How Can AI Analyze an Image?
Artificial intelligence can analyze an image by processing visual information and identifying patterns, objects, text, colors, shapes, people, scenes, and relationships between different elements. Modern AI vision systems can go beyond simple image recognition. Depending on the system and the task, they can describe what appears in a picture, read text from an image, answer questions about visual content, classify objects, compare images, identify visual patterns, and combine image information with natural-language reasoning.
AI image analysis is used in many areas, including search engines, smartphones, security systems, healthcare research, manufacturing, education, accessibility tools, retail, robotics, document processing, creative applications, and scientific research. A person can also use an AI assistant to upload an image and ask questions about what it contains.
The basic idea is to convert visual information into representations that an AI model can process. The system can then use those representations to identify features and relationships and generate an interpretation or response. The exact technology varies between AI systems, but computer vision, machine learning, deep learning, neural networks, image recognition, optical character recognition, and multimodal AI are important parts of modern image analysis.
AI image analysis is powerful, but it is not infallible. An image may be blurry, poorly illuminated, cropped, ambiguous, misleading, or difficult to interpret. AI can also misidentify objects or infer information that cannot actually be established from the image. Important decisions should therefore include human verification and, where appropriate, additional evidence.
This guide explains how AI analyzes images, how computer vision works, how AI recognizes objects and text, how multimodal models interpret pictures, what image analysis can be used for, how to write useful image-analysis prompts, and what limitations users should understand.
What Is AI Image Analysis?
AI image analysis is the use of artificial intelligence to extract meaningful information from visual data. Instead of treating an image as a collection of pixels only, an AI system attempts to identify structures and patterns that correspond to objects, text, scenes, or other visual concepts.
For example, if an AI system receives a photograph of a bicycle beside a road, it may identify the bicycle, road, surrounding vegetation, sky, and other visible elements. If the image contains a sign, an AI system with suitable visual and text-recognition capabilities may also read the words on the sign.
Image analysis can be performed for a specific task or as part of a broader question-answering process. A specialized computer-vision model may classify an object, while a multimodal AI assistant may describe the entire scene and answer follow-up questions about it.
How Does AI Analyze an Image?
The process can be understood as a sequence of related stages. First, the system receives the image. The image is represented digitally as visual data. A trained neural network then processes patterns within that data and creates internal representations that can be used for recognition or reasoning.
The system may detect visual features such as edges, textures, shapes, colors, spatial relationships, and increasingly complex structures. Higher-level representations can correspond to objects, faces, written characters, scenes, or activities.
Once useful visual representations have been created, the AI can perform the requested task. That task might involve classification, object detection, segmentation, text extraction, image captioning, comparison, visual question answering, or another form of analysis.
Generative multimodal systems can then translate the result into natural language. This is why a user can upload an image and ask a question in ordinary language rather than having to operate a specialized computer-vision application.
What Is Computer Vision?
Computer vision is the field of artificial intelligence and computer science concerned with enabling computers to process and interpret visual information. It includes techniques for recognizing objects, detecting patterns, understanding scenes, analyzing images and video, and extracting information from visual data.
Traditional computer-vision systems often relied heavily on manually designed features and rules. Modern systems commonly use machine learning and deep neural networks that learn useful representations from large collections of training examples.
Computer vision is now used in applications such as image search, autonomous systems, quality inspection, medical-image research, document processing, facial recognition, agriculture, robotics, and accessibility technology.
How Do Neural Networks Help AI Understand Images?
Neural networks are computational models that can learn patterns from data. In image analysis, neural networks can learn representations of visual structures through training.
Earlier layers of a vision network can respond to relatively simple patterns such as edges or textures, while deeper layers can represent increasingly complex structures. The exact internal behavior is more complicated than this simplified explanation, but the general idea is that the model builds increasingly useful representations from visual input.
Convolutional neural networks became particularly important in computer vision because they can efficiently process spatial patterns. More recent vision systems may use transformer-based architectures or combinations of different model components.
What Is Image Recognition?
Image recognition is the process of identifying what an image contains or determining which category an image belongs to. For example, an image-recognition model may determine whether an image contains a dog, a car, a building, or a landscape.
Classification generally assigns one or more labels to an image. If an image contains several relevant objects, more advanced systems can identify multiple objects rather than assigning a single overall label.
