DoView drawing prompt
See. Plan. Do.
Quick Overview
This page provides the following prompts and related resources:
A general prompt package for drawing a DoView Board - you download it and drag it into ChatGPT, Codex or Claude to create a DoView Board.
A Claude Skill - you download it, create a Claude Skill, and then use it to create DoView Boards.
The Ten Setup Choices - the ten choices that determine the characteristics of a DoView Board you want your AI system to build. A standard set of the Ten Choices for creating a standard DoView Board for any organization or initiative is already included in the prompts above and is shown below. This standard set of choices was used to create most of the organization or initiative boards here.
Changing the Ten Setup Choices - you can change any of the Ten Setup Choices by just putting in the number of the choice and what you want for it at the time when you tell your AI to create the board.
Other sets of customized Ten Setup Choices for creating boards with different characteristics - you can copy and paste customized sets of the Ten Setup Choices into your AI after you have dragged in the prompt package, or when using the Claude Skill to build a DoView Board. Two customized sets below are:
Research Domain Ten Setup Choices Set (Set 3) - for modeling a research domain. Examples of boards created using this customized set are here.
AI Agent Building Project Ten Setup Choices Set (Set 4) - for planning the building of an AI agent system of any sort. Examples of this type of board are here.
Use of these prompts constitutes acceptance of the Conditions of Use at the bottom of the page.
1. To create an empty DoView Board
3. Prompt for ChatGPT, Codex and Claude
If you just want to open a new empty DoView Board, you do not need a prompt. You can simply open one by clicking here.
2. To update a board to the next version
If you are using a DoView Board, check that it is using the latest version. You can see the version number at the top right of the board. If the version number is less than the latest version (1.3.7), then drag your DoView Board file into your AI system. Download the latest DoView Prompt Package from below, drag it into your AI system and give it the following instruction:
‘Use this DoView Prompt Package to update the engine in this DoView Board to the latest version’.
The prompt below has been tested with ChatGPT, Codex and Claude. At the time of testing with Claude, it can take a significant amount of time to create a board when using Opus 4.6 Max. If you want it to go faster, then use a less powerful model or turn the Opus effort level down. Always regard what is produced by your AI system as a draft, check it, ask your AI to improve it in one or more additional runs, and remember that you can always also edit it by hand.
- Download the DoView Board Prompt Package. Drag it into your AI system. Download The DoView Board Prompt Package And Drag Into Your AI System
- Type 'Build a DoView Board' and then after it responds, give it the name of the board you want to build.
- The prompt will then create a DoView Board using the built-in standard organization or initiative Ten Setup Choices. If you want to build a board with different setup choices just tell your AI system the setup choices you want changed, or use a set of setup choices copied and pasted from those below. Always check the board your AI system builds. For instance if it has not sufficiently mapped projects/activities onto This-Then boxes ask it to redo the board with a richer mapping. For more complex boards, build them in a series of iterations.
- DoView Boards are a free prototype, so read the disclaimer below. Using the prompts on this site constitutes acceptance of the conditions of use.
4. Use this Claude Skill
- Download the DoView Board Claude Skill. Download The DoView Board Claude Skill
- In Claude, turn on Settings > Capabilities > Code Execution and File Creation.
- Go to Customize > Skills, click Add > + Create Skill > Upload a Skill, Drag what you downloaded into the Upload Skill box.
- Start a new Claude chat and type /doview boards. Then type 'Build a DoView Board' and then after it responds, give it the name of the board you want to build.
- The prompt will then create a DoView Board using the built-in organization or initiative Standard Ten Setup Choices which are listed immediately below. If you want to build a board with different setup choices just tell your AI system the setup choices you want changed, or use a set of setup choices copied and pasted from those below. Always check the board your AI system builds. For instance, if it has not sufficiently mapped projects/activities onto This-Then boxes ask it to redo the board with a richer mapping. For more complex boards, build them in a series of iterations.
- DoView Boards are a free prototype, so read the disclaimer below. Use of the Skill constitutes acceptance of the conditions of use.
The Ten Setup Choices
1. Board subject and name
a. What is the DoView Board about?
b. What do you want the DoView Board called?
2. Board scale and level of detail
a. Do you want the This-Then pages to be simple or detailed?
b. How many This-Then pages do you want?
