HBM Flowers Kenya Ltd, a Dutch–Kenyan B2B summer-flower grower with 25 hectares of its own land, is introducing a dedicated AI agent into an operational foundation built over time.
For years, the business has collected selected operational data and developed its own software layer for planning, monitoring, process control and KPI steering. That work created records, workflows, dashboards, integrations and management routines around a grow-to-order, year-round flower programme.
The next step is HBMaster Kenya: a dedicated agent that helps prepare analysis, reporting, planning context and communication from approved information. “Human-led” means a person reviews its output and approves external action or customer commitments. The farm team and management set direction and remain responsible.
A concrete example
An existing Figor API connection provides shipment temperature, humidity and location information. HBMaster Kenya can help organise available shipment-condition context for human review when a quality question arises.
Core distinction: HBM did not begin with AI. It began by building records and tools for its people. The Agent is the layer that helps bring that work together.
Four intended areas of value
Quality: clearer review of available crop, test, shipment and follow-up context.
Continuity: stronger connection of requirements, actions and planning context.
Control: more visible source gaps, exceptions and management follow-up.
Cost price: more consistent investigation of relevant input, labour, packing, transport and output drivers.
H U M A N – L E D B Y D E S I G N
What the Agent does — and does not do
HBMaster Kenya supports selected information work: organising approved sources, connecting verified signals, preparing explanations and supporting human review. External publication and outreach remain approval-led.
The Agent does not autonomously run the farm or make customer commitments. Results depend on source quality and human follow-up; outcomes are not guaranteed.
A longer digital journey
The company describes its software and data direction at hbmkenya.com/future/ and its human-led Agent approach at hbmkenya.com/ai/. Together, the pages show the sequence: build records, develop workflows, translate information into management steering, then add an AI layer that can connect approved context.
The HBM-specific software layer includes internally developed workflows, reports, dashboards, integrations, controls and agent configuration. The wider stack also uses third-party and open-source infrastructure and external AI models; the complete underlying technology stack is not presented as fully proprietary.