Image recognition is useful for organizing image libraries, searching visual databases, filtering content, assisting accessibility applications, and automating repetitive visual tasks.
What Is Object Detection?
Object detection goes beyond recognizing that an object exists. It attempts to locate individual objects within an image, commonly by identifying their approximate positions with bounding boxes or other spatial representations.
For example, a traffic image could contain several cars, pedestrians, bicycles, and traffic signs. An object-detection system can identify multiple objects and determine where they appear in the image.
This capability is useful in robotics, manufacturing, traffic analysis, retail systems, agriculture, and many other computer-vision applications.
What Is Image Segmentation?
Image segmentation divides an image into meaningful regions or assigns labels to individual pixels or groups of pixels. Instead of simply drawing a box around an object, segmentation can identify the exact visible area associated with that object.
For example, a segmentation system may distinguish the road from vehicles, pedestrians, buildings, and vegetation. In scientific and medical-image research, segmentation can help researchers identify particular structures or regions within an image.
Segmentation is especially useful when the exact shape or boundary of an object matters.
How Can AI Read Text From an Image?
AI can extract text from images through optical character recognition, commonly called OCR. OCR systems analyze visual representations of characters and convert them into machine-readable text.
This allows software to process information from scanned documents, photographs of signs, receipts, forms, labels, books, screenshots, and other images containing text.
Modern multimodal AI systems can combine OCR-like capabilities with broader visual understanding. Instead of simply extracting every word, they may also explain what the text means in the context of the image.
However, OCR accuracy can be affected by handwriting, unusual fonts, low image quality, shadows, distortion, small text, and complex backgrounds.
Can AI Understand a Screenshot?
AI can analyze many screenshots when the system supports image input. It may identify visible text, interface elements, buttons, charts, error messages, icons, and layout relationships.
For example, a user may provide a software error screenshot and ask the AI to explain what the visible message means. A user can also ask an AI system to identify information displayed in a dashboard or describe the structure of an application interface.
For technical troubleshooting, the AI's interpretation should be combined with the actual software documentation and system information because a screenshot may not contain enough context to determine the underlying cause of an issue.
Can AI Describe an Image?
Yes. Image captioning and multimodal AI systems can generate descriptions of images. A simple description might identify the major objects and setting, while a more detailed description can discuss visible relationships, activities, text, colors, and other relevant features.
The desired description should be specified by the user. A person preparing accessibility text may need a concise objective description. A researcher may want a detailed account of visible features. A designer may want a description of composition and visual elements.
How Does AI Answer Questions About an Image?
Visual question answering combines image understanding with natural-language processing. The user asks a question about the image, and the AI attempts to identify the relevant visual information before generating an answer.
For example, a user could ask how many visible vehicles appear in an image, what color a particular object is, what text appears on a sign, or where an object is located relative to another object.
The accuracy of the answer depends on image quality, model capabilities, ambiguity, and the complexity of the question.
What Is Multimodal AI?
Multimodal AI refers to systems that can work with more than one type of information, such as text and images. A multimodal model can receive an image and a written instruction together and use both forms of information to perform a task.
This is important because human communication frequently combines visual and textual information. A person might upload a chart and ask for an explanation, provide a photograph and ask what it shows, or submit a document image and ask for specific information.
Multimodal systems make these interactions more natural because users can communicate about visual information using ordinary language.
How Does AI Analyze a Chart?
AI can analyze charts by identifying labels, axes, legends, plotted values, visual patterns, and relationships among data points when these elements are sufficiently clear.
A user might ask an AI to explain the trend shown in a line chart, identify the largest category in a bar chart, or summarize the information displayed in a dashboard.
Charts require particular caution because visual interpretation can sometimes produce mistakes. Exact numerical analysis should preferably be performed using the underlying data rather than relying only on a picture of a chart.
Can AI Analyze a Photograph?
AI can analyze photographs by identifying visible objects, scenes, colors, text, composition, and other features. For example, an AI system might describe a landscape, identify common objects in a room, or explain the general scene represented in a photograph.
However, the system should distinguish between what is directly visible and what is merely inferred. A photograph may show an object without providing enough information to determine its exact model, history, ownership, location, or purpose.
Can AI Identify Objects in an Image?
Object identification is one of the most established computer-vision tasks. AI models can be trained to recognize categories of objects and, in more advanced applications, distinguish between multiple objects in the same scene.
