Published On: Thu, Apr 30th, 2020

Index and Credo lead a $2.75M seed in anti-fraud tech, Resistant AI

Prague formed Resistant AI has nabbed a $2.75M seed round. The confidence startup’s appurtenance training record is designed to be deployed on tip of AI systems used for financial preference creation to strengthen business in markets such as financial services and ecommerce from attacks such as targeted manipulation, adversarial appurtenance training and modernized fraud.

The seed turn was co-led by Index Ventures (Jan Hammer) and Credo Ventures (Ondrej Bartos and Vladislav Jez). Seedcamp also participating, along with Daniel Dines, CEO of UiPath; Michal Pechoucek, CTO of Avast and other unnamed angel investors. Bartos joins a house of directors on interest of a investors.

The startup sells an additional covering of insurance that’s privately designed for tightening confidence around programmed functions such as credit risk scoring and anti-money laundering by regulating tech to detect feign papers that feed such systems. Its tech is also directed during uncovering questionable patterns of exchange that competence prove a vital conflict on a indication itself or an try to duplicate supportive data.

“Historically, all systems that make high-value financial decisions turn targeted. This is already function with a programmed systems deployed by a fintech and financial business and we are here to strengthen them,” pronounced Martin Rehak, owner and CEO, in a statement.

The seed turn is Resistant AI’s initial tranche of outmost funding, with a founders bootstrapping a association given starting adult in Feb 2019.

“We have onboarded a initial business in 2019 and a appropriation will assistance us scale a sales organization to accommodate a rising direct from banks and fintechs,” Rehak told us. “We are safeguarding a AIML systems used in financial automation from strategy or injustice by intelligent attackers.”

Resistant AI has dual products it offers a business during this stage: First, request inspection. It offers a appurtenance training complement that’s designed to dwindle and reject “malicious documents” submitted for programmed processing. “Bank statements, payslips, invoices, squeeze orders and KYC papers submitted to fintechs and banks are frequently manipulated or totally falsified,” explained Rehak. “Resistant Documents, a initial service, identifies and rejects a questionable or antagonistic inputs.”

A second charity — Resistant Transactions — relates AI to mark cryptic transaction patterns.

“We work with a fact that many attacks on AI systems need endless communication to learn a vulnerability,” he said. “Our complement is singular by inspecting all a patron queries (which can take form or payments, income transfers or credit applications assessed by a complement we protect) in context of identical queries. By looking during a tide of queries statistically, we can recognize and retard a attacks that find to take a information embodied in a indication (information stealing) or, worse, aim to poke a complement into creation a wrong preference by exploiting an existent disposition in a system.”

Resistant AI isn’t violation out patron numbers nonetheless though Rehak pronounced it onboarded a initial business final year. “The appropriation will assistance us scale a sales organization to accommodate a rising direct from banks and fintechs,” he added, observant also that it will be spending on building out product facilities and fluctuating functionality, as good as on beefing adult a sales and go-to-market team.

“Right now, a aim business are financial and fintech startups, as good as other companies deploying a programmed routine (both program and RPA) in their financial processes,” he added. “The financial systems are a stream focus, though a attacks on appurtenance training are applicable in many other areas: routine automation, e-commerce, strategy of ‘trend detection’ algorithms in amicable media and other opportunities.”

It’s regulating a SaaS indication — preferring a value proceed to pricing, per Rehak. “Our problem and proceed is new, and we feel that a value pricing indication aligns a incentives between us and a patron in a optimal way,” he pronounced on that.

Asked who he sees as a categorical competitors for a business, he cited Google Brain and a tech giant’s activities in adversarial appurtenance learning.

The infancy of work in this area is now finished in-house by a vast tech companies building their possess exclusive systems — such as Google and Microsoft, he added.

Other competitors he mentioned were Inpher, that is enabling appurtenance training on encrypted data; Sentilink, that is doing showing of fake identities in a US; and Bullwall (Denmark) and YC-backed Inscribe (US/Ireland) that are focused on request forgery.

Resistant AI’s founders have a credentials in appurtenance training practical to cyber confidence problems carrying founded Cognitive Security, an progressing startup that they subsequently sole to Cisco in 2013. Over some 12 years operative in a confidence attention Rehak pronounced they saw how enemy targeting AI systems were removing increasingly worldly in avoiding showing — that gave them a thought for their latest business.

Commenting on a seed appropriation in a statement, Jan Hammer, ubiquitous partner during Index Ventures, added: “Automation, potency and trustworthiness are cornerstones of financial innovation. As appurtenance training takes some-more and some-more nuanced financial decisions, it needs to be protected. And this is not loyal usually in finance, though a attacks will fast widespread to other domains as well. More of a activity currently takes place online, a trend accelerated by COVID-19, and one we trust will last. With criminals prepared to take advantage of each vulnerability, a need for solutions such as those from Resistant AI has never been greater.”

About the Author