GFoundry Intelligence - Overview
GFoundry Intelligence is the AI layer of the GFoundry platform: Gi's agents, which write, teach, simulate and answer, and People Intelligence, which computes Insights about your organisation.
What it is
GFoundry Intelligence has two halves, and they work in different ways.
Gi's agents are generative. They write, teach, simulate and answer: an assistant for employees, a conversational analyst for the backoffice, a course generator, a role-play coach and a writing assistant inside backoffice forms.
People Intelligence is analytical. It is a catalogue of Insights, authored by GFoundry, that you assemble into your own dashboard, plus a set of Insight cards on an individual employee's profile. These are computed by the platform, not generated by a model.
Most organisations use both. Knowing which one you are talking about matters, because they answer to different rules on privacy, on repeatability and on how much weight the output can carry.
Gi's agents
Gi: AI Assistant for Employees and Managers. On-demand answers from your company documents and platform content.
Data Analyst Agent. Ask questions in natural language in the backoffice and get answers on talent, performance, learning, recognition and activity, including behavioural archetypes where your organisation has them. It answers what you ask.
Role-Play Coach Agent. Builds role-play scenarios where learners practise a conversation with an AI persona, and evaluates each session against the scenario's criteria.
Learning Designer Agent. Turns a briefing or an uploaded PDF into a course with a structured content page and, as you choose, slides, a quiz, a downloadable PDF and a curated YouTube video.
AI Writing Assistant. In the text fields of backoffice forms: generate, improve, shorten, fix spelling and grammar, apply your own instruction, or translate into your configured languages.
People Intelligence
A catalogue of ready-made Insights rather than a fixed dashboard or a blank report builder. Every Insight carries a metric contract, an honesty classification and privacy protection applied on the server, and states what it does not show as clearly as what it does.
People Intelligence: the Insight dashboard for your organisation. Start here: what an Insight is, what each one carries, and why your catalogue may be smaller than another organisation's.
How to read a People Intelligence Insight. The badges, the freshness stamp, the filters, and what a suppressed or unavailable card means.
Building your People Intelligence dashboard. Adding Insights, filtering a card, comparing segments, refreshing and removing.
The People Intelligence Insight catalogue. The reference list: every organisation-level Insight, what it answers, and what it refuses to claim.
Talent Insights: People Intelligence on an employee's profile. The cards in the Talent Insights tab of an employee's profile, and the rules for using them responsibly.
People Intelligence: privacy, limits and responsible use. What is protected, what is deliberately withheld, and the questions a works council or DPO will ask.
How to judge what you are looking at
Every Insight carries one of four labels, and the label is a statement about how much weight the Insight can hold.
Factual: a count, a rate or a status read from a record someone created.
Diagnostic: a composite or derived reading. It points you at where to look and is never sufficient on its own for a decision about an individual.
Predictive: an estimate of something that has not happened yet. Only admitted to the catalogue with validation information attached.
Experimental: offered because it is useful, still under validation, and never the sole basis for a decision about a named person.
Who is it for?
Employees: on-demand answers from company knowledge, and role-play practice.
Managers: practice and answers from company knowledge.
HR and people analytics teams: faster content creation, natural-language analytics, and a defensible read of the organisation with the caveats attached rather than discovered later.
HR directors and CHROs: a dashboard of Insights assembled for their questions.
Requirements
An active GFoundry account with the Intelligence features enabled for your organisation.
Backoffice permissions for the features you want to use. For example, Ask AI (the Data Analyst Agent) is shown to administrators and to users given access to the assistant.
For People Intelligence, the modules that feed each Insight. An Insight that reads pulse surveys, recognition, learning or an evaluation cycle is only offered once that source exists, which is why two organisations see catalogues of different sizes.
Data and privacy
The two halves handle data differently, and the distinction is worth having ready when the question comes up.
Gi's agents send prompts to external AI providers (OpenAI and Google Gemini). What each agent sends is different. The Data Analyst Agent replaces employee names with pseudonymous identifiers before data reaches the model. Gi, the Learning Designer Agent and the Role-Play Coach Agent send the text they work with as written: the user's question, the relevant document excerpts, your course briefing and document, or the role-play conversation. Each agent's article details what it sends.
People Intelligence Insights are produced by deterministic code running inside the platform, so the same inputs produce the same card. Computing an Insight does not call an AI provider.
No agent mixes one organisation's data with another's. AI output, including the Role-Play Coach Agent's evaluations, is there to inform people, and decisions about an employee stay with the people who own them. For the full picture, see our Data Protection and Privacy Policy. The Data Processing Agreement, including the list of sub-processors, is available on request.