Cost & ROI assumptions

Make the assumptions visible before making a decision.

Assumption — change it

UsersDeployment packAI usage per working dayModel tierLocal model capacityDisplay currency

Illustrative annual time and value

Assumption — change it

Average monthly salary (THB)Hours per week on documents, decks and finding informationShare of those hours AI could help with (%)

Editable assumptions

Source / assumption: call costs measured on 2026-09-23 traffic. VM: 4 vCPU / 16 GB in Bangkok or Singapore (AWS/GCP). GPU and VM ranges are planning assumptions. Working days: 22; exchange rate: 33. Hours per day: 8 and weeks per year: 52 are editable modelling assumptions.

Efficient model: cost per call (USD)Assumption — change itFrontier model: cost per call (USD)Assumption — change itWorking days per monthAssumption — change itVM monthly minimum (USD)Assumption — change itVM monthly maximum (USD)Assumption — change it24 GB GPU monthly minimum (USD)Assumption — change it24 GB GPU monthly maximum (USD)Assumption — change it80–96 GB GPU monthly minimum (USD)Assumption — change it80–96 GB GPU monthly maximum (USD)Assumption — change itTHB per USDAssumption — change itWorking hours per dayAssumption — change itWorking weeks per yearAssumption — change it

Estimated monthly components

SaaS includes AI. This model shows estimated components, not a bill or a quote. BYOTOK AI is billed directly by your provider; local-model cost is represented by the GPU row.

Estimated monthly components
Component	Range
Kairos	Per quote
AI usage	—
Cloud / server	—
Local GPU	—
Operating subtotal · excludes licence and services	—
Services	Per quote

Illustrative annual time and value

Hours freed per year—

Value of those hours (THB)—

Annual operating estimate (THB)—

This is a comparison with operating costs, not full payback. Licence and services need a quote. Time freed is not guaranteed savings.

Get a quote for this