Many AI instruments in the present day can carry out particular person duties with excessive accuracy, usually surpassing 98%. Nonetheless, in advanced multi-step workflows widespread in monetary providers, authorized, and insurance coverage sectors, small errors accumulate, inflicting total accuracy to drop to 30–40% by the ultimate step. This “compound error” challenge prevents firms from adopting AI at scale as a consequence of unreliable outcomes.
Curvestone AI, a London-based startup, addresses this important drawback by offering a workflow automation platform that maintains constant accuracy all through complete document-heavy and compliance workflows.
The corporate simply secured $4 million in seed funding led by MTech Capital, with participation from Enhance Capital Companions, D2 Fund, and Portfolio Ventures, to speed up product growth and increase go-to-market efforts, rising its library of confirmed workflows.
Earlier than elevating exterior funding, Curvestone AI reached profitability and grew income 7x in 12 months. The platform now processes billions of tokens quarterly throughout authorized and monetary providers workflows.
Fixing AI’s reliability drawback
Curvestone was based by brothers Dawid and Sebastian Kotur, who beforehand led AI automation applications at PwC, Metro Financial institution, and GKN, and have been concerned in Fb Developer Circles London. Recognising the potential of generative AI in skilled providers, they determined to pivot from consulting to constructing an answer that basically improves AI reliability in regulated industries.
Curvestone AI’s expertise integrates with all main language fashions and present methods equivalent to CRMs, doc administration software program, and mortgage origination platforms. Its no-code workflow configuration empowers groups to adapt processes simply as laws evolve.
“In regulated industries, high quality and scale have at all times been at odds. You may assessment every thing and go broke, or reduce corners and hope for the perfect. AI that truly works adjustments that equation by dealing with routine validation at scale whereas people concentrate on the advanced circumstances that want skilled judgement,” stated Dawid Kotur, Co-Founder and CEO of Curvestone AI.
Not like many AI options that see accuracy fall off steeply in multi-stage processes, Curvestone’s platform ensures steady, excessive accuracy at each workflow stage. This makes Curvestone the primary actually reliable AI automation answer tailor-made to the nuanced necessities of monetary providers, authorized, and insurance coverage compliance workflows.
Direct opponents embody AI startups centered on workflow automation; nonetheless, Curvestone differentiates itself by fixing the “compound error drawback” with precise proof of constant efficiency in regulated sectors.
What’s subsequent?
Curvestone AI plans to speed up product growth and increase its library of confirmed workflows.
“Curvestone is fixing the exhausting technical drawback of automating advanced workflows whereas attaining excessive accuracy, and accuracy is paramount in regulated industries like monetary providers. Their answer works throughout monetary providers, authorized, and insurance coverage, significantly for compliance workflows that desperately want automation. The early traction they’re seeing validates actual market demand,” stated Kevin McLoughlin, Companion at MTech Capital, who’s becoming a member of Curvestone AI’s board.
