Background & Opportunity
Appkeep is an Italian integrated facilities and property management provider covering facilities maintenance, construction and project management, property management, energy efficiency, renewable energy, and ESG-related services. It delivers projects through its own teams and a nationwide technical network. The company website presents 8 operating locations in Italy, more than 180 local technicians, and more than 350 customers. These figures describe overall operating scale and service coverage; they cannot be directly attributed to the AI project. For a business spanning management, administration, operations, field technical services, and customer delivery, process quality affects handoffs, document handling, inspection results, and compliance performance. Appkeep’s opportunity was therefore not simply to add another software tool, but to reconsider how everyday work was divided, transferred, checked, and confirmed. Existing workflows contained manual steps and repetitive tasks, while quality control depended largely on employees’ time and attention. Public materials provide no implementation baseline for time, error rates, rework, costs, or ROI. The size of the original problem therefore cannot be inferred from undisclosed figures, and later changes cannot be precisely attributed to AI. Directly placing AI on top of an unexamined workflow could have made the problem worse. A tool might process unclear information faster while amplifying duplication, omissions, and unclear responsibility boundaries. Appkeep consequently began with process understanding rather than tool purchasing. The project analyzed real workflows department by department, identified wasted time, rework, errors, and weak quality controls, and assessed which issues should be solved through process redesign and which checkpoints were suitable for AI. This sequence shifted the question from “Where can AI be used?” to “Where is there a verifiable business need?” At selected points, AI was used for process control, data-consistency checks, anomaly alerts, result validation, and employee support. Employees remained involved in judgment and retained final decision authority. AI was therefore framed not as a workforce-reduction tool, but as a second layer of control over human work. That arrangement suits facilities and property management, where standardization must coexist with field judgment. Repetitive checks can receive technical support, while complex situations still require people to understand context, handle exceptions, and own decisions. When the case was published, the project was still ongoing. Management said that visible results had appeared after several months, including fewer manual repetitive tasks, improved management, administrative, and operational efficiency, shorter work cycles, improved business performance, and fewer non-compliance issues. These outcomes come from management statements and supplier case materials, not from an independent audit. Public sources also disclose no specific time savings, reduction percentages, implementation costs, or ROI. The case’s main value is therefore its adoption path and control logic, rather than a set of transferable financial figures. Appkeep’s experience suggests that a scaling service company should first create a shared understanding of its processes, then place technology where it can be checked, reviewed, and continuously adjusted. Only when responsibilities, inputs, outputs, and quality standards are clearer can AI become a stable operating control layer rather than another source of complexity.
Original Problem / Model
The original model contained manual and repetitive steps, with error prevention and quality control depending heavily on employee time and attention. Automating an unexamined workflow could have amplified existing confusion. Public materials disclosed no quantified baseline.
Turning Point
The turning point was to analyze actual processes department by department, identify needs, problems, and improvement opportunities, and then decide which activities required redesign and which were suitable for AI.
Core Decision
Appkeep chose to continue its digitalization and AI adoption program, positioning AI as a second layer of control over human work rather than as a direct replacement for employees.
Execution / Model / Acquisition Changes
AI was used for process control, data-consistency checks, anomaly alerts, and result review. Employees continued to interpret situations and retain final decision authority, combining technical support with human accountability in service delivery.
Measurable Results
After several months, management reported shorter work cycles, improved business performance, and fewer non-compliance issues. Public materials disclosed no specific time savings, reduction rates, costs, or ROI, and the website’s scale metrics cannot be directly attributed to AI.
Lessons
Understand and restructure the process before deploying AI to control critical points. Use verifiable checks to strengthen human work, rather than simply adding a tool on top of existing confusion.