Tech • AI • Robotics • Game

VIDEO
ENFR
TodayPlay for freeFury BloxLeaderboardShortsTop StoriesFor youTopicsVideosYT channelsArchivesSearchFavorites

Claude Project debuts as Europe banks on ScalIT and AI agencies

AI CodingSaturday, October 10, 2026· 3 videos

Briefing

Audio player
0:00 / 0:00

Claude Project targets long-running coding

Anthropic's Claude Project expands Claude Desktop from a chat interface into a persistent coding workspace. Projects can keep shared goals, repository access, instructions and memory across sessions, reducing the need to restate context. The feature is positioned around long-running software work rather than one-off prompts. Early hands-on results suggest the concept is promising but still uneven in execution.

Claude Desktop adds parallel threads

Claude Project is designed to spin up multiple work threads against the same codebase and objective. In principle, it can assign separate branches, keep sub-tasks isolated and continue progress in the background. That pushes the product closer to an autonomous development assistant than a standard chatbot. The pitch is orchestration: one persistent project coordinating many coding actions.

Cloud sessions meet repo workflows

The new workflow combines local files, GitHub repositories and cloud execution inside Claude Desktop. Users create a project, define a broad objective and attach a folder or repo for ongoing work. Projects can then be pinned and reopened as durable workspaces instead of disposable chats. The approach aims to centralize code, memory and execution in one environment.

Claude Project shows early friction

Despite the broader ambition, early testing indicates Claude Project remains slow and inconsistent in practice. The system can struggle to deliver the fluid, autonomous behavior implied by its product framing. That leaves a gap between the vision of background software development and current reliability. For now, the feature looks more like an important first step than a finished replacement for existing coding workflows.

ScalIT bets on bank core rebuilds

ScalIT is positioning itself for a structural overhaul of European banking infrastructure. The company sells software to banks, fintechs and corporates moving into embedded finance, focusing on regulated, mission-critical operations. Its thesis is that decades-old core systems will give way to more modern, cloud-style architectures. If that transition accelerates, vendors with deep implementation capacity stand to win unusually sticky contracts.

ScalIT carries private equity risk

The business sits in a demanding middle ground between classic SaaS and infrastructure-heavy financial software. An investment fund committed roughly €50 million, backing both existing product-market fit and operational change. Leadership shifted about three years ago as a founder prepared to retire, with the incoming chief executive having first joined through the investment. The setup offers large contract potential, but long sales cycles and high execution risk remain central.

Real estate becomes automation beachhead

A separate trend in 2026 automation services is pushing agencies to specialize by industry and problem rather than sell generic AI promises. Real estate stands out because delayed responses to buyer inquiries can directly destroy commission revenue. That makes lead handling, qualification and staff workload easier to quantify in ROI terms. In this model, one closed transaction can justify a full year of service fees.

Niche agencies ditch broad AI pitches

The recommended operating model is narrow: solve one expensive workflow inside one vertical, then reuse templates across similar clients. Markets cited include real estate, e-commerce, healthcare, dental, legal and financial services, with real estate favored for its common stack of Gmail, calendars, CRMs and spreadsheets. The argument is that buyers respond to business pain, not abstract automation language. Specialist positioning also reduces custom rebuilds and improves credibility in outbound sales.

Videos covered

Previous briefings · AI Coding