICP and customer "hook" before the category commoditizes. The threat: every customer interviewed runs Claude in parallel, switching costs are near-zero, and some customers are trying to leave.

Viktor

Viktor hit $20M ARR in 4 months with minimal sales team. The bull case: 50%+ of customers interviewed already run operationally-embedded workflows, in a large TAM. The threat: Claude Tag now lives in Slack, and Claude can port Viktor workflows in minutes.

Baseten

Baseten — "best-in-class for model deployment," with customers like Cursor and Notion — reports 400% net dollar retention and zero churn among its top 30, riding open-weight inference's rise. But buyers multisource and switch easily, raising the question of whether that retention holds as inference commoditizes.

Modal

Modal hit an estimated $300M ARR in roughly eight months by collapsing messy, heterogeneous AI workloads into a single Python decorator — positioning it more as a hyperscaler competitor than a Baseten one. The open question is whether that pull holds as GPU supply normalizes.

Fireworks AI

Fireworks AI — "the most bang for GPU," built by the team that ran 50 trillion daily inference instances at Meta — rides open-weight inference's surge from 1% to 30% of tokens. But buyers see few differences across providers and switch easily, threatening the edge the whole bull case rests on.

Together AI

Together AI scaled from $44M to $1B ARR in under two years by owning its GPU fleet rather than renting — a vertical-integration bet that aims to capture margin and availability beyond what asset-light peers typically offer as open-weight adoption accelerates. The open question is how that owned compute fares if AI demand cools.

Campfire

Campfire customers report 2-3x cost savings versus NetSuite with implementations taking weeks instead of months, earning +60 NPS from finance teams who say it 'feels like getting in a race car.' The AI-native ERP has largely closed Rillet's early lead and is winning direct replacements from legacy systems. The question is whether Campfire can continue to develop broader mid-market traction as AI ERPs converge and legacy players add AI features.

DualEntry

DualEntry customers praise its multi-entity consolidation and see close times fall from 15-20 days to 8-10 days. But some evaluators call it 'barely better than QuickBooks,' and it trails Rillet and Campfire on the SaaS modules CFOs actually buy: RevRec, SaaS metrics, and billing. The question is whether a last mover can close a widening gap before the market settles on the two names everyone already cites.

Rillet

Rillet customers are cutting close cycles from 20 days to 5-7 days while saving 3-5x vs. NetSuite. But in head-to-head deals with Campfire, Rillet wins only 50% of the time as product parity narrows and both chase the same SaaS ICP. The company needs to expand beyond SaaS verticals or develop sustainable pricing power to grow beyond its SAM.

Day AI

Day AI has a strong founder team in CRM (built HubSpot) and $20M from Sequoia, but scored low on data quality in our survey. The trust gap forces customers to run legacy CRMs in parallel — a pattern that questions whether Day AI can become a system of record

Reevo

Reevo's conversational intelligence reportedly handles "80-90% of Gong," and one buyer called it "a net price reduction" versus their Apollo + Gong stack. Having raised $80M at a $500M valuation, the question is whether its unified Revenue OS can match best-in-class point solutions or stay limited to simpler greenfield deployments.

Attio

Attio earns a high NPS (+29), with 100% of switchers happy with the move. Its edge is superior UI/UX and embedded AI legacy players can't easily match. But the case is fragile: 53% would have stayed with comparable AI. Attio works best for teams under 20 sellers with simple GTM motions, pushing stronger companies to Salesforce as they scale.

Monaco

Monaco combines AI-native CRM with forward-deployed sales executives, charging $25-50K annually. Sam Blond's 'Sam-as-a-service' model creates strong early traction among well-funded startups, but scaling the high-touch approach without eroding differentiation presents the core challenge for this premium-positioned platform.

Fulcrum

Fulcrum customers report 3x ROI and 75% cost savings versus BPOs, but some customers flagged accuracy issues - a table-stakes criterion for brokerages. The company sits squarely in Tier 2 brokers while Tier 1 players build in-house and the long tail adopts Claude's insurance plug-in. To justify venture scale, Fulcrum must resolve accuracy gaps, win Tier 1 logos, and prove it can expand from its brokerage TAM into the much larger carrier/MGA opportunity.

Pace AI

Pace raised $10M from Sequoia to replace offshore BPO labor with AI agents across insurance workflows. The sweet-spot TAM is in carrier and MGA outsourcing. But Accenture restructured to embed AI into BPO contracts, and customers are scaling productivity tools that could shrink BPO volumes. Early enterprise traction is real, but the window to capture value may be narrowing.

FurtherAI

FurtherAI customers report 95% submission time reduction and 30% throughput gains, with one MGA clearing a 500-submission backlog. The company has strong traction at MGAs and raised a $25M Series A from a16z, but venture-scale returns require proving it can convert this into carrier adoption

Outmarket

Outmarket is beating Fulcrum in head-to-head evaluations despite Fulcrum's funding advantage and market awareness, validating strong product-market fit in Tier 2 segment. The question is whether this Tier 2 traction translates into the Tier 1 logos, or if competitive pressure from DIY solutions above and Claude's insurance plugins below boxes them into a narrowing middle market.

