Skip to content

markoblogo/decision-map

Repository files navigation

DecisionMap

DecisionMap Logo

License: MIT GitHub stars Last commit

A practical protocol for using AI to map complex business and product decisions.

DecisionMap is a protocol + prompt toolkit, not a SaaS product and not a new AI model.

It is a structured way to think through decisions where there is no single correct answer, only trade-offs under uncertainty.


What it is

DecisionMap helps you turn messy, high-stakes business, product, market, and marketing situations into a map of strategic options.

It is especially useful when the problem is not lack of intelligence, but a fragmented or distorted picture of reality.

Instead of asking:

“What should we do?”

it reframes the problem as:

“What are our real options, and what does each of them cost?”

Each option is mapped by:

  • expected upside
  • cost (money, time, reputation, risk)
  • required resources
  • likely reactions from competitors, customers, or partners
  • short / mid / long-term effects
  • assumptions
  • breakpoints (where it fails)
  • signals to monitor

The core output is not a final answer.

It is a working strategic hypothesis, with visible trade-offs.


What it is not

DecisionMap is not:

  • a chatbot that gives advice
  • a prediction engine
  • a replacement for decision-makers
  • a “smart agent” that thinks for you

It does not:

  • guarantee outcomes
  • remove uncertainty
  • make decisions on your behalf

And it is intentionally out of scope for:

  • military or political conflict
  • legal or medical advice
  • financial investment decisions
  • M&A, layoffs, or HR restructuring

When to use it

Use DecisionMap when:

  • you have multiple plausible strategies and no clear winner
  • the decision involves trade-offs, not right vs wrong
  • competitors or external reactions matter
  • you feel stuck because the picture is unclear

How it differs from a normal AI chat

Many AI chats jump from context to a clean recommendation too quickly.

DecisionMap forces a disciplined process:

  1. clarify the problem
  2. separate facts, assumptions, interpretations, and unknowns
  3. ask only questions that can change the strategy map
  4. build multiple realistic strategies
  5. compare trade-offs, resources, risks, and breakpoints
  6. pressure-test shortlisted options
  7. produce a decision record and working hypothesis

This is slower, but better aligned with real strategic decisions.


Usage

For a full copy-paste workflow with any LLM, see USAGE.md.

Quick start (manual)

You don’t need any app.

  1. Set the system prompt from prompts/system_prompt.md
  2. Run 01_intake.md with your situation
  3. Answer 02_clarifying_questions.md
  4. Generate options via 03_strategy_map.md
  5. Deep dive with 04_deep_dive.md
  6. Finalize with 05_decision_summary.md

Optional: track updates using schemas/cascade_log.schema.json.

For the full step-by-step version, use USAGE.md.


Examples


ABVX ecosystem

DecisionMap can be used as a standalone protocol with any LLM.

Inside the ABVX ecosystem:

  • lab.abvx lists it as a decision/strategy protocol artifact
  • agentsgen can maintain repo-local agent docs for contributors
  • SET can track/audit the repository as part of orchestration flows
  • ID can optionally provide portable user context for long-running decision work

None of these integrations are required for manual use.

Integration status

  • lab.abvx: added as supporting tool in Decision & Strategy Protocols (repo + landing card).
  • SET: added as registry entry for tracking/audit (repo-docs baseline only; no runtime orchestration).
  • ID: added as optional reference/link integration (no hard dependency).
  • agentsgen: repo-local docs generated (AGENTS.md, RUNBOOK.md, .agentsgen.json).

Optional ABVX integration

  • DecisionMap stays standalone and usable with any LLM.
  • ID is an optional context layer, not a dependency.
  • SET should track and audit this repo, but not run runtime orchestration yet.
  • agentsgen is the preferred path for repo-local agent-facing docs (AGENTS.md, RUNBOOK.md, .agentsgen.json).
  • lab.abvx should position DecisionMap as a supporting tool in Decision & Strategy Protocols, not as core stack infrastructure.

What is included in v0.1

  • protocol
  • reusable prompts
  • JSON schemas
  • manual usage workflow
  • example cases
  • Codex notes for a future mini-tool

What is not included yet

  • web app
  • hosted service
  • authentication
  • persistent project memory
  • LLM integration
  • local model runner

Repository structure

decision-map/
├── README.md
├── USAGE.md
├── protocol.md
├── prompts/
│   ├── system_prompt.md
│   ├── 01_intake.md
│   ├── 02_clarifying_questions.md
│   ├── 03_strategy_map.md
│   ├── 04_deep_dive.md
│   └── 05_decision_summary.md
├── schemas/
│   ├── strategy_map.schema.json
│   └── cascade_log.schema.json
├── examples/
│   ├── full_run_product_launch.md
│   ├── product_launch.md
│   └── competitive_response.md
├── codex/
│   ├── build_prompt.md
│   └── implementation_notes.md
├── LICENSE
└── .gitignore

Privacy

DecisionMap does not require storing data.

However:

  • most hosted LLM APIs process data externally
  • your input may be processed by third-party providers

For sensitive work:

  • anonymize names, companies, exact financials, customer data, and internal documents
  • or run in a local model / approved internal environment

Future direction (optional)

Future versions may include:

  • project mode
  • cascade logs for long-running decisions
  • structured memory of decisions, assumptions, signals, outcomes, and revisions

But the core value is already here:

  • options, not answers
  • working hypothesis, not final truth
  • clarity under uncertainty

Version

Current status: v0.1 protocol draft.