Your learning profile
Artificial intelligence
Using AI well at work: how models work, prompting, workflows and agents, applied AI in finance, accounting, design, content, marketing and data, responsible AI and security (added 3 Nov 2026 with the AI at work courses). 235 terms, page 2 of 2.
- Synthetic mediaImages, video or audio that have been generated or substantially changed by AI, especially when they look or sound realistic.
- Synthetic personaAn AI-generated imitation of a type of customer that can help prepare research questions and ideas but is never a substitute for research with real people.
- Synthetic voiceSpeech generated by software rather than recorded, sometimes modelled on a real person's voice, which should only happen with that person's clear consent.
- System promptA system prompt is a set of standing instructions placed before a conversation starts, which shape every later response, such as rules on tone, format or what…
- Task decompositionTask decomposition means breaking a complex request into smaller steps that are handled one at a time, so each step is simpler to do and easier to check.
- Task exposureTask exposure describes how much a particular work task could be sped up, changed or taken over by AI, judged one task at a time rather than for a whole job ti…
- Task fitTask fit is how well a particular task suits an AI assistant, judged mainly by how much is at stake and how easily a person can check the output.
- Task inventoryA task inventory is a list of the repeated tasks a team performs, recording how often each occurs and how long it takes, so that time can be seen and automatio…
- TemperatureTemperature is a sampling setting that controls how varied a model's output is: lower values give more predictable wording, higher values give more varied and…
- Test setData kept aside and never used to build or tune a model, so it gives an honest estimate of how the model performs on cases it has not seen.
- Text and data miningText and data mining is the automated analysis of large amounts of text, images or data to find patterns, including collecting published work to train AI model…
- Text-to-image modelAn AI model that generates a new image from a written description, starting from random noise rather than from an existing picture.
- Text-to-SQLTurning a question written in plain language into a database query, often with the help of an AI assistant, so the question can be run against stored data.
- Theme codingGrouping open-ended comments, such as survey answers or reviews, into labelled themes using clear definitions, then counting how many comments fall into each t…
- Third-party riskThird-party risk is the risk that comes from relying on an outside supplier, such as an AI model provider, whose failure, change or weak controls could affect…
- Threat modelA threat model is a structured description of what could go wrong with a system, who might cause it, how, and what defences are in place.
- Time auditA time audit is recording how working hours are actually spent over a typical period, task by task, so that the tasks taking the most time can be found and imp…
- TimeoutA timeout is the longest time a workflow waits for a step or an outside system to answer before treating the attempt as failed and following its retry or error…
- TokenA token is the small unit of text a language model reads and writes, such as a short word, part of a longer word, a number or a punctuation mark.
- Tool useTool use means an AI model calling outside functions, such as a search, a calculator or a database lookup, to get information or take actions it cannot manage…
- Training dataThe examples an AI or machine-learning model learns from; whatever is missing, outdated or biased in the training data tends to show up in the model's answers.
- TriggerA trigger is the event that starts an automated workflow, such as a new email arriving, a form being submitted, a file appearing in a folder or a set time of d…
- Two-person ruleThe two-person rule is a control that requires two authorised people to approve a sensitive action, such as a large payment or a change of bank details.
- Unsupervised learningMachine learning without known answers, used to find structure in data such as groups of similar items, hidden factors or unusual points.
- UpscalingIncreasing the pixel resolution of an image, with AI adding plausible detail that it invents rather than recovering detail that was really there.
- Use case inventoryA use case inventory is a register of every place an organisation uses AI, recording the owner, purpose, data, risk level and controls for each use.
- User testingWatching real people try to complete tasks with a design, in order to find where it confuses, slows or fails them.
- Vector databaseA vector database stores embeddings alongside their text and quickly finds the stored items closest in meaning to a new question or passage.
- Vendor lock-inVendor lock-in is the situation where leaving a supplier or tool becomes so costly or difficult that a business stays even when it no longer suits.
- Voice guideA short document describing how a writer or brand sounds, with examples of phrasing to use and to avoid, so drafts by anyone stay consistent.
- Walk-forward validationTesting a time-series model by training on an earlier period, testing on the next one, then rolling both windows forward, so the model is never tested on data…
- WebhookA webhook is a message that one system sends automatically to another's web address the moment an event happens, so a workflow can start at once instead of che…
- WireframeA simple outline of a page or screen that shows layout, content and order without colour, imagery or final visual design.
- Workflow mappingWorkflow mapping is writing down each step of a recurring task, who does it, what it needs and how long it takes, before deciding what to change or automate.
- Zero-shot promptingZero-shot prompting means asking a model to carry out a task using instructions only, without including any worked examples of the inputs and outputs you want.
Other subjects
- Finance
- Markets
- Investing
- Business
- Accounting
- Technology
- Fintech
- Regulation
- Financial law
- Crypto
- Cybersecurity
- Political economy
- Financial ethics
- Economics
- Personal money
- Marketing
- Media
- Advertising
- Public relations
- Brand
- Audit and assurance
- Risk management
- Strategy and management
- Project management
- Consulting
- Human resources
Education, not advice: the Stockertim Dictionary explains what terms mean. It never recommends buying or selling anything.