# Decisions Lab

> Decisions Lab (Almita Limited) provides simulated buyer research for B2B messaging decisions. Product Marketing, Sales, Growth, and leadership teams simulate how specific buyer groups respond to the positioning options they are considering before Sales and Marketing commit to one direction. Built on 230M+ B2B profiles available to identify and build relevant buyer groups across markets, industries, company sizes, and roles. A six-month Hong Kong government study measured a 41.9% weekday cut in restaurant food waste; the simulation returned 40%. Across buyer personas, accuracy varied 7.6x less than the average of five frontier models. Supported by HKSTP, AppWorks, Founders Inc., and AWS.

## Positioning

Decisions Lab helps teams make the four messaging decisions that determine whether the right buyers recognize, understand, and believe the value of their product. Every engagement starts with a conversation.

## Core Pages

- [Home](https://www.decisionslab.io/): Simulated buyer research for B2B messaging decisions. After proof points (Hong Kong government pilot 41.9% live vs 40% simulated, 7.6x consistency vs frontier models, 230M+ B2B profiles), the page covers the four messaging decisions, how it works (Define, Simulate, Compare, Activate), who it is for (scaling beyond early adopters, product launch, market expansion, repositioning), and FAQ with FAQPage JSON-LD.

- [Articles](https://www.decisionslab.io/articles): Research, case studies, and thought leadership on buyer simulation, messaging decisions, and AI-assisted market research.

- [Outreach](https://www.decisionslab.io/outreach): Product page for AI-powered cold outreach and prospect-level message simulation. Distinct from the enterprise research homepage; targets sales teams optimizing outbound email.

## Key Facts

- **Category:** Simulated buyer research for B2B messaging decisions
- **Profiles:** 230M+ B2B profiles for building relevant buyer groups
- **Consistency:** Across buyer personas, accuracy varied 7.6x less than the average of five frontier models
- **Pilot validation:** Hong Kong government restaurant study measured 41.9% weekday food-waste reduction; simulation returned 40%
- **Four messaging decisions:** Choose which buyer group deserves the next push; choose which message or value proposition should lead; find the objection weakening the message; identify the proof that makes the claim credible
- **Methodology:** Identify comparable prospects using professional profiles, company context, company news, public professional activity, social context, and customer evidence, then simulate how each buyer group responds to the same messages, research questions, and scenarios
- **Outputs:** Segment and messaging recommendation, buyer response comparison, buyer language bank, messaging playbook and guardrails, Sales and campaign activation brief, and next market test
- **Partners:** AWS, Founders Inc., AppWorks, HKSTP

## Case Studies and Articles

- [Analytics Firm 20× Reply Rate Case Study](https://www.decisionslab.io/articles/how-one-analytics-firm-achieved-a-20-reply-rate-lift): B2B analytics consultancy improved cold outreach reply rates using prospect simulation.
- [Tamjai Case Study](https://www.decisionslab.io/articles/tamjai-case): Customer behavior prediction and outreach optimization.
- [DTIA 2025](https://www.decisionslab.io/articles/dtia-2025): Louis Cheung (Co-founder and CEO) on the human simulation engine.
- [HKU Guest Lecture](https://www.decisionslab.io/articles/hku-guest-lecture): Academic talk on AI and human behavior prediction.
- [Responsible AI](https://www.decisionslab.io/articles/responsible-ai): Panel on ethical AI development and data privacy.

## Trust and Legal

- [Trust](https://www.decisionslab.io/trust): Security overview, subprocessors, and compliance documentation.
- [Privacy Policy](https://www.decisionslab.io/privacy-policy): Data collection, usage, and user rights.
- [Terms of Service](https://www.decisionslab.io/terms-of-service): Service terms and legal obligations.

## Contact

- Email: hello@decisionslab.io
- Demo: https://cal.com/louisdlab/30min