Accuracy varies according to the model, training data, image quality, object visibility, and similarity between possible categories. Objects that are partially hidden or photographed from unusual angles may be more difficult to identify.
Can AI Identify People in an Image?
AI systems can detect or analyze human figures in images, but identifying a specific person is a different and more sensitive task from simply detecting that a person is present.
A system may be able to describe visible characteristics or determine that a photograph contains people without being able to establish their identity reliably. Users should also consider privacy, consent, and applicable laws when using facial-recognition or identity-related technologies.
Can AI Analyze Medical Images?
AI is used in medical imaging research and in certain clinical technologies to assist with the analysis of images. Medical AI research can involve X-rays, CT scans, MRI images, pathology images, ultrasound images, and other forms of visual medical data.
These applications can involve detecting patterns, segmenting structures, classifying findings, or assisting qualified professionals with image interpretation.
Medical image analysis is a high-stakes application. A general-purpose AI assistant should not be treated as a substitute for a qualified medical professional or validated clinical system. Important medical decisions require appropriate clinical evaluation.
How Can AI Analyze a Document Image?
AI can combine OCR and visual understanding to analyze document images. It may extract text, recognize headings, identify tables, locate fields, and summarize the information contained in the document.
This is useful for receipts, invoices, forms, scanned records, certificates, printed notes, and other documents that are stored as images rather than editable text.
Document quality matters. Skewed pages, shadows, handwriting, low resolution, damaged paper, and unusual layouts can reduce the accuracy of extraction.
How Can AI Analyze an Image for Accessibility?
AI can assist accessibility by generating descriptions of images for people who cannot easily see them. A system may identify the main subject, setting, visible actions, and important text.
Alternative text should generally be concise and relevant to the purpose of the image. Decorative images may require little or no descriptive information, while an informational diagram may need a more detailed explanation.
Human review is valuable for important accessibility content because AI may omit information that is meaningful in a particular context.
How Can AI Analyze an Image for Education?
Students can use AI vision systems to understand diagrams, charts, maps, textbook illustrations, handwritten notes, and other visual learning materials.
For example, a student could upload a diagram and ask the AI to explain each labeled component in simple language. A chart could be converted into a written explanation, or a photographed page could be transformed into study notes.
The student should still compare the explanation with the original educational material, particularly when preparing for examinations or learning technical subjects.
How Can AI Analyze an Image for Business?
Businesses can use computer vision for inventory management, quality inspection, document processing, retail analysis, product classification, logistics, and other operational tasks.
For example, a retail system may analyze product images to organize a catalog, while a manufacturing system may inspect products for visible defects.
Business applications should be tested against representative real-world data because performance measured in controlled examples may not match performance in changing environments.
How Can AI Analyze Product Images?
AI can analyze product images to identify product categories, visible attributes, colors, shapes, labels, and other characteristics. E-commerce systems can use these capabilities to organize catalogs, improve search, generate descriptions, or identify similar products.
A product-image system may also detect whether an image contains a product against a background, identify multiple items, or extract visible text such as model numbers.
Exact product identification should be verified when small visual differences between models are important.
How Can AI Analyze an Image for Search?
AI can help search systems understand image content so that users can find images based on visual concepts rather than only filenames or manually entered tags.
For example, an image library may automatically identify categories such as vehicles, buildings, nature, food, or people. More advanced systems can connect visual concepts with natural-language search queries.
This approach is useful for large photo libraries where manually labeling every image would require significant time.
How Can AI Compare Two Images?
AI can compare images by examining visual similarities and differences. Depending on the application, comparison may focus on objects, layout, color, text, structural changes, or other visual features.
Image comparison is useful for detecting changes between versions of a design, inspecting products, comparing documents, checking photographs, and identifying differences in visual records.
A meaningful comparison requires a clear definition of what counts as a relevant difference. Two images can look different while representing the same object or information.
How Does AI Detect Objects in Complex Scenes?
Complex scenes can contain many overlapping objects, different lighting conditions, partial occlusion, reflections, and background clutter. Modern computer-vision systems are designed to learn patterns that help distinguish relevant objects from surrounding visual information.
Object detection models may produce confidence estimates or scores associated with detected categories. These values should not automatically be interpreted as guarantees of correctness.
