From First AI Experiments to 100% Adoption

iOpt is a Glasgow-based PropTech and IoT company serving UK social housing providers. Its sensors and  algorithms monitor damp, mould, occupancy and energy performance across tens of thousands of homes. Vaul  Labs has worked with iOpt since 2024, and in 2026 turned to putting AI to work across the business. 

The challenge 

iOpt’s leadership wanted AI working across every team in a way people could trust on real work. 

• Employees were already trying AI tools on their own initiative, with no shared view of what was in use or  what for. 

• There was no agreed platform and no written rules on what data could go into an AI tool - an important question for a business whose working material includes tenants’ names and addresses. 

• What worked for one person stayed with that person, so prompts and context were rebuilt individually rather than shared. 

iOpt’s leadership chose to invest in AI early and at pace, and engaged Vaul Labs to bring the whole business along together: one platform, one set of rules, and training that made sense to each team. 

Our approach 

Vaul Labs led the programme as iOpt’s fractional Chief Data Officer through spring 2026, with iOpt’s leadership backing it from the start. We brought every department along at the same time through an AI Working Group with one AI Champion from each team, meeting weekly during the rollout. 

We compared the four leading platforms against iOpt’s requirements and recommended Claude. It was rolled out across the organisation, and engineering adopted Claude Code. 

We wrote the AI policy with the Champions. An early draft required routine personal data to be anonymised, and the Champions pointed out that this would stop people using AI for real housing work and push them towards free tools. The final policy instead classifies data as allowed, allowed with care or prohibited. Every department then had a 30-minute session built around its own work, so each person could see what Claude  would do for them, followed by weekly drop-ins and a shared Teams channel. 

The investment went in up front. Everyone learned at the same time, with training that related to their own role and leadership visibly behind it, so that using Claude became part of everyday work within weeks. 

Scope & constraints 

• Claude was connected to iOpt’s Microsoft 365 tools, including Teams, SharePoint and Outlook, through each user’s own permissions. 

• Special-category personal data under UK GDPR is allowed only with care, and any client-specific restriction  overrides the general policy. 

• People stay human-in-the-loop, and AI-assisted work is reviewed before it is shared with clients.

Outcomes 

“We decided to invest in AI properly from the start and bring everyone with us at once. Vaul Labs gave us a platform decision we could stand behind, a policy our teams could follow, and training built around each team’s own work. Months later, everyone here still uses it.” 
Dane Ralston, CEO, iOpt 

From a spring rollout to everyone at iOpt using Claude 

• Five months after rollout, every employee at iOpt uses Claude, and it remains the business’s AI tool of choice. 

• That level of use has held since the rollout ended, which shows AI has become part of how iOpt’s teams work day to day. 

From hours of manual analysis to work delivered in minutes 

• Within weeks of rollout, iOpt’s tracker recorded more than ten named wins across five departments, with active users reporting around five hours a week saved on average. 

• A solar PV and heat pump analysis that took two to three hours now takes around ten minutes, and it  underpins iOpt’s battery-storage business case across 17,000 homes. 

• Account management turned 39 condition-survey drawings into a full monitoring strategy and two client-ready reports in under 24 hours, and operations built a weekly scan across 5,500 properties that saves around a day a week. 

From individual setups to shared standards 

• iOpt works on a single approved AI platform, with one policy that tells everyone at iOpt what data can go where. 

• The design team built an iOpt branding skill that anyone can use to produce on-brand documents and presentations. It saves people the time they used to spend finding templates and making presentations look polished. 

iOpt’s experience reinforces something we see often: when an organisation invests in adoption up front, with leadership behind it and training that relates to each person’s own work, AI use takes hold quickly and lasts.  Shared standards, clear ownership, and practical judgement are what allow capability to build over time.