Wordsmith

Wordsmith combines contract negotiation with intake workflows, differentiating from pure-play assistants. Success hinges on deeper expansion into workflows and reaching teams beyond legal faster than commoditization arrives.

Crosby

Crosby raised $60M at $458M to automate contract reviews. Our analysis shows they need ~50-60% AI automation for unit economics to start to work -- below that threshold, this is a law firm with venture expectations.

DeepJudge

DeepJudge is a retrieval-first legal AI platform focused on internal knowledge search for law firms, a TAM estimated at $0.3B. The company was founded in 2020 by three ex-Google AI researchers with PhDs from ETH Zürich and raised a $41M Series A led by Felicis (backed by Coatue) in November 2025, bringing total funding to $52M at a valuation that reflects strong enterprise pull-through in European markets.

GC AI

GC AI is the rare legal AI company built for in-house teams, not law firms — and buyers love it (NPS +58, 63% eval win rate). The bet is whether that wedge grows into a platform before better-capitalized players close the gap.

Sandstone

Sequoia just backed a two-month-old company to unify the fragmented in-house legal stack. GCs love the vision; the question is whether Sandstone can deliver before Ironclad, Claude, and GC AI get there first.

Harvey vs. Legora

Near-identical products in a distribution race, not a technology battle. The category is real but smaller than the hype — can either justify their valuation before the wrapper thins out?

Vibe Coding

Explosive growth and massive TAM, but near-zero switching costs and Frontier Labs closing in. Can app builders build moats before the models they wrap make them redundant?

Numeric

AI-native close management with real enterprise logos, but wedged between BlackLine's installed base and ERP-native automation. Can Numeric pull financial logic out of the ledger before the ledger fights back?

David AI

Picks and shovels winner in the AI voice boom, or strong operator in a niche and commoditizing segment?

Cartesia

Best-in-class latency on a novel architecture, but trailing on quality and brand. Is their state-space model edge durable or temporary?

ElevenLabs

60%+ Fortune 500 penetration, but voice models are commoditizing fast. Is there enough headroom after an $11B valuation?

Poke

Poke is an early mover on iMessage and the only agent Apple approved for Messages for Business -- your 'Gen Z bestie' that texts first. But NPS is 0 and most users sit on the free tier. The deal: convert light users to daily habit before Siri AI ships free this fall and Claude Tag reaches iMessage.

Town

Town came out of beta in June 2026 and reached ~10,000 users. Users split: one founder calls it '100 times faster' than a $45K assistant, while some users plan to churn because Claude does the same thing. The bet: Town converts hype-stage trials into embedded retention, and can hold a preference-based edge as competitors close in.

Dust

French-born, now setting its sights on Silicon Valley. The edge: productized RAG, which Claude and other AI coworkers have yet to ship. The bet: converting adoption into operational dependency before Claude ships persistent memory, and expanding TAM from Europe to the US

Tasklet

Tasklet hit $7M ARR in 7 months on almost no marketing. The bull case: Tasklet successfully defines a defensible ICP and customer "hook" before the category commoditizes. The threat: every customer interviewed runs Claude in parallel, switching costs are near-zero, and some customers are trying to leave.

Viktor

Viktor hit $20M ARR in 4 months with minimal sales team. The bull case: 50%+ of customers interviewed already run operationally-embedded workflows, in a large TAM. The threat: Claude Tag now lives in Slack, and Claude can port Viktor workflows in minutes.

Baseten

Baseten — "best-in-class for model deployment," with customers like Cursor and Notion — reports 400% net dollar retention and zero churn among its top 30, riding open-weight inference's rise. But buyers multisource and switch easily, raising the question of whether that retention holds as inference commoditizes.

Modal

Modal hit an estimated $300M ARR in roughly eight months by collapsing messy, heterogeneous AI workloads into a single Python decorator — positioning it more as a hyperscaler competitor than a Baseten one. The open question is whether that pull holds as GPU supply normalizes.

Fireworks AI

Fireworks AI — "the most bang for GPU," built by the team that ran 50 trillion daily inference instances at Meta — rides open-weight inference's surge from 1% to 30% of tokens. But buyers see few differences across providers and switch easily, threatening the edge the whole bull case rests on.

Together AI

Together AI scaled from $44M to $1B ARR in under two years by owning its GPU fleet rather than renting — a vertical-integration bet that aims to capture margin and availability beyond what asset-light peers typically offer as open-weight adoption accelerates. The open question is how that owned compute fares if AI demand cools.

Campfire

Campfire customers report 2-3x cost savings versus NetSuite with implementations taking weeks instead of months, earning +60 NPS from finance teams who say it 'feels like getting in a race car.' The AI-native ERP has largely closed Rillet's early lead and is winning direct replacements from legacy systems. The question is whether Campfire can continue to develop broader mid-market traction as AI ERPs converge and legacy players add AI features.

DualEntry

DualEntry customers praise its multi-entity consolidation and see close times fall from 15-20 days to 8-10 days. But some evaluators call it 'barely better than QuickBooks,' and it trails Rillet and Campfire on the SaaS modules CFOs actually buy: RevRec, SaaS metrics, and billing. The question is whether a last mover can close a widening gap before the market settles on the two names everyone already cites.

