INPUT – Adventures in AI

TEXT
Plain or Markdown-formatted text is the most common input. It might be instructions, questions, ideas, or early drafts. It’s how you tell the AI what you want, whether by exploring a topic, refining a draft, or brainstorming possible solutions.

Question to reflect on: What types of text do you write or read — like drafts or notes — or questions you ask, such as exploring a topic or brainstorming solutions?

STRUCTURED TEXT
This is text arranged in a fixed format, often exported from another system. Examples include CSV survey results, XML files, or JSON pulled from a website. AI can interpret and work with these structures, which makes them useful when you’re dealing with lists, tables, or other organised data.

Question to reflect on: Do you often work with tables, like contact lists or survey results? What types of structured data could you export from your current tools?

FILE
You can upload documents for the AI to process — Word reports, PDFs, presentations, Markdown blog drafts, or even code files. These let you bring in real materials from your work and have the AI help you review, summarise, edit, or extend them.

Question to reflect on: What documents do you work with — such as reports, presentations, or contracts?

CODE
Sometimes the input is instructions for software: a code snippet, an Excel formula, or a VBA macro. The AI can help you understand, debug, or extend it. Even if you’re not a coder, you may encounter small pieces of code inside reports or spreadsheets that you’d like to make sense of.

Question to reflect on: Do you ever encounter existing code — like formulas in reports, macros in Excel, or scripts — that you need to debug or understand?

IMAGE
AI can also take in visuals, such as product photos, scanned documents, charts, or diagrams. It might then analyse, extract details, or enhance them. Screenshots or photos that feel messy to handle manually can become usable inputs for new insights or outputs.

Question to reflect on: What visual materials — like screenshots, charts, or photos — do you work with that could benefit from being analysed or enhanced?

AUDIO
Recorded sound, like a podcast, a lecture, or a voice memo, can also serve as input. AI can transcribe it, summarise it, or reformat it for different uses. This helps when you’d rather capture ideas verbally, or when spoken content needs to be made searchable or shareable.

Question to reflect on: What audio materials — like podcasts, lectures, or voice memos — do you work with that could benefit from being analysed or enhanced?

VIDEO
Video is another input option: demos, tutorials, or recorded lectures. AI can break these down into transcripts, summaries, highlight reels, or insights. This saves time when long recordings contain important details but are impractical to watch in full.

Question to reflect on: What video materials — like product demos, tutorials, or lectures — do you work with that could benefit from being analysed or enhanced?

INTERACTIVE INPUT
Not all inputs are static files — some come from live interaction. This might include chat messages in a conversation, keystrokes in a form, or clicks in an online tool. They provide a stream of information that AI can respond to in real time, making the process dynamic rather than one-off.

MIXED / MULTIMODAL INPUT
Sometimes the starting point is a combination: a report with embedded charts, a training video with slides, or a research pack of documents, images, and transcripts. AI can take in these mixed formats together, preserving their context and connections while analysing or transforming them.

PHYSICAL INPUT (DIGITAL CAPTURE)
Physical materials can also become inputs once digitised. A scanned contract, a photographed whiteboard, or a 3D scan of a prototype are examples. AI works with the captured digital representation, extending its usefulness beyond the physical original.