Paperplane helps builders', plumbing, and timber merchants mitigate their credit risk by predicting which customers will default 4+ weeks before anyone else. Instead of relying solely on credit reports and lagging public data, Paperplane uses the company's own transaction data and customer knowledge directly from sales reps to make more accurate predictions in real-time.
Computational physics + ML PhD dropout.
I left Cambridge with a post-graduate degree in Advanced Computer Science and peer-reviewed research in generative language models receiving almost 100 citations in the field. I started a company applying LMs to API security, where I met Daniel Kwak, who led my Seed Round. I left this company to join Daniel and Ben in building NLP for Sales. In my spare time I enjoy rowing long distances. Once I rowed unsupported from Africa to America in 40 days.
Previous ML Ops tech lead at Xometry and software engineer at the Naval Nuclear Laboratory. Attained his PhD in Nuclear Engineering at MIT specializing in reactor coolant modeling and simulation.
TLDR: ✈️ Paperplane automatically updates Salesforce after sales calls. Sales teams at fast-growing B2B companies do 14%+ more calls per week because they no longer have to spend hours updating their CRM.
There are over 3 million B2B salespeople in the US, and each of them spends 3-6 hours each week manually updating their CRM with data about their customers and live deals.
This creates a problem at every layer of large and fast-moving sales teams.
Reps at companies like Nextdoor, Twilio, Checkr, and Front spend 10-15% of their time on data entry versus actually selling; Sales leaders spend hours chasing reps for updates; and sales-ops / rev-ops teams deal with their CRM missing ~50% of opportunity data. This has always made it difficult for sales teams to know what is really happening and which deals are at risk.
In an environment where many B2B companies have to do more with less, something has to give.
Paperplane solves this problem by automatically updating Salesforce after every sales call.
We plug into a sales team’s conversational intelligence platform (e.g. Gong, Chorus, Wingman) and Salesforce.
After the call, Paperplane transcribes and extracts the required Salesforce fields (eg. next steps, follow-ups, pain points, MEDDIC) and notes. No keywords, tracking words, or scripts to adhere to. Paperplane’s NLP model is built to retrieve the correct data regardless of phrasing and is grounded to mitigate hallucinations that are common with language models today.
Once done, sales leaders get the updates and notes they need in minutes and reps no longer spend hours moving data from their notes into Salesforce.
We’re currently working with a handful of fast-growing B2B companies to increase their pipeline visibility, improve sales productivity, and help them understand why they’re winning/losing.