REF17 terms
Glossary
Plain-English definitions of the terms that come up in websites, marketing, software and AI projects, so a conversation with us, or anyone else, starts from the same words.
REF.1
Terms
- Technology roadmap
A step-by-step plan for what to build, buy or automate in your business, and in what order.
A technology roadmap turns a vague sense that things could work better into a short list of practical steps. Each step says what problem it solves, whether to build something, buy a product or automate a task, roughly what it takes, and what comes next. Good roadmaps start small, so every step is useful on its own.
- AI automation
Using AI to do routine work, such as sorting emails or drafting replies, with a person checking what matters.
AI automation joins an AI model to the tools you already use, so routine tasks such as reading enquiries, filling in forms, chasing invoices or drafting replies happen on their own. Good automation is narrow, logged and has a person approving anything with real consequences.
- API (application programming interface)
A way for two pieces of software to talk to each other automatically.
An API is a set of rules that lets one system ask another for information or ask it to do something, without a person copying data between screens. Most modern business software, from accounting to booking tools, has one.
- SEO (search engine optimisation)
Work that helps your website show up when people search on Google and other search engines.
SEO covers the technical health of a website (speed, mobile use, structure), the content on it (clear answers to the questions customers ask) and signals from elsewhere (consistent business details, reviews and links). Nobody can honestly guarantee a ranking.
- GEO (generative engine optimisation)
Helping AI assistants such as ChatGPT, Gemini and Perplexity understand your business and mention it accurately.
AI assistants answer questions by reading and summarising the web. GEO makes sure the facts about your business are clear, consistent and easy for them to find and cite: well-structured pages, structured data, plain answers and, where useful, an llms.txt file. It overlaps heavily with good SEO.
- MCP (Model Context Protocol)
An open standard for connecting AI assistants to tools and data.
The Model Context Protocol is an open protocol, built on JSON-RPC 2.0, that defines how AI applications (hosts and clients) talk to services (servers) that expose tools, resources and prompts. Anthropic introduced it in November 2024, and it is now governed under the Linux Foundation's Agentic AI Foundation.
- MCP server
A service that exposes tools and data to AI assistants over MCP.
An MCP server wraps a system, such as a database, an API or a file store, and publishes a set of tools an AI client can discover and call. A good MCP server exposes only what is needed, checks permissions on every call and logs what happened.
- AI agent
An AI system that takes several steps on its own to reach a goal.
An agent combines a language model with tools and a loop: it decides what to do next, calls a tool, reads the result and continues until the task is done or it needs a person. Agents are only as safe as the permissions they are given.
- LLM (large language model)
A model trained on large amounts of text to predict and generate language.
Large language models power chat assistants and many AI features. They are good at drafting, summarising and classifying text, and can be wrong with confidence, which is why grounding and human review matter.
- RAG (retrieval-augmented generation)
Answering questions by first retrieving relevant passages from your own documents.
Instead of relying on what a model learned in training, a RAG system searches your documents for relevant passages and asks the model to answer from them, with citations. It is the usual basis of a knowledge assistant.
- Human in the loop
A design where a person reviews or approves AI output before it takes effect.
Human-in-the-loop design puts a person at the decision points that matter: sending a customer email, making a payment, changing a record. It reduces risk and supports the safeguards UK GDPR requires for significant automated decisions.
- DPIA (data protection impact assessment)
A documented risk assessment for processing personal data.
Under UK GDPR, a DPIA is required for processing that is likely to result in a high risk to individuals. Many AI projects that use personal data need one. The ICO publishes guidance and a template.
- Automated decision-making (ADM)
Decisions made by technology without meaningful human involvement.
UK GDPR regulates solely automated decisions with legal or similarly significant effects. The Data (Use and Access) Act 2025 replaced Article 22 with Articles 22A to 22D, which took effect on 5 February 2026, widening the lawful bases while keeping safeguards such as the right to human intervention.
- MVP (minimum viable product)
The smallest version of a product real users can use and learn from.
An MVP tests whether a product is worth building further. It should be small in scope but not careless in quality.
- SaaS (software as a service)
Software customers access online and pay for by subscription.
SaaS products are hosted by the provider and used through a browser or app, usually paid monthly or annually per user or per account.
- CI/CD
Continuous integration and continuous delivery: automated testing and deployment.
CI runs automated checks on every code change; CD automates getting approved changes to production. Together they make releases routine and reversible.
- Day rate
The price for one person's working day on a project.
Software work is often estimated in person-days and priced by day rate. Fixed quotes are usually built from an estimate of days plus a contingency.
T+99Next step
Talk it through with us
A free 30-minute first chat, with no obligation. Tell us what is getting in the way; we tell you honestly what we would do, roughly what it would take, and whether it is worth doing at all.