Human and AI Forecasts, Side by Side

Different Information, Same Question

Your staff, partners, and invited experts hold context that never reaches the public web: what a regulator signaled in a meeting, how a supplier is behaving this quarter, or what a colleague in the region is seeing on the ground.

Hinsley's AI Forecasting brings a different advantage. It researches the question, reads more source material than an analyst has time for, applies the same method every time, and never forgets to update its forecast.

Hinsley collects and shows you both human and AI forecasts, rationales, and sources. You can see where your network and Hinsley agree, where they diverge, and which evidence explains the difference. And when you need it, Hinsley also provides a combined, human+AI forecast.

Two Forecasts, Numerous Benefits

Running your forecasting question on Hinsley provides the benefits of both:

The human, AI, and combined forecasts are scored under the same rule, so once your questions resolve you can compare accuracy on the topics you care about. Forecast Accuracy Scoring explains how the scoring works, and relative Brier scores covers comparing forecasters against each other.

The Questions Come From Your Analysis, and the Answers Go Back Into It

Forecasting on its own is helpful, but its usefulness multiplies when it connects to the broader strategic analyses that your team is already working on.

When your team runs an analysis in Hinsley, you're already developing scenarios, running decompositions, and stress testing the analysis via red teams. When connected to a forecaster network, the network provides forecasts on questions derived from your decomposition's drivers and indicators. The scenarios you've developed get fed forecasting results to help inform their direction and likelihood. It becomes a fully connected lifecycle with probabilities and reasoning informing your analyses.

The Probability Arrives, and the Next Question Is "Why"

A probability prompts the next question: why? Human rationales surface private signals, local experience, and assumptions that other participants can challenge. Hinsley's AI rationale shows which public sources, base rates, and arguments shaped its estimate.

Read together, those rationales show where the network agrees, where it splits, and which evidence explains the difference. Individual forecasts remain available beneath the aggregates.

Hinsley organizes both sets into the same structure: a bottom line, historical precedent, arguments for the outcome, arguments against it, and what moved recently.

Who Does the Forecasting?

The roster is yours to compose, and the four sources mix freely on one site.

• Your own staff - people across functions, regions, and levels of seniority who hold context your analysts do not.
• Invited experts and partners - subject matter experts, partner organizations, and external stakeholders you invite by email.
• Professional forecasters - experienced forecasters from Cultivate Labs' own network, added to your site alongside your people.
• Hinsley's AI forecaster - available on every question, with its own forecast.

Who have we worked with?

Organizations including Shell, Ford, RAND, DARPA, and the UK Cabinet Office have run crowd forecasting programs with Cultivate Labs.

Cultivate Labs has built and run forecasting programs since 2006, including platforms used in IARPA-supported forecasting research. Crowdsourced Forecasting covers the method behind them.

If you'd like to learn more about how forecaster networks work in Hinsley, check out our documentation.