Agents are making more and more product and marketing decisions. Our Claude skill gives yours a way to check those decisions with simulated users as it works.
Joseph Wright
We can now build user models of people who speak languages other than English. They're built from real human data in that language and validated to the same state-of-the-art accuracy as English models.
Joseph Wright
Projects puts an agent in charge of simulating your users. Describe a decision or upload a spec and the agent will plan multiple simulations to predict the consequences of making the change. Apply the findings to v2 and re-simulate to find an even better solution.
Joseph Wright
Get a custom model of your users, with state-of-the-art accuracy in predicting how individuals answer new questions.
Joseph Wright
Anacreon predicts how individual people answer new questions, not just population averages, preserving diverse and minority group opinions. It achieves a state-of-the-art ordinal alignment of 0.775, the individual-level accuracy measure on which the field has converged.
Pranav Dahiya
MONEYME launches credit products and changes its messaging more often than research with real people can keep up with. Its teams now have a model of 450 everyday Australians, each one its own fine-tuned LLM, that they can consult in minutes.
Fabian Teichmueller
We gave Claude Fable one prompt and access to simulated users, then walked away. It shipped a working debt planner, and every decision behind it traces back to a simulation it ran.
Joseph Wright
The work that decides what to build, who for, and how to reach them did not get cheap when the code did.
Fabian Teichmueller
Sometimes a Semilattice prediction lands with one answer at 99.8%, or two answers stacked at 99.9% combined. We decided to run an audit to diagnose what was happening.
Joseph Wright
We built the simulation first. We're now starting to think that simulation is what makes a context engine useful, and context is what makes simulation useful. They need each other.
Fabian Teichmueller
We redesigned audiences in Semilattice to make accuracy fully inspectable.
Joseph Wright
207 questions from the FCA Financial Lives Survey, tested out-of-sample. Here's where the model is strong, where it struggles, and why the pattern matters.
Joseph Wright
Today we're launching 16 UK Consumer Finance audience simulation models in Semilattice.
Fabian Teichmueller
Early-stage product decisions are made in a noise-rich, data-poor environment, often informed more by assumption than data. Semilattice helps product managers cut through the noise of internal alignment, letting them share their audience’s voice with teams and execs.
Fabian Teichmueller
Test batches are a new feature in the Semilattice API which let you test population model accuracy against specific use cases
Joseph Wright
Predict user behaviour like you make database queries.
Joseph Wright
Semilattice believes significant advances in decision-making, management, and governance are possible, but only if we can accurately model the complex human systems which run the world.
Joseph Wright
Despite its importance for marketing, product development, and sales, most organisations lack knowledge about their users for decision-making. Semilattice changes this.
Fabian Teichmueller
Semilattice lets you rapidly answer emerging user & market questions as your audience evolves through awareness, consideration, purchase, and usage stages.
Joseph Wright
Semilattice gives you an AI model of your audience which answers questions like they do in seconds, reducing decision uncertainty without delaying action.
Joseph Wright