Applications that require reliable detection should be tested with realistic examples and monitored for errors.
Why Does Image Quality Matter?
Image quality strongly affects AI image analysis. High-resolution images generally provide more visual information than severely compressed or blurry images. Good lighting can make objects and text easier to distinguish.
Problems such as motion blur, glare, shadows, extreme cropping, low contrast, distortion, and very small objects can make analysis harder.
If an AI system produces an uncertain result, providing a clearer image or a closer crop of the relevant area may improve the analysis.
Can AI Analyze Blurry Images?
AI may be able to interpret some information from blurry images, but accuracy decreases when important visual details are unavailable. A system may infer what an object probably is based on surrounding context, but inference is not the same as reliable observation.
Users should avoid treating an uncertain interpretation of a blurry image as definitive evidence. If accuracy matters, obtaining a clearer original image is preferable.
How Can You Write a Good AI Image Analysis Prompt?
A good image-analysis prompt clearly explains what the user wants to know. Instead of saying “analyze this image,” the user can specify the task.
For example, a prompt could ask: “Describe the main objects in this image, identify any visible text, explain the scene in simple language, and distinguish clearly between what is visible and what you are inferring.”
For a chart, the prompt might request the main trend, highest and lowest visible values, labels, and important comparisons. For a document, it might request text extraction, headings, dates, and key points.
Example AI Image Analysis Prompts
A general prompt can ask: “Describe what is visible in this image and organize the answer into subjects, objects, text, setting, and notable details.”
For text extraction, a useful prompt is: “Read the visible text in this image and reproduce it as accurately as possible. If any section is unclear, mark it as uncertain rather than guessing.”
For education, a user could ask: “Explain this diagram step by step for a beginner and define each important label.”
For a chart, a user could ask: “Explain the main trends shown in this chart and identify the largest visible differences. Do not invent values that cannot be read from the image.”
How Can AI Distinguish Facts From Inferences in an Image?
This distinction is extremely important. A visible fact is something that can be directly observed in the image. An inference is a conclusion drawn from what is visible.
For example, an image may visibly contain a person holding an object. It may not provide enough evidence to establish why the person is holding it, where they obtained it, or what they intend to do.
Users can ask AI to separate direct observations from interpretations. This makes the response more transparent and reduces the risk of treating assumptions as facts.
What Are AI Vision Models?
AI vision models are machine-learning systems designed to process visual information. Some are specialized for tasks such as classification, detection, segmentation, OCR, or image generation. Others are multimodal models designed to process images alongside text.
Modern AI systems can combine vision capabilities with language generation. This allows users to ask questions about images and receive natural-language responses rather than raw classification labels.
The capabilities of individual models vary significantly, so users should check what image formats, sizes, tasks, and features a particular system supports.
How Does AI Learn to Analyze Images?
AI models learn visual patterns during training. Training typically involves large collections of examples and optimization processes that adjust model parameters so that the system becomes better at performing specified tasks.
For image classification, the model may learn associations between visual patterns and categories. For object detection, it can learn both categories and locations. For image-text systems, training can establish relationships between visual content and language.
The quality and diversity of training data affect what the model can recognize and how well it generalizes to new images.
Why Can AI Make Mistakes When Analyzing Images?
AI can make mistakes for many reasons. The image may contain an unfamiliar object, unusual perspective, poor lighting, hidden details, ambiguous visual evidence, or patterns that differ from the examples used during training.
Another problem occurs when a user asks a question that cannot be answered reliably from the image alone. The AI may attempt to provide an answer even when the visual evidence is insufficient.
A good practice is to ask the AI to state uncertainty and identify what additional information would be needed to answer the question confidently.
Can AI Analyze Images Without Internet Access?
Some AI vision systems can operate locally on a device, while others rely on cloud-based processing. The available capabilities depend on the software, hardware, and model.
Local image analysis can offer advantages for privacy and offline operation, while cloud systems may provide access to larger or more capable models. Organizations should evaluate security, performance, cost, and data-handling requirements when choosing between these approaches.
What Is OCR and How Does It Work With AI?
Optical character recognition converts visual representations of written characters into text. OCR is especially useful when information exists only in an image or scanned document.
AI-based OCR can use learned visual patterns to recognize characters under different fonts, layouts, and conditions. More advanced systems can combine OCR with document understanding to identify the role of the text, such as a title, date, address, table entry, or product number.
