Use AI on your own documents without handing them to public AI services. HLBNGA runs its own private large language model on infrastructure we own and operate. We set up your documents, connect them to our model, and deliver answers drawn from your data through a custom chatbot, our APIs or products like Compliance Logger.
| Component | Set-up | Status |
|---|---|---|
| Model | HLBNGA private LLM | ACTIVE |
| Document library | Policies & procedures | INDEXINGSYNCED |
| Answers drawn from | Your documents | ENFORCED |
| Public AI & web search | Not used | NONE |
Public AI tools such as ChatGPT are useful, but every prompt sends information to an outside provider, and the answers come from general internet knowledge rather than your own records. Ask one about your own policy and it may simply make one up.
HLBNGA's private LLM runs on infrastructure we own and operate. Your documents and questions are processed there, not by public AI services, and the model answers from your documents rather than the open web.
Our private LLM does not answer from general internet knowledge. When someone asks a question, the system first searches your document library for the most relevant passages, then asks the model to answer using those passages.
Because every answer is drawn from your documents, it is far less likely to be invented, and easy to check against your own policies. It is the same reason Compliance Logger references the policies you upload instead of guessing what they say.
We build chatbots that answer from your organisation's own documents. Your people stop searching shared drives and policy manuals: they ask a question in plain English and get a clear answer drawn from your own documents.
For one client, we loaded their complete document set into a private assistant that references their own files instead of searching the web. The same approach powers the policy checks in Compliance Logger, where answers reference the policies you upload rather than inventing them.
We handle the set-up from start to finish, so your team gets the benefit of AI on its own data without having to run any AI infrastructure.
We agree the use case, the documents involved and who will use the assistant, whether that is your staff, a chatbot or one of our products.
You receiveA clear scope and document list
We load your documents into a private library on our infrastructure, organise them and prepare them for fast, accurate search.
You receiveYour private document library
We set the assistant to answer from your documents, then test it with your real questions before anyone relies on it.
You receiveA tested, ready assistant
We launch it as a chatbot, through our API or inside an HLBNGA product, