Static Documents Are Dead: Why Educators Should Start Building Websites, Dashboards, and Apps
Key Takeaways
Documents, spreadsheets, and slide decks still matter but they should no longer be the automatic endpoint for most deliverables.
The important shift is from files people receive to systems people can use.
Educators and researchers can now turn public, nonconfidential material into interactive websites, dashboards, self-audits, decision tools, and resource hubs.
AI can help build and maintain those systems, but professional review, accessibility, privacy, security, and institutional policy still matter.
Static documents are dead - or at least they should no longer be our default.
Educators, researchers, and consultants still exchange information primarily through documents, PDFs, Excel spreadsheets, and slide decks. We create a file, attach it to an email, and consider the work finished.
But a static file is often a poor interface for information.
A spreadsheet gives someone rows and columns. A dashboard lets them explore the data.
A report tells someone what you found. An interactive website lets them compare findings, filter results, follow different paths, and return when the information changes.
A checklist tells someone what to consider. An application can collect responses, calculate scores, identify risks, generate recommendations, preserve progress, and create a report.
The important shift is not from Word to a prettier webpage.
It is from files people receive to systems people can use.
I wanted to test whether this shift was real
When GPT-5.6 became available, I wanted to test what the newest generation of AI-assisted development tools could actually do.
OpenAI describes GPT-5.6 as improving frontend design, including layout, visual hierarchy, usability, and design judgment. It also describes the model as more capable in complex production workflows and more efficient in its use of output tokens. (Learn ChatGPT (https://learn.chatgpt.com/api/docs/guides/latest-model))
Rather than run another collection of abstract prompts, I gave myself a practical challenge:
Could I build a professional website and interactive portfolio in a day?
Using Codex and ChatGPT Sites, I created a working website for my consulting practice. I organized my services, research, selected projects, professional experience, and ways that organizations can work with me. I added separate pages for different kinds of work and created a tailored discussion page for a particular audience.
You can explore the interactive dashboards, readiness audit, and related examples in my selected work portfolio (https://lockwood-consulting-portfolio-review.adamblockwood.chatgpt.site/work).
More importantly, I incorporated interactive tools rather than relying exclusively on descriptions.
The site includes examples such as:
Interactive survey dashboards
Data visualizations
A browser-based ethical AI readiness audit
A working early AI-assisted report-writing prototype
Research and consulting case studies
Audience-specific briefing pages
Links to live professional and research resources
The result was not perfect after one day.
That was not the point.
The point was that I could move from an idea to a working, editable, hosted system in hours rather than waiting weeks to hire a developer, define every detail in advance, or learn an entire web-development stack.
A website is no longer just a website
Most educators hear “build a website” and picture a homepage, a biography, several links, and perhaps a contact form.
That is an increasingly narrow view.
ChatGPT Sites can create, host, refine, and share websites, web applications, and games from a prompt or compatible project. Sites remain editable after they are created, and the underlying project can be revised and tested through Codex before a new version is deployed. (Learn ChatGPT (https://learn.chatgpt.com/docs/sites))
That means an educator can now build something closer to a small application than a traditional webpage.
Consider the difference.
Instead of a static research report
Build an interactive dashboard where users can:
Explore the findings
Compare groups
Filter by variables
Examine trends
Review qualitative themes
Download selected information
Return when new data are added
Instead of a checklist
Build a self-audit that can:
Collect responses
Calculate scores
Flag areas of concern
Generate recommendations
Save progress
Export a report
Link each item to supporting evidence
Dashboards change what research deliverables can be
Researchers frequently devote enormous effort to designing studies, collecting data, cleaning files, conducting analyses, and interpreting results.
Then we often deliver the final product as a collection of tables in a manuscript, a spreadsheet, or a report.
Those outputs may be necessary, but they do not always make the findings easy to use.
A dashboard can turn research into a decision tool.
For example, survey findings can be organized so that an organization can explore:
Provider experiences
Customer needs
Adoption patterns
Training priorities
Workflow bottlenecks
Differences across professional groups
Implementation challenges
Changes over time
The goal is not simply to make the data more attractive.
The goal is to make the evidence more accessible and actionable.
My preferred model is:
Collect → Analyze → Visualize → Interpret → Automate → Act
The spreadsheet may still exist underneath. The technical report may still be necessary. But neither must remain the primary way that people encounter or use the findings.
We are moving toward agentic information systems
Interactive sites and dashboards are only part of the change.
The next step is agentic.
In a traditional workflow, someone must:
Export the data
Clean the spreadsheet
Conduct the analysis
Create the charts
Write the report
Email the attachment
Repeat the process when new data arrive
An AI-enabled system can increasingly support that process continuously.
It can help:
Monitor new submissions
Identify missing or inconsistent information
Update analyses
Refresh visualizations
Detect emerging patterns
Generate summaries for different audiences
Flag issues requiring professional review
Maintain an evolving resource site
Prepare decision-ready updates
We should stop thinking about AI only as something that writes documents.
We should start thinking about AI as something that helps us build and maintain systems.
What educators could build
Every educator should consider whether something they currently distribute as a file would work better as an interactive resource.
Potential examples include:
A course website that evolves throughout the semester
A family-facing guide organized around questions
A searchable intervention library
A research dashboard
A workshop companion site
An interactive rubric
A simulation
A readiness assessment
A public project portfolio
A professional-development toolkit
A continuously updated policy or resource hub
An application that scores and interprets nonconfidential information
This does not require every educator to become a professional software developer.
It requires educators to understand their audience, define the desired experience, organize the source information, establish boundaries, and iteratively review what the system produces.
Those are already familiar instructional and research skills.
How I built mine
My process was iterative rather than magical.
1. I clarified the purpose
I identified:
Who the site was for
What I wanted visitors to understand
What kinds of work I wanted to demonstrate
Which information was public
Which ideas required a restricted or unlisted page
2. I created a local project in Codex
Rather than generate one disposable webpage, I asked Codex to build a maintainable project with:
Reusable page components
Structured content
Documentation
Version control
Testing
Accessibility requirements
Clear confidentiality rules
3. I added source materials
I supplied:
My curriculum vitae
Approved descriptions of consulting work
Presentation evaluations
Project files
Dashboard code
A prototype pitch deck
Links to professional profiles and research materials
4. I reviewed the output repeatedly
The first version was not the final version.
I found:
Pages that were too crowded
Sections that were too vague
Inconsistent terminology
Navigation problems
Broken images
Cards that were visually unbalanced
Internal drafting notes appearing publicly
Descriptions that sounded like consultant jargon
I gave Codex focused revision instructions, tested the changes, and reviewed the next version.
5. I added working tools
I did not want the portfolio merely to say that I could build dashboards, applications, and decision tools.
I wanted it to show them.
That meant adding:
Live dashboards
An interactive organizational audit
A video of an earlier proof of concept
Project overviews explaining what I contributed
6. I published through Sites
Sites allowed the project to be hosted, reopened, refined, versioned, and shared. OpenAI notes that every deployed Sites URL is a production deployment, so review and access decisions still matter. (Learn ChatGPT (https://learn.chatgpt.com/docs/sites))
What educators should be cautious about
The fact that something can be built quickly does not mean it should be published carelessly.
The easiest place to begin is with:
Public information
Nonconfidential resources
De-identified or aggregate data
Materials you have permission to share
Fictional demonstration data
Content that does not involve protected student, client, patient, or employee information.
Educators also need to consider:
Privacy
Data security
Accessibility
Accuracy
Copyright
Institutional policy
Records requirements
Professional responsibility
Who can access the site
Whether the system stores or transmits data
A hidden link is not the same as secure access.
A polished interface is not evidence that a tool is valid.
An AI-generated application still requires expert review.
Static files will not disappear tomorrow
There will continue to be good reasons to use PDFs, Word documents, and spreadsheets.
They remain useful for:
Archival records
Formal submissions
Legal documentation
Technical appendices
Raw data
Portability
Offline access
Standardized reporting
My argument is not that these formats must literally vanish.
It is that they should no longer be our automatic endpoint.
The question should be:
What format will make this information most useful to the people who need it?
Sometimes that answer will be a document.
Increasingly, it may be a website, dashboard, application, or continuously updated system.
Stop delivering files. Start building systems.
The barrier between an idea and a working interactive resource has dropped dramatically.
That creates an opportunity for educators, researchers, clinicians, and professional organizations to rethink what they produce.
Do not begin with:
“How can AI help me write this document?”
Begin with:
“What should the user be able to understand, explore, decide, or do?”
Then build the experience around that goal.
I built my site in a day because I wanted to see what was possible.
What I found was not simply a faster way to create a webpage.
It was a different way to think about professional work.
Static documents are no longer enough.
The future deliverable is often not a file.
It is a system.
Final Key Takeaways
Keep files where they serve archival, legal, technical, or compliance needs. Do not let format become the constraint on usefulness.
Start with the user experience: what should someone be able to understand, explore, decide, or do?
Begin with public, nonconfidential, appropriately approved information, then test and improve the experience with real users.
AI can lower the barrier to building useful systems. It does not remove the need for professional judgment and accountability.
AI Disclosure: Generative AI was used to assist with drafting and editing this post and to create the accompanying image. I reviewed and revised the content and take responsibility for its accuracy and final form.