How Can AI Analyze Handwritten Text?
Some AI systems can recognize handwritten text, although handwriting can be substantially more difficult than printed text. Differences in writing style, connected characters, unusual abbreviations, image quality, and page structure can affect recognition accuracy.
Users should verify extracted handwritten information when exact transcription matters.
Can AI Analyze Images in Real Time?
AI can analyze visual input in near real time when appropriate hardware, software, and model infrastructure are available. Real-time computer vision is used in areas such as robotics, industrial inspection, interactive applications, and certain monitoring systems.
Real-time systems must balance accuracy, processing speed, memory use, and energy consumption. A highly accurate model may require more computational resources than a lightweight model designed for fast responses.
How Is AI Image Analysis Used in Robotics?
Robots can use computer vision to identify objects, understand their surroundings, estimate positions, navigate environments, and interact with physical objects.
For example, a warehouse robot may use cameras to detect packages and determine where they are located. A robotic system may combine camera information with other sensors to build a more complete understanding of its environment.
Robotic vision is particularly challenging because the environment can change continuously and errors can have physical consequences.
How Is AI Image Analysis Used in Manufacturing?
Manufacturers can use computer vision for automated quality inspection. Cameras can capture products as they move through a production process, while AI models analyze them for visible defects or inconsistencies.
This can help identify problems more consistently and at high speed. However, production systems need careful validation because false positives and false negatives can both have operational costs.
How Is AI Image Analysis Used in Retail?
Retail applications can include product recognition, shelf analysis, inventory monitoring, visual search, catalog organization, and automated product classification.
For example, AI can analyze product images to generate structured attributes that improve search and filtering. Computer vision can also help identify whether shelves contain expected products, although real-world lighting and product placement can affect performance.
How Is AI Image Analysis Used in Science?
Scientists can use image analysis to process large collections of visual observations. Applications can include microscopy, astronomy, environmental monitoring, materials research, biological imaging, and other fields.
AI can help identify patterns that would be difficult or time-consuming to locate manually across thousands or millions of images. Researchers still need appropriate experimental design, validation, and domain expertise to interpret results.
How Is AI Image Analysis Used in Agriculture?
Computer vision can analyze crops, leaves, fields, fruits, soil surfaces, and other agricultural images. Potential applications include crop monitoring, plant classification, disease research, weed detection, yield estimation, and automated inspection.
Environmental conditions can vary significantly between locations and seasons, so models need to be tested under realistic conditions rather than assuming that performance in one setting will automatically transfer to another.
How Can AI Analyze Images for Content Moderation?
AI vision systems can classify images according to predefined categories and assist platforms with large-scale content moderation. Automated systems can identify potentially problematic visual material for further review.
Because visual context can be ambiguous, automated classification may produce errors. Human review can be important for difficult or sensitive cases.
What Are the Privacy Issues With AI Image Analysis?
Images can contain personal information, faces, locations, documents, addresses, identification numbers, and other sensitive details. Uploading an image to an AI service may therefore involve privacy considerations.
Users should understand how a service processes uploaded images and avoid sharing sensitive information unnecessarily. Organizations should establish appropriate policies for handling personal and confidential visual data.
Can AI Analyze an Image Without Knowing Its Original Context?
AI can analyze visible information without knowing the complete history of an image, but this limitation matters. A photograph may not reveal when it was taken, where it was taken, who created it, or what happened immediately before or after the captured moment.
Visual evidence should therefore be separated from contextual claims. An AI can describe what appears in an image while remaining uncertain about events that are not directly represented.
How Can AI Analyze Multiple Images?
When a system supports multiple image inputs, AI can compare or synthesize information across several images. This can be useful for before-and-after comparisons, document collections, product variations, multiple photographs of the same object, or visual research.
The user should specify how the images should be compared. For example, the task may involve identifying changes, common elements, unique elements, or differences in text.
Can AI Analyze an Image and Text Together?
Yes. Multimodal AI can combine visual and textual input. A user can upload an image and provide a written question or instructions describing the desired analysis.
This makes image analysis more flexible because the same image can support different tasks. A photograph can be described, a diagram can be explained, and a document can be summarized depending on the user's request.
How Can AI Analyze an Image More Accurately?
