Thrive AI: From Claude-Powered Upskilling to Production AI, at Scale.
Client Context
Thrive AI is a dedicated smart building technology business and sister company to Reds10, the industrialised construction specialist redefining how the UK designs, manufactures and operates buildings. Thrive AI’s mandate is distinct and demanding: apply AI to the continuous stream of data generated by building sensors – temperature, water, energy and occupancy among them – and turn it into insight that keeps buildings healthier, more comfortable and more efficient for the people who use them. It is a mission at the leading edge of construction’s shift from static assets to intelligent, responsive buildings, and one that is increasingly being delivered on Claude.
Even with a software team already fluent in AI, adoption across the wider business was inconsistent before Digital Futures’ Athena platform arrived: AI was useful for routine tasks such as email drafting, but its potential to transform day-to-day operations remained largely untapped.
Problem
Manual, repetitive work was consuming disproportionate time – hours spent extracting, analysing and reporting on operational data – while the wider business lacked a structured way to scale the AI fluency its engineering team already possessed. As a business built around AI, Thrive AI had the technical foundations in place; the expertise and the appetite were both there. The opportunity was to extend that fluency beyond the engineering team and embed it consistently across the whole organisation – turning individual capability into a shared, company-wide operating culture, and building the foundation from which production-grade Claude use cases could be delivered.
Approach
Thrive AI’s team was onboarded onto the Athena Apply persona – Digital Futures’ AI upskilling platform, built on Claude, and tailored to employees applying AI to enhance day-to-day productivity rather than build or govern it at a technical level. Each learner began with the Skills Lab assessment to establish an individual AI capability baseline, generating a personalised Skills Fingerprint. Athena’s Recommendation Engine, powered by Claude, then built each learner a tailored path through the content, completed at their own pace alongside the day job.
Throughout, Digital Futures’ delivery team worked alongside Thrive AI, using managerial dashboards to give a live, cohort-wide view of engagement and skills across every learner – so support could be directed where it mattered most. The rollout culminated in a hands-on hackathon, translating individual learning into applied, team-based problem solving with Claude, the same model underpinning Athena, as the working tool. Several of the concepts that emerged have since moved beyond the hackathon: piloted, refined, and deployed into live production – powered by Claude.
Working with Digital Futures gave us something training alone never does, momentum. Athena took the AI fluency that lived inside our engineering team and spread it across the whole business, and within weeks that fluency was showing up in how we serve customers. Support tickets are answered faster and more consistently, and ideas that started as learning exercises are now live products. That is real commercial value, not a course completion certificate.
Impact
Materials Database & Pricing Automation
A Claude-powered agent now identifies and prices components from a centralised materials database, replacing what was previously a manual lookup process. The concept extends to auto-generating O&M manuals and producing a Bill of Materials directly from CAD drawings – removing hours of repetitive lookup and pricing work and redirecting that time to higher-value engineering work.
This has cut administrative overhead by a day a month, plus three days per live project. Across the financial year that adds up to a full working week of admin removed every month, winning back time for the team to focus on innovative, work-winning tasks. For a small, growing team, that matters enormously.
Intelligent Support Triage
Deployed into production, this is a use case Thrive AI relies on every day – not once or twice a project – and that daily cadence is exactly what makes it valuable. Claude now screens each incoming support ticket for completeness, drafts a reply chasing any missing information, and – once a ticket is complete – retrieves the answer via an internal MCP server to prepare a response. What was once a manual, interruption-heavy task handled ticket by ticket is now a largely hands-off workflow: it frees the team from constant context-switching and gives customers a faster, more consistent reply every time – a daily improvement to customer experience and brand reputation that compounds well beyond the hours saved.
The workflow now delivers approximately £20,000 in operational savings, with a market-positioning benefit alongside it. Where the average ticket previously took 99 minutes to investigate, Claude now generates a response in around eight minutes, followed by 30 minutes of human validation. Comparable AI triage tools are priced from £0.35 to £1.50 per ticket, rising to £20,000 to £37,000 and above for enterprise deployments, yet Thrive AI’s approach delivers savings that are already proven. That is a stronger position than most competitors can claim, and it supports charging a premium rather than competing on cost alone.
Automated naming Sync
Built and deployed on Claude, this tool automatically renames objects and devices across the Logic Machine (LM) and ETS platforms whenever a client changes bedroom naming within a building – removing a time-consuming manual update that previously had to be done by hand, project by project.
Previously this process took a minimum of one week depending on the size of the project, and had to be carried out separately on each LM and within the ETS software. The new approach completes the same task in approximately 30 minutes. Accuracy has improved significantly too, removing the 20% error margin that repetitive manual work inevitably invites and freeing valuable time for higher-priority, revenue-generating work.
The measure of any upskilling programme is what runs in production afterwards. With Digital Futures and Athena, the answer is three live deployments: automated materials pricing, intelligent support triage, and synchronised building data. Hours of manual, repetitive work have been engineered out of our operations, and the team’s time now goes to the engineering problems that actually move us forward.
Summary
From a single training rollout, Thrive AI has built a company-wide AI operating culture and a working pipeline for turning ideas into production. Adoption has been near-universal: the team reports a 90% uptake in AI use, with Claude now the everyday tool of choice, and Thrive AI’s lead developer describes Claude as delivering the output of an additional software engineer, effectively doubling his own productivity. Athena, built on Claude, gave every employee – not just the engineering team – the fluency to spot where AI could remove friction from their day. The hackathon that followed put that fluency to work, and three of its concepts have already moved from pilot to live deployment on Claude: automating materials pricing, triaging customer support, and keeping building data in sync in production.
For a business tasked with using AI to make buildings healthier and more comfortable for the people inside them, embedding Claude this deeply – into how the team learns and into what the business runs – moves Thrive AI, and its sister company Reds10, further ahead in transforming how the construction industry designs, builds and operates.