AI, in plain language.

AI agent

Software that does a task from start to finish, not just answers a question. You give it a goal in plain language and it carries out the steps, where a chatbot would only reply. AI for small business.

Agentic AI

AI that acts instead of only answering: it plans the steps toward a goal, carries them out, and adjusts as it goes. A business runs on agentic AI when its tools do the work rather than draft it for someone else to finish. Intelligent multi-agent AI.

AI employee

An AI agent, or a team of them, that you brief like a colleague and that does ongoing real work for your business rather than one-off answers. What is an AI employee.

AI team (agent fleet)

Several specialised agents working together, each on its part of the work, coordinating and reviewing each other instead of one model trying to do everything. How the platform works.

AI workforce

Your business's ongoing AI capacity: agents that cover real roles across marketing, sales, support, and operations the way a team of employees would, without the headcount. Also called an AI team or AI employees. AI for small business.

AI sales agent

An AI agent that runs sales outreach from end to end: it builds the prospect list, writes each email in your voice, follows up without dropping anyone, and books the meeting when someone replies, instead of just drafting a template for you to send. What is an AI sales agent.

AI onboarding assistant

An AI agent that already knows your company and guides a new hire from day one: it answers their questions in chat or Slack, points them at the right documents and people, and walks them through their first weeks, instead of leaving them a folder of PDFs. What is an AI onboarding assistant.

AI support agent

An AI agent that handles customer support on your product: it reads the page the customer is on, answers from your own help docs, fills in forms through signed actions, and files a structured bug report when something is broken, instead of deflecting to an article like a chatbot. What is an AI support agent.

Agent memory

Context that persists and builds up over time, so the AI remembers your business across sessions instead of starting from zero every time. Why agents need memory.

Multi-agent workflow

A way of getting work done where agents hand tasks to each other, check each other's output, and escalate to a human only when judgement is needed. Multi-agent workflows explained.

Autonomy

An agent working unattended, finding and finishing work on its own schedule, while you stay in control of anything that ships. Memory, coordination, autonomy.

Human in the loop

A way of working where the AI does the work but a person approves anything that goes out, so you keep control of what ships. The agent proposes and prepares; you review and sign off. How the platform works.

Plain-language briefing

Telling an AI what you want in ordinary words, the way you would brief a colleague, instead of writing careful prompts. It means no prompt engineering: you describe the outcome and the agent works out the steps. AI for small business.

Model-agnostic AI

A setup that is not tied to a single AI model or vendor, so the system keeps working and can use the best model as that changes over time. Model-agnostic AI.

Brand-consistent AI

AI whose output stays on-brand every time because every agent reads from the same shared memory of your brand and your data. Brand-consistent AI.

AI for small business

Using AI agents to get a team's worth of work done across marketing, sales, support, and operations, without hiring the headcount. AI for small business.

EU-resident AI

AI where the data and the processing stay inside the EU, rather than being sent wherever a provider happens to run. EU and sovereign AI.

Cookieless analytics

Web analytics that measures visits without cookies or cross-site tracking, so there is no consent banner and nothing following people around the web. You see real traffic while respecting privacy. AI web analytics.

Ticket deflection

The share of support questions handled before they reach your team. The honest way to raise it is resolution: an agent that answers the question end to end, so the ticket never lands on a human and does not come back as a reopen. Deflection tools that just point at a help article drop the count without answering anyone. AI customer support.

AI SDR

An AI sales development rep: it builds the prospect list by role, industry, and fit, writes each outreach email for the person receiving it, follows up on schedule, stops the moment someone replies, and books the meeting. The work of an SDR, run by an agent with your approval level set per action. AI sales agents.

Codebase map

An interactive knowledge graph built from a repository: files, classes, and functions as nodes, real imports and calls as edges, clustered into labeled modules with guided tours through the important paths. Documentation extracted from the code itself, so it cannot drift from what the code does. A system map from existing code.

Company brain

Your business knowledge made answerable where your team already works: an agent connected to your own docs, decisions, and history that answers questions in Slack instead of making people dig through folders or interrupt a colleague. AI assistant on Slack.

Autonomy dial

Per-action control over how much an agent does alone: draft-only, queue for your approval, or run end to end. You raise it per job as trust builds, and anything that spends money always waits for a human yes, whatever the dial says. How to build an AI team.

GATE (gate.software)

The platform this glossary belongs to: an AI team for small and mid-sized businesses, built by Gate Intelligence AB in Stockholm. Not affiliated with the crypto exchange, the large-model management product, or any other company that happens to share the name. If you searched for one of those and landed here, welcome anyway. About GATE.

These are not ideas. They run real businesses.

Every term above is something GATE does in production today. See the workloads it runs, or request an invite.