Blokk Logo
Login

Consulting
Powered by AI

We advise leadership teams on where AI creates measurable value — then stay to make it happen. Strategy, architecture and hands-on implementation from a team that has seen what actually ships.

The Blokk Dashboard

Rehearse decisions
before you make them.

Our dashboard runs multi-agent simulations of your market. Describe your business and a scenario — a price hike, a launch, a competitive crisis — and watch dozens of AI agents with distinct personas react, argue, shift opinions and commit to decisions across multiple rounds.

Multi-Agent Simulation

40+ unique AI personas — satisfied customers, at-risk accounts, prospects and competitors — react to your scenario round by round. You see the full conversation graph, who influenced whom, and the decision each agent ended on.

Pricing, launch and crisis modules
40 to 60 diverse agent personas per run
Phased rounds: awareness → decision → aftermath
Live streaming: agents appear as they're born

Strategic Report

After the simulation ends, you get a report written against real agent behaviour: revenue impact, churn estimate, segment breakdown, competitor reactions, a 3-period timeline, ranked recommendations and a confidence score.

Revenue impact & churn estimate
Segment-by-segment risk breakdown
Competitor reaction forecast
Ranked recommendations with confidence

Explore & Interrogate

The simulation isn't a static PDF. Chat with any agent in character, survey the whole population with one question, or inject a what-if variable and re-run the decision logic to see who changes their mind.

1:1 chat with any agent, in persona
Whole-population surveys with sentiment
What-if re-simulation with deltas
Network graph of every interaction
How a simulation runs

From your scenario
to a strategic read.

Every simulation follows the same four-stage pipeline. You see progress streaming in real time — agents, conversations and the final report appear as the engine produces them.

01

Describe

Tell the dashboard about your company — industry, products, segments, competitors, pricing — and propose a scenario. A price hike, a new product, a competitor's aggressive move. Anything you would actually need to decide.

02

Populate

The engine generates 40–60 agents with distinct personas: role, budget, priorities, price sensitivity, satisfaction, alternatives. Distribution is tuned for your market — satisfied, at-risk, prospects and competitors.

03

Simulate

Agents talk to each other across multiple rounds under shifting phases — awareness, discussion, evaluation, coalition, decision, aftermath. Every message, effect and sentiment is recorded and rendered as a live network graph.

04

Analyze

The engine writes the final report grounded in who actually said what, who changed their mind, and who committed to churn. Then you stay in the dashboard to chat, survey, or re-run the scenario with a tweak.

Under the hood

What powers
the simulation engine.

The dashboard is built to stream results as they happen. Every agent and every conversation is persisted the moment it's generated — you watch the market react, you don't wait for a batch job to finish.

AI & Models

OpenRouterGemini 2.5 FlashClaudeGPT-4oZod validationSelf-correcting retries

Simulation Engine

Phased roundsAgent personasInteraction graphRound-level memoryWhat-if replaysSentiment tracking

Backend & Data

Convex (reactive)Node actionsStreaming insertsPer-agent decisionsSurvey responsesChat history

Frontend

Next.js 15React Flow graphLive subscriptionsClerk authTailwind + shadcnSpanish / English

Want to see your market
argue with itself?

Book a demo and we'll run a live simulation against your real scenario. You walk away with a report, a transcript, and a feel for what your market actually does when you poke it.