3. Information sources
a. Should the AI use internet research, information you supply, general sector knowledge, or a mixture?
4. This-Then links and evidence
a. Do you want links between This-Then boxes?
b. Do you want information put into the links between This-Then boxes?
5. How pages
a. How many hierarchical Vertical Links How pages do you want?
b. Do you want any Cross-Link How pages?
6. Links involving How pages
a. Do you want links between the How pages and the This-Then pages?
7. Extra display features
a. Do you want any extra Display Text in boxes?
b. Do you want traffic lights?
c. Do you want priorities?
d. What Page View items do you want selected when the board opens?
8. Measurement and evaluation content
a. Do you want measures, indicators or KPIs?
b. Do you want evaluation questions?
9. Documentation pages
a. Do you want Documentation pages?
10. Board note and language style
a. What note should appear on the board?
b. What spelling style, language style or local terminology should be used?
V1.2 Standard Set Number 1
The Standard Ten Setup Choices for a DoView Board for an Organization or Initiative
1. Board subject and name
a. What is the DoView Board about?
[Name of the organization or initiative. This can include a collaboration, partnership, programme, project, service system, reform, strategy, network or other organized work.]
b. What do you want the DoView Board called?
[Name] I will base it on the name you give.
2. Board scale and level of detail
a. Do you want the This-Then pages to be simple or detailed?
Detailed content structure. This means more subject-specific This-Then boxes, richer causal logic, useful page coverage, and relevant measures and evaluation questions where requested. It does not mean opening the board in Detailed Page View or turning on traffic lights, priorities, Display Text, or under-box details unless separately requested.
b. How many This-Then pages do you want?
Normal-sized board. Do not treat this as a fixed number of This-Then pages. Recommend the number of This-Then pages based on the actual scope and complexity of the organization or initiative. Small initiatives may need only a few pages; large agencies, companies, nonprofits, schools, clinics, reform programmes or broad initiatives may need more.
3. Information sources
a. Should the AI use internet research, information you supply, general sector knowledge, or a mixture?
Use public internet information where available, plus general sector knowledge. If internet access or source retrieval is not available in the AI environment, use general sector knowledge, tell the user this has been done, and record the limitation in the assumptions and cautions documentation.
4. This-Then links and evidence
a. Do you want links between This-Then boxes?
Yes.
b. Do you want information put into the links between This-Then boxes?
Yes. Include concise rationale and evidence notes where useful with URLs where appropriate.
5. How pages
a. How many hierarchical Vertical Links How pages do you want?
Two. Level 1: activities, services, projects or workstreams. Level 2: teams, functions, units etc.
b. Do you want any Cross-Link How pages?
Yes. One Cross-Link How page: Competencies.
6. Links involving How pages
a. Do you want links between the How pages and the This-Then pages?
Yes. Put in all relevant links, including:
vertical links between Level 1 How boxes and This-Then boxes;
vertical links between boxes on different How page levels;
cross-links between non-Level 1 How boxes and This-Then boxes;
cross-links between Cross-Link How boxes and This-Then boxes;
cross-links between Cross-Link How boxes and other How boxes.
7. Extra display features
a. Do you want any extra Display Text in boxes?
No.
b. Do you want traffic lights?
No. Do not assign or display traffic lights by default. Leave traffic-light settings unset or off unless the user specifically asks for them. The board must still include the traffic-light feature if supported by the engine.
c. Do you want priorities?
No. Do not assign or display priorities by default. Leave priority settings unset or off unless the user specifically asks for them. The board must still include the priority feature if supported by the engine.
d. What Page View items do you want selected when the board opens?
For This-Then pages only, select: Show link Display Text/Traffic Lights on mouse over of black link arrows.
8. Measurement and evaluation content
a. Do you want measures, indicators or KPIs?
Yes. Include a small number of measures.
b. Do you want evaluation questions?
Yes. Include a small number of evaluation questions.
9. Documentation pages
a. Do you want Documentation pages?
Yes. Include a general documentation, source notes, and cautions Documentation page. Also include an illustrative monitoring and evaluation plan with relevant clones on it.
10. Board note and language style
a. What note should appear on the board?
Proof of Concept - Not Endorsed.
b. What spelling style, language style or local terminology should be used?
Use the spelling and terminology normally used in the relevant country, sector or organization unless the user says otherwise.
V1.2 Customized Set 2
The Ten Setup Choices for a DoView Board for a Research Domain
1. Board subject and name
a. What is the DoView Board about?
[Name of the research domain. This can include a disease area, scientific field, policy research area, technology domain, cross-disciplinary research area, research programme, evidence field or other research domain.]
b. What do you want the DoView Board called?
[Name] Research Domain DoView Board. I will base it on the research domain name you give.
2. Board scale and level of detail
a. Do you want the This-Then pages to be simple or detailed?
Detailed content structure. This means more subject-specific This-Then boxes, richer causal logic, useful page coverage, and relevant measures and evaluation questions where requested. For a research domain, the This-Then pages should cover the main research pathways, evidence-building routes, translation routes, capability-building routes, and intended domain outcomes. It does not mean opening the board in Detailed Page View or turning on optional overlays except where separately requested below.
b. How many This-Then pages do you want?
Normal-sized board. Do not treat this as a fixed number of This-Then pages. Recommend the number of This-Then pages based on the actual scope and complexity of the research domain. Narrow research domains may need only a few pages; broad research domains may need more, but usually fewer than 10 This-Then pages.
3. Information sources
a. Should the AI use internet research, information you supply, general sector knowledge, or a mixture?
Use public internet information where available, plus general research-domain knowledge. Use authoritative and current public sources where possible, including research agendas, systematic reviews, major reports, scholarly sources, public health or policy sources, funder strategy documents, and other relevant domain sources. If internet access or source retrieval is not available in the AI environment, use general research-domain knowledge, tell the user this has been done, and record the limitation in the assumptions and cautions documentation.
4. This-Then links and evidence
a. Do you want links between This-Then boxes?
Yes.
b. Do you want information put into the links between This-Then boxes?
Yes. Include concise rationale and evidence notes where useful with URLs where appropriate. Link Display Text should summarize current research evidence, remaining uncertainties, relevant source URLs, and a traffic light for evidence quality.
5. How pages
a. How many hierarchical Vertical Links How pages do you want?
One. Level 1: main research questions or research priorities for the domain, with Display Text, URLs where available, maturity traffic lights, and priorities.
b. Do you want any Cross-Link How pages?
Yes. One Cross-Link How page: Competencies needed. This should set out the main capabilities, disciplines, methods, infrastructure, partnerships, and other competencies needed in the research domain, prioritized and traffic lighted for current maturity.
6. Links involving How pages
a. Do you want links between the How pages and the This-Then pages?
Yes. Put in all relevant links, including:
vertical links between Level 1 research-question How boxes and This-Then boxes;
cross-links between competency How boxes and research-question How boxes;
cross-links between competency How boxes and This-Then boxes.
Do not include a Level 2 How page unless the user specifically asks for one.
7. Extra display features
a. Do you want any extra Display Text in boxes?
Yes. This-Then boxes, research-question boxes, and competency boxes should include Display Text where useful. Display Text should include concise source notes, assumptions, current evidence or uncertainty notes, and URLs where applicable.
b. Do you want traffic lights?
Yes. Research-question boxes and competency boxes should be traffic lighted for current maturity. This-Then links should be traffic lighted for evidence quality. Ordinary This-Then boxes should not be traffic lighted unless there is a clear reason. Whenever you traffic light provide the rationale for your assessment of the traffic light level in the display text.
c. Do you want priorities?
Yes. Level 1 research-question boxes and Cross-Link competency boxes should be prioritized. Whenever you prioritize provide the rationale for your assessment of the priority in the display text.
d. What Page View items do you want selected when the board opens?
For This-Then pages only, select: Show link Display Text/Traffic Lights on mouse over of black link arrows. For How pages, show traffic lights and priorities where supported.
8. Measurement and evaluation content
a. Do you want measures, indicators or KPIs?
Yes. Include a small number of measures. Measures should be suitable for a research domain and should focus on evidence production, research maturity, uptake or translation, capability development, equity or access where relevant, and progress toward the domain’s intended outcomes.
b. Do you want evaluation questions?
Yes. Include a small number of evaluation questions. Evaluation questions should focus on whether the domain is addressing the most important knowledge gaps, strengthening needed capabilities, producing usable evidence, supporting translation into policy or practice, and contributing to intended outcomes.
9. Documentation pages
a. Do you want Documentation pages?