Start with the highest-quality image available. Make sure important areas are visible and not unnecessarily cropped. If the image contains small text, provide a clear version or a close-up where appropriate.
Next, write a specific prompt. Explain exactly what you want the AI to identify, extract, compare, or explain. If accuracy is important, ask the AI to distinguish visible evidence from inference and to identify uncertain areas.
For numerical or technical information, verify the result against the original data whenever possible.
What Should You Do If AI Cannot Understand an Image?
If an AI system cannot interpret an image correctly, try improving the input. A higher-resolution version, better lighting, a different crop, or a close-up of the relevant section may help.
You can also simplify the task. Instead of asking the AI to analyze everything, ask one specific question at a time. For example, first request text extraction, then ask about a particular section.
If the visual information remains ambiguous, do not force a definitive conclusion. Some images simply do not contain enough information to answer a question reliably.
Best Practices for AI Image Analysis
Use clear, high-quality images whenever possible. Give the AI a specific task instead of a vague request. State whether you want a description, text extraction, object identification, comparison, explanation, or another form of analysis.
Ask the system to preserve uncertainty rather than guessing. Verify important results against the original image or another reliable source. Avoid uploading sensitive images unless the service's privacy and security practices are appropriate for the information.
For professional applications, test the system using representative examples and measure both correct results and errors. A model that performs well on simple images may behave differently under real-world conditions.
What Are the Main Benefits of AI Image Analysis?
AI image analysis can process visual information quickly, automate repetitive tasks, organize large image collections, extract text, support accessibility, assist research, and help users understand visual information.
It can also make visual data searchable and easier to combine with other information. Multimodal AI allows users to ask natural-language questions instead of learning specialized computer-vision interfaces.
For organizations with large image collections, automation can reduce the manual effort required to classify, label, inspect, or organize visual information.
What Are the Limitations of AI Image Analysis?
The most important limitation is that AI does not have perfect visual understanding. It can misclassify objects, misread text, overlook small details, or make unsupported inferences.
Image quality, unusual viewpoints, occlusion, complex scenes, ambiguous context, and unfamiliar objects can all affect performance.
There are also privacy and ethical considerations when images contain people or sensitive information. High-stakes applications require stronger validation than casual image description.
AI Image Analysis vs Human Visual Analysis
AI and humans have different strengths. AI can process large numbers of images quickly and consistently for narrowly defined tasks. Humans can use broader contextual knowledge, common sense, domain expertise, and judgment when evidence is ambiguous.
In many practical applications, the two approaches can complement each other. AI can perform initial analysis or screening, while a person reviews important or uncertain cases.
The appropriate balance depends on the task, consequences of errors, required speed, and available resources.
How AI Image Analysis Is Changing Digital Work
Image analysis is becoming an important component of modern digital workflows. Businesses can process documents automatically, developers can build applications that understand photographs, educators can create visual learning tools, and researchers can analyze large image collections.
The combination of computer vision and natural-language interaction is especially significant because it allows users to interact with visual information using ordinary questions and instructions.
As multimodal AI develops, image analysis is increasingly becoming a general-purpose capability rather than a specialized feature available only through dedicated computer-vision software.
Final Answer: How Can AI Analyze an Image?
AI can analyze an image by converting visual information into representations that machine-learning models can process and then using those representations to identify objects, text, patterns, scenes, relationships, and other features. Computer vision provides many of the underlying techniques, while multimodal AI can combine visual analysis with natural-language reasoning.
Depending on the system, AI can classify images, detect objects, segment regions, read text with OCR, describe photographs, interpret diagrams, analyze charts, compare images, summarize document images, and answer questions about visual content.
The quality of the result depends on the AI model, training data, image quality, task definition, and amount of available context. Clear images and specific prompts generally make analysis easier. Users should also distinguish between what is directly visible and what the AI is inferring.
AI image analysis is useful across education, research, business, manufacturing, retail, accessibility, robotics, agriculture, science, and many other fields. At the same time, it can make mistakes and should not automatically be treated as authoritative evidence, particularly for medical, legal, financial, safety, identity, or other high-impact decisions.
The most reliable approach is to use AI as an image-analysis assistant: provide a clear image, describe the exact task, request uncertainty where appropriate, review the response, and verify important findings against the original image or another reliable source. Used this way, AI can make visual information much easier to understand, search, organize, and use.
