AI agents & agentic AI
Meet the agent inside your academy
Most people have met a chatbot. An agent is different: it takes steps — works out what it needs, opens the right report in your own platform, reads what is actually there, and only then answers. Here is what that means in plain words, and what it changes for the people who run your academy.
In plain words
What is an AI agent?
No jargon, and nothing you have to be technical to follow. Two short answers, then an example of the real thing.
A chatbot talks. An agent takes steps.
A plain chatbot answers from whatever it already knows, in one go. An agent works more like a colleague with the keys to the filing room: it works out what it needs, opens the right report, reads the rows that are actually in there, and only then writes an answer. The figures come from your platform, not from a model's memory.
“Agentic AI” is AI that acts within limits you set.
Agentic simply means the AI can act — use a tool, run a lookup, go back for one more piece of information — instead of only producing text. The limits matter as much as the ability: which tools it may use, whose data it may see, what it may do with them. In EduGears AI LMS those limits are the design. The agent can read, and reading is all it can do.
Example
What that looks like
How is enrolment across our courses this month?
Across your 18 published courses, 214 learners enrolled this month — 27% more than last month. Business English leads with 46. Three courses took no new enrolments at all: Food Safety Level 2, Workplace First Aid and Retail Operations. Shall I show you how those three did last term?
Where this came from
Enrolment report · 18 coursesCog opened the same enrolment report the owner can open, counted the rows, and linked it. Follow the source and you land on that report with the filters already applied — so the figure in the answer and the figure on the page are the same figure.
Across the platform
How Cog works in your academy
Cog is the assistant surface your staff and students already use — one panel reachable from every page, now able to look things up as well as explain them. It answers in the language the reader is using, under your own branding.
One panel, every page
Staff open Cog from the corner of any page in the platform. It already knows the course, class or report on screen, so a question rarely needs setting up, and the conversation carries across pages.
Ask in plain words
No report builder, no filter syntax. “Which classes are behind this term?”, “How did the last assessment go?”, “Who hasn't finished the induction course?” — asked the way one colleague asks another.
Answers from your own reports
Cog draws on the platform's own reports and dashboards: enrolment, course and module progress, gradebook and category results, certificates issued. The same figures your reporting pages show, gathered into a sentence.
And how-do-I answers too
The same panel walks staff through the product itself, from EduGears AI's own manual — how to publish a course, how to set up a grade category, how to issue a certificate — and names the help article it drew on.
Student information, where it is switched on
Academies running EduGears AI SIS can ask about student records as well — cohorts, attendance, report cards. Where the SIS is not switched on, those questions simply have no data behind them, and Cog says so rather than guessing.
Your administrators decide
Data questions are a capability the owner or an administrator switches on or off for the whole organization, in the platform's own settings — on for the office, off while you take a look at it, changed again at any time.
Student records, attendance and report cards live in EduGears AI SIS, built into the same platform rather than bolted onto it. Switch it on and the questions your office asks about students have data behind them; leave it off and everything else here works exactly as described.
Explore EduGears AI SISGuardrails
Guardrails, not promises
The limits are built into how the agent runs — not merely written into its instructions.
Read-only, always
Cog can look; it cannot change anything. No grade touched, no record edited, no enrolment altered, no message sent on anyone's behalf. Reading is the only thing the agent is able to do — every change stays a human action in the normal screens.
It sees only what the person asking can see
The agent works as the person in front of it: their role, their classes, their organization. An instructor asking about a class gets their class; an administrator gets the academy. If a report is not open to the person asking, it is not open to Cog while they are the one asking.
Every figure is traceable
Answers name the report each number came from, and the source opens that real report with the same filters — so anything that looks odd can be checked in one click. Where the data isn't there, Cog says so instead of filling the gap.
Switched on by your administrators
Data questions are turned on or off for the whole organization by an owner or administrator, at any time. Underneath, tenant isolation still applies: one academy's data is never visible to another.
Questions schools ask
The things every academy wants settled before its staff start asking an AI about their students.
What is an AI agent?
An AI agent is an AI that takes steps to answer you instead of replying from memory alone. It works out what it needs, uses a tool to look it up — in our case one of the reports your platform already produces — reads the result, and then writes an answer based on what it found.
What does “agentic AI” mean?
Agentic AI is AI that can act — use tools, run a lookup, go back for one more piece of information — within limits somebody sets. The limits matter as much as the ability. In EduGears AI LMS the agent's only tools are read-only lookups into reports the person asking can already open.
Can it change grades, enrolments or student records?
No. Cog is read-only by design and by construction: it has no tool that writes, edits, sends or deletes anything, and the connections it reads through are read-only as a backstop. Changing a grade, an enrolment or a record stays a human action in the normal screens.
Can an instructor see another class, or another academy's data?
No. The agent answers as the person asking, with their role and their permissions, inside their own tenant. Instructors get the classes they teach; administrators get their own organization. Tenant isolation is enforced in the database itself, so one academy's data is never visible to another.
Who decides whether staff can ask data questions?
The academy does. An owner or administrator switches data questions on or off for the whole organization in the platform's own settings, and what each person can then ask about is decided by their own role and permissions. The rest of Cog — coursework help for learners and product walkthroughs for staff — carries on either way.
Is our data used to train AI models?
No. Your courses, submissions and grades are not used to train AI models. Answers are assembled from your own reports inside your own tenant, and an academy that would rather run the AI on its own provider account can bring its own key.
See it answer a question about your own academy
Book a walkthrough and we will open the assistant panel on a real course, ask it something about enrolment or progress, and follow the source straight back to the report the figures came from.