Yes. Include a general documentation, source notes, assumptions and cautions Documentation page. Also include an illustrative monitoring and evaluation plan with relevant clones on it. Include the standard Using DoView Boards and Disclaimer Documentation page where this is part of the board package.
10. Board note and language style
a. What note should appear on the board?
Proof of Concept - Not Endorsed.
b. What spelling style, language style or local terminology should be used?
Use American spelling unless the user specifically asks for another spelling style. Use terminology appropriate to the research domain, and avoid overstating certainty where the evidence is incomplete, contested, emerging, or context-dependent.
V1.2 Customized Set 3 - Research Domain
The Ten Setup Choices for a DoView Board for Building an AI Agent
1. Board subject and name
a. What is the DoView Board about?
[Name of the AI agent project. This can include an AI agent build, workflow automation agent, customer-facing agent, internal staff-support agent, research agent, operations agent, service agent, decision-support agent, or other AI agent project.]
b. What do you want the DoView Board called?
[Name] AI Agent Project DoView Board. I will base it on the AI agent project name you give.
2. Board scale and level of detail
a. Do you want the This-Then pages to be simple or detailed?
Detailed content structure. This means more subject-specific This-Then boxes, richer causal logic, useful page coverage, and relevant measures and evaluation questions where requested. For an AI agent project, the This-Then pages should cover the main design, build, data, tool, testing, deployment, safety, adoption, monitoring and improvement pathways. It does not mean opening the board in Detailed Page View or turning on traffic lights, priorities, Display Text, or under-box details unless separately requested.
b. How many This-Then pages do you want?
Normal-sized board. Do not treat this as a fixed number of This-Then pages. Recommend the number of This-Then pages based on the actual scope and complexity of the AI agent project. Simple agents may need only a few pages; broader agents involving multiple workflows, systems, user groups, risks, integrations or governance requirements may need more.
3. Information sources
a. Should the AI use internet research, information you supply, general sector knowledge, or a mixture?
Use general AI agent build knowledge plus credible public internet information where available. Use authoritative and current sources where possible, including AI safety, evaluation, privacy, security, software engineering, human-centered design, workflow automation, product management and sector-specific sources. If internet access or source retrieval is not available in the AI environment, use general AI agent project knowledge, tell the user this has been done, and record the limitation in the assumptions and cautions documentation.
4. This-Then links and evidence
a. Do you want links between This-Then boxes?
Yes.
b. Do you want information put into the links between This-Then boxes?
Yes. Include concise rationale and evidence notes where useful with URLs where appropriate. Link Display Text should explain the build logic, implementation rationale, testing or evaluation basis, known risks, assumptions and relevant source notes.
5. How pages
a. How many hierarchical Vertical Links How pages do you want?
Two. Level 1: AI agent build workstreams, components or project phases. Level 2: roles, systems, data sources, tools, models, interfaces, governance groups, users and other participants.
b. Do you want any Cross-Link How pages?
Yes. One Cross-Link How page: AI Agent Project Competencies. This should set out the main capabilities needed for the project, such as product ownership, workflow design, prompt and agent design, data engineering, integration engineering, evaluation, safety, privacy, security, change management, user support and operational monitoring.
6. Links involving How pages
a. Do you want links between the How pages and the This-Then pages?
Yes. Put in all relevant links, including:
vertical links between Level 1 AI agent build workstream boxes and This-Then boxes;
vertical links between boxes on different How page levels;
cross-links between non-Level 1 How boxes and This-Then boxes;
cross-links between AI Agent Project Competency boxes and This-Then boxes;
cross-links between AI Agent Project Competency boxes and other How boxes.
7. Extra display features
a. Do you want any extra Display Text in boxes?
No.
b. Do you want traffic lights?
No. Do not assign or display traffic lights by default. Leave traffic-light settings unset or off unless the user specifically asks for them. The board may still include the traffic-light feature if supported by the engine.
c. Do you want priorities?
No. Do not assign or display priorities by default. Leave priority settings unset or off unless the user specifically asks for them. The board may still include the priority feature if supported by the engine.
d. What Page View items do you want selected when the board opens?
For This-Then pages only, select: Show link Display Text/Traffic Lights on mouse over of black link arrows.
8. Measurement and evaluation content
a. Do you want measures, indicators or KPIs?