Rillet

Rillet customers are cutting close cycles from 20 days to 5-7 days while saving 3-5x vs. NetSuite. But in head-to-head deals with Campfire, Rillet wins only 50% of the time as product parity narrows and both chase the same SaaS ICP. The company needs to expand beyond SaaS verticals or develop sustainable pricing power to grow beyond its SAM.

Day AI

Day AI has a strong founder team in CRM (built HubSpot) and $20M from Sequoia, but scored low on data quality in our survey. The trust gap forces customers to run legacy CRMs in parallel — a pattern that questions whether Day AI can become a system of record

Reevo

Reevo's conversational intelligence reportedly handles "80-90% of Gong," and one buyer called it "a net price reduction" versus their Apollo + Gong stack. Having raised $80M at a $500M valuation, the question is whether its unified Revenue OS can match best-in-class point solutions or stay limited to simpler greenfield deployments.

Attio

Attio earns a high NPS (+29), with 100% of switchers happy with the move. Its edge is superior UI/UX and embedded AI legacy players can't easily match. But the case is fragile: 53% would have stayed with comparable AI. Attio works best for teams under 20 sellers with simple GTM motions, pushing stronger companies to Salesforce as they scale.

Monaco

Monaco combines AI-native CRM with forward-deployed sales executives, charging $25-50K annually. Sam Blond's 'Sam-as-a-service' model creates strong early traction among well-funded startups, but scaling the high-touch approach without eroding differentiation presents the core challenge for this premium-positioned platform.

Fulcrum

Fulcrum customers report 3x ROI and 75% cost savings versus BPOs, but some customers flagged accuracy issues - a table-stakes criterion for brokerages. The company sits squarely in Tier 2 brokers while Tier 1 players build in-house and the long tail adopts Claude's insurance plug-in. To justify venture scale, Fulcrum must resolve accuracy gaps, win Tier 1 logos, and prove it can expand from its brokerage TAM into the much larger carrier/MGA opportunity.

Pace AI

Pace raised $10M from Sequoia to replace offshore BPO labor with AI agents across insurance workflows. The sweet-spot TAM is in carrier and MGA outsourcing. But Accenture restructured to embed AI into BPO contracts, and customers are scaling productivity tools that could shrink BPO volumes. Early enterprise traction is real, but the window to capture value may be narrowing.

FurtherAI

FurtherAI customers report 95% submission time reduction and 30% throughput gains, with one MGA clearing a 500-submission backlog. The company has strong traction at MGAs and raised a $25M Series A from a16z, but venture-scale returns require proving it can convert this into carrier adoption

Outmarket

Outmarket is beating Fulcrum in head-to-head evaluations despite Fulcrum's funding advantage and market awareness, validating strong product-market fit in Tier 2 segment. The question is whether this Tier 2 traction translates into the Tier 1 logos, or if competitive pressure from DIY solutions above and Claude's insurance plugins below boxes them into a narrowing middle market.

Wordsmith

Wordsmith combines contract negotiation with intake workflows, differentiating from pure-play assistants. Success hinges on deeper expansion into workflows and reaching teams beyond legal faster than commoditization arrives.

Crosby

Crosby raised $60M at $458M to automate contract reviews. Our analysis shows they need ~50-60% AI automation for unit economics to start to work -- below that threshold, this is a law firm with venture expectations.

DeepJudge

DeepJudge is a retrieval-first legal AI platform focused on internal knowledge search for law firms, a TAM estimated at $0.3B. The company was founded in 2020 by three ex-Google AI researchers with PhDs from ETH Zürich and raised a $41M Series A led by Felicis (backed by Coatue) in November 2025, bringing total funding to $52M at a valuation that reflects strong enterprise pull-through in European markets.

GC AI

GC AI is the rare legal AI company built for in-house teams, not law firms — and buyers love it (NPS +58, 63% eval win rate). The bet is whether that wedge grows into a platform before better-capitalized players close the gap.

Sandstone

Sequoia just backed a two-month-old company to unify the fragmented in-house legal stack. GCs love the vision; the question is whether Sandstone can deliver before Ironclad, Claude, and GC AI get there first.

Harvey vs. Legora

Near-identical products in a distribution race, not a technology battle. The category is real but smaller than the hype — can either justify their valuation before the wrapper thins out?

Vibe Coding

Explosive growth and massive TAM, but near-zero switching costs and Frontier Labs closing in. Can app builders build moats before the models they wrap make them redundant?

Numeric

AI-native close management with real enterprise logos, but wedged between BlackLine's installed base and ERP-native automation. Can Numeric pull financial logic out of the ledger before the ledger fights back?

David AI

Picks and shovels winner in the AI voice boom, or strong operator in a niche and commoditizing segment?

Cartesia

Best-in-class latency on a novel architecture, but trailing on quality and brand. Is their state-space model edge durable or temporary?

ElevenLabs

60%+ Fortune 500 penetration, but voice models are commoditizing fast. Is there enough headroom after an $11B valuation?