Simulate any decision.
Rendering South African granularity.
Kento simulates South African populations to predict market response, policy adoption, and political behaviour before the world answers in public.
We're building simulation software that recreates South Africa using a multi-agent approach.
Money can't reach the unreachable
Humans are unreliable narrators of their own behaviour. Memory fails, incentives distort answers, and social pressure warps what people say away from what they actually do. Kento grounds simulation in measurable outcomes, contextual datasets, and calibrated personas rather than stated preference alone.
Interrogate every assumption
The most valuable South African populations are often the hardest to understand with traditional tools. Township shoppers, LSM bands, policymakers, informal traders, multilingual households, youth voters, and rural communities shape markets and institutions differently. We represent those distinctions directly.
See around every corner
You can't survey a future that hasn't happened. New policy, coalition politics, load-shedding scenarios, market entry, pricing shifts, and public communication strategies create conditions that do not yet exist in historical data. Simulation makes those futures testable.
Built for the operator
Research too often stops at description instead of action. Kento is built for operators in business, government, and politics who need decision-ready outputs across segmentation, forecasting, campaign strategy, policy design, and market execution.
If prepaid mobile data prices fell by 20% in Gauteng townships, how would switching behaviour change over the next 90 days?
What is the minimum amount of information a voter needs about a ward candidate in a coalition election to feel confident voting for them?
How do different load-shedding communication strategies perform across language groups, municipalities, and income bands?
If a prominent political figure endorses a local candidate, does that increase turnout or drive backlash?
Do consumers feel pressure from friends, family, or colleagues to try a new financial or health product when data costs are still high?
How likely is each segment to engage with different product messages or offers before launch?
If a municipal by-election were held today, who would win and where does turnout move most sharply?
What is the likely adoption rate of a national service or ID programme among older citizens in rural provinces?
Academic foundation
In an independent study, Stanford researchers showed generative agents replicate real human attitudes at 85% accuracy — comparable to humans re-taking their own surveys two weeks later. Kento applies that architecture to South Africa, where no global platform is truly calibrated.