natural language processing Startups

Alex Kilka built an automated marketplace arbitrage system that identified price discrepancies between eBay and Amazon, scaling it from a side project into a three-year business selling 5,000+ power tools and generating $4,000-$10,000 in monthly gross profit. He shut down the business after Amazon changed its return policy, making inventory management too complex, and has since pivoted to Austin Data Solutions, a five-person consulting firm using data analysis, NLP, and computer vision to solve business problems.

Pegasystems, founded by Alan Treffler in 1983, is an enterprise SaaS platform for digital process automation and customer engagement. The company took 24 years to reach $100M in revenue, IPO'd in 1996 at a $32M run rate (raising over $100M), but struggled until Treffler resumed leadership in 2005. Through disciplined product development, a 2010 acquisition of predictive analytics capabilities ($160M), and geographic/market expansion, the company has grown at 20%+ annually and now operates at an $847M ARR run rate with a ~$5B market cap.

Keytext, founded in 2010 by Narjesu Buffadin, started as a professional services firm focused on natural language processing before pivoting to a SaaS product in 2015. The platform helps businesses understand customer and employee disengagement by analyzing unstructured text feedback from multiple sources. With 30 customers, 20 employees in Montreal, and $4M raised, they're approaching $1M ARR with 5X year-over-year growth, driven primarily through partnerships with survey platforms and marketing agencies.

MedStack is a SaaS platform for digital health companies that provides infrastructure, security, and privacy compliance for handling patient health data in the cloud. Founded by Balaji Gopalan, the company grew from $300k to $672k ARR across ~70 customers by pivoting to a self-service product (MedStack Control) and shifting to an inbound-focused sales strategy. With $2.3M raised and seeking a $500k extension, MedStack aims to reach $1M ARR while maintaining lean operations with a 7-person team.

Stitched Insights is a SaaS platform using deep learning and NLP to analyze customer feedback from support tickets, emails, and surveys to generate ranked SWOT analysis and market insights. Founded by Demetri Pavlov in 2018, the company grew from ~$60K in 2018 revenue to ~$100K in 2019 while bootstrapping with founder capital and strategic partnerships through GSV Labs, with a goal of reaching $1.5M ARR by end of 2020.

Rippleworks is a people analytics SaaS platform launched in 2018 that helps organizations understand employee sentiment and optimize workforce motivation and retention. Starting with athletic teams, they've scaled to 50 customers including police forces, production studios, and high-performing teams, generating over $100k MRR with an average customer paying ~$2,000/month for $4-8 per user licenses. The team of 35 (including 15 engineers) is expanding their sales organization and planning to launch NLP-powered features to drive toward their 10x revenue scaling goal.

Joe Coll is a 25-year-old founder who built Pulse, an AI-powered SaaS platform that predicts emotional responses to ad creative with 97% accuracy before publication. He bootstrapped the business with over £1.48 million of his own capital from his successful marketing agency (Oncore), which generates over £1M in revenue annually with 25 employees. With 50 users in beta testing and pricing ranging from £500-£5,500/month, Pulse is preparing to launch paid offerings and raise a Series A round in early 2022.

Moth and Flame is an award-winning immersive VR learning platform founded in 2015 that helps enterprises drive behavior change through virtual reality training experiences. After launching with custom innovation projects, they pivoted to a SaaS model in 2019 when Accenture approached them, and have grown to $5M in total revenue with $2M recurring (40% of revenue). The team of 39 is targeting a Series A of $10-12M to build out synthetic media and NLP capabilities, positioning themselves as the Squarespace for VR.

Abstract is a real-time conversational intelligence platform that uses Google's NLP and machine learning to coach sales reps during calls—telling them what to say and how to handle objections in the moment, rather than analyzing failures after the deal is lost. Founded by Greg Raffner in early 2020, Abstract launched in April 2021 and reached $178k MRR ($2.1M ARR) within 5 months through a laser-focused SEO strategy targeting 'conversational intelligence software' and related keywords, even outranking $10B competitor Gong. Operating with just 2 US employees plus contractors in India and Indonesia, Greg has raised only $630k total while achieving exceptional capital efficiency and preserving equity.

Leedsift is a B2B sales intelligence platform that mines social media data to identify buying intent signals and deliver qualified leads to sales teams. Founded by technical co-founders in 2012, the company spent three years pivoting through consumer-facing and agency-focused models before finding product-market fit in the B2B SaaS lead generation space. Today, with 105 customers and growing 13% month-over-month, they are approaching $1M ARR.

Rob Schall is a serial SaaS entrepreneur who sold his first real estate website company for $80M and his second vacation rental platform for $15M. He's now building CN, an AI-powered sales productivity platform that measures lead quality, seller attributes, and macro factors to help sales teams scale more effectively. The company has raised just over $2M and is targeting the trillion-dollar opportunity of AI applied to CRM systems.

X.ai is an AI-powered personal assistant that schedules meetings via email on behalf of users. Founded by serial entrepreneur Dennis Mortensen after he sold his previous analytics company and discovered he had 1,019 meetings in a year (65% requiring rescheduling), X.ai has raised $44 million and uses a novel approach of testing market viability before building the full product—starting with a concierge MVP and only then raising seed funding to validate the technical approach through data labeling.