Render Unstructured Content in StratML

A practical prompting guide for using AI/LLM services to convert plans, reports, webpages, articles, dialogue, and other narrative content into open, standard, machine-readable StratML.

Quick Start

The goal is to make StratML conversion as easy as possible without pretending that AI-generated markup should go unreviewed.

1. ChooseSelect StratML Part 1 for strategic intentions or Part 2 when stakeholder roles, value chains, indicators, targets, and actual results are also needed.
2. CopyUse one of the tested prompts below, or generate a shorter customized prompt with the prompt builder.
3. ProvideGive the AI service a URL, file, pasted text, or other source material to render.
4. ReviewCheck fidelity to the source, then validate the XML against the applicable StratML schema.
Core principle: Treat StratML as a semantic representation of the source, not merely an XML container for its prose. Preserve intentions, stakeholders, relationships, qualifications, uncertainties, and measurable results faithfully.

StratML Part 1 — Reusable Prompt

Render the supplied source content in StratML Part 1 format conforming to ISO 17469-1. Preserve the intentions expressed in the source as faithfully as possible. Do not silently invent goals, objectives, stakeholders, values, or other substantive content. Where structure must be inferred from narrative material, make reasonable inferences only when necessary and document significant inferences in the appropriate OtherInformation elements. Apply these conventions: • Use nouns or noun phrases for Goal and Objective Names. • Begin Goal and Objective Descriptions with active verbs expressing the intended result. • Distinguish objectives (intended results) from activities, methods, organizations, topics, and resources. • Identify stakeholders explicitly mentioned or reasonably implied by the source and classify them appropriately as Person, Organization, or Generic_Group. • Place stakeholders at the plan, goal, or objective level where their roles are most directly relevant. • Capture important source detail that does not fit naturally elsewhere in OtherInformation rather than discarding it. • Preserve useful source URLs and references. • Assign globally unique identifiers to elements requiring identifiers. • Provide SequenceIndicators for Goals and Objectives. • Do not include HTML markup in the XML. • Do not fabricate facts merely to populate optional elements. • Preserve distinctions, qualifications, uncertainties, inconsistencies, and limitations rather than silently resolving them. If the source is lengthy, provide the XML in manageable sequential segments so it can be reviewed and validated incrementally. Begin with the opening elements, organization, vision, mission, values, and first goal. Wait for “Next” before continuing. Before finalizing, check XML well-formedness and conformance to the supplied or referenced ISO 17469-1 schema.

StratML Part 2 — Reusable Prompt

Render or extend the supplied source content in StratML Part 2 format using the applicable StratML Part 2 schema. Preserve the source faithfully. Do not invent performance claims, targets, baselines, dates, actual results, stakeholder roles, or quantitative values merely to populate the schema. In addition to the applicable Part 1 conventions: • Identify stakeholder roles relevant to each objective. • Represent inputs, processing, outputs, outcomes, and other value-chain information where supported by the source. • Derive performance indicators only from measurable results stated or reasonably supported by the source. • Use UnitOfMeasurement only for quantitative indicators. • Use a short noun or noun phrase for MeasurementDimension. • When a numeric value is supported, capture it in NumberOfUnits. • Preserve useful percentage or explanatory wording in the Target or Actual Description while also capturing machine-processable numeric values where possible. • Distinguish targets from actual results. • When the source provides a baseline actual result, preserve its applicable date. • Do not manufacture missing targets, baselines, dates, or actual values. • For qualitative indicators, do not fabricate quantitative units. • Document calculations, conversions, ambiguities, source inconsistencies, inferred values, and other material qualifications in OtherInformation. If percentages, rates, or ratios can legitimately be converted into NumberOfUnits from other source values, calculate them when useful, preserve the original percentage or rate in the Description, and document the calculation in OtherInformation. If the source is lengthy or structurally complex, provide the XML in manageable sequential segments and wait for “Next” before continuing. Before finalizing, check XML well-formedness and conformance to the supplied Part 2 schema.

Simple Prompt Builder

For users who do not want to edit a long prompt, choose the options below. The page will assemble a compact instruction that can be copied into an AI service.

Generated Prompt

Choose your options and click “Generate Prompt.”

Open an AI Service

For now, the safest service-independent workflow is copy prompt → open service → provide source. Some AI services also support reusable configured assistants that can eventually reduce this to nearly one click.

ChatGPT

Use the copied prompt in a normal ChatGPT conversation. Reusable GPT configurations may also be shareable by link where account and workspace settings permit.

Open ChatGPT

Claude

Use the copied prompt with Claude and provide the source as a URL, pasted text, or supported attachment.

Open Claude

Gemini

Use the copied prompt with Gemini. Google also supports shareable Gems, which may provide a convenient reusable StratML-specific configuration.

Open Gemini
Why not inject the full prompt automatically into every service? Public deep-link mechanisms differ among providers and can change. Keeping the master prompts on this page provides a stable, provider-independent starting point while allowing direct configured-assistant links to be added when available.

Review & Validate

AI can accelerate the conversion, but human review remains important. Check the rendition against the source, especially where the AI has inferred structure or calculated values.

Part 1 Form

Use the StratML Part 1 form to review or edit strategic-plan content.

Open Part 1 Form

Part 2 Form

Use the Part 2 form to flesh out stakeholder roles, value chains, performance indicators, targets, and actual results.

Open Part 2 Form

Discover Existing Plans

Search existing StratML goals, objectives, stakeholders, values, sources, and related content before creating something duplicative.

Search StratML

OBVO Example

See the Objective-Based Voluntary Organization plan for an example of a machine-readable plan that also identifies supporting tools and services.

View OBVO Plan

Final Review Prompt

Review the complete StratML draft against both the source material and the applicable StratML schema. Check particularly for missing or duplicated goals, objectives, stakeholders, indicators, or source information; Goal and Objective Names that are not noun phrases; Descriptions that fail to express intended results with active verbs; activities or methods incorrectly represented as objectives; stakeholder roles represented at inappropriate levels; missing or duplicate SequenceIndicators; identifiers that are not globally unique; unsupported AI-generated assertions; source qualifications, inconsistencies, or uncertainties that have been lost; and XML well-formedness or schema-validity problems. Do not rewrite valid content merely for stylistic preference. Recommend only corrections that materially improve fidelity, validity, completeness, interoperability, or clarity.