Yes. Include a small number of measures. Measures should be suitable for an AI agent project and may include task completion rate, correct handoff or escalation rate, human-review rate, average response or action time, user helpfulness score, outputs requiring correction, unresolved or reopened task rate, privacy or safety incident rate, knowledge-gap or tool-gap closure rate, adoption rate, and operational reliability.
b. Do you want evaluation questions?
Yes. Include a small number of evaluation questions. Evaluation questions should focus on whether the agent is solving the right workflow problems, using approved knowledge and tools accurately, routing high-risk or uncertain cases to humans quickly enough, producing outputs staff and users can trust, detecting privacy and safety risks before harm occurs, improving service or workflow quality, and avoiding hidden unresolved work or inappropriate automation.
9. Documentation pages
a. Do you want Documentation pages?
Yes. Include a general documentation, source notes, assumptions and cautions Documentation page. Also include agent design notes, risk and guardrail notes, testing and evaluation notes, monitoring and improvement notes, and an illustrative monitoring and evaluation plan with relevant clones on it.
10. Board note and language style
a. What note should appear on the board?
Proof of Concept - Not Endorsed.
b. What spelling style, language style or local terminology should be used?
Use American spelling unless the user specifically asks for another spelling style. Use plain language appropriate to the AI agent project, the relevant sector and the intended users. Avoid overstating agent capability, reliability, autonomy or safety where these depend on testing, deployment context, monitoring or human oversight.
V1.2 Customized Set 4 - AI Agent Project
Disclaimer and Conditions of use
License, attribution, and disclaimer
These conditions apply when you use the DoView Boards prompt package, the DoView PowerPoint prompt, or materials created from them.
You may use these prompts to create DoView Boards, DoView PowerPoints, DoViews, and related DoView-based planning, outcomes, implementation, monitoring, evaluation, and learning materials. You may use, adapt, share, present, teach from, consult with, and build on the materials you create, including for internal work, client work, training, research, reports, presentations, software, apps, platforms, systems, and AI tools. No licence fee is required for these uses.
Please acknowledge DoView Planning and Dr Paul Duignan when you use the approach or create DoViews, DoView Boards, DoView PowerPoints, or related DoView-based models. A suitable acknowledgment is:
This work uses the DoView Planning approach and DoView Boards developed by Dr Paul Duignan. See DoViewPlanning.org.
DoView developer resources, including the DoView Boards GitHub developer package, are released under Apache-2.0. Other prompt-page materials are made available under these conditions unless a specific file says otherwise.
You may accurately describe your work as using DoView Planning, creating DoViews, supporting DoView Boards, or being DoView-compatible, provided this is accurate, but please do not imply official DoView® endorsement, certification, approval, quality assurance, affiliation, or badge rights.
Attribution to DoViewPlanning.org, or use of the DoView® name, does not mean that DoViewPlanning.org, Dr Paul Duignan, The Ideas Web Limited, or DoView Corporation Limited endorses your adaptation, your use of it, or any outputs produced using it. Use of DoView® Marks and Materials is governed by the DoView® Planning Attribution & Trademark Use Policy: https://doviewplanning.org/trademarkuse. A short explanation is here: https://doviewplanning.org/doviewnameuse.
The prompts and related materials are provided free “as is” for general information only. They are not professional advice, including legal, financial, management, evaluation, security, or other professional advice. No warranties of accuracy, completeness, security, reliability, suitability, or fitness for purpose are given.
You are solely responsible for how you use the prompts and any outputs, including checking outputs, exercising professional judgment, complying with applicable laws and policies, and avoiding disclosure of confidential, proprietary, personal, sensitive, or regulated data.
Generated standalone DoView Board .html files are active HTML/JavaScript files. Treat them like executable web content, not passive documents. Do not use them for sensitive, confidential, regulated, public, multi-user, enterprise, or production purposes unless you have security, privacy, hosting, access-control, audit, and deployment arrangements appropriate to that context.
Dr Paul Duignan, The Ideas Web Limited, and DoView Corporation Limited and associated parties accept no liability for any loss or damage arising from use of the prompts, related materials, adaptations, or outputs produced using them.
Use of these prompts constitutes acceptance of these conditions and does not create any professional, advisory, client, partnership, employment, endorsement, certification, approval, quality-assurance, or agency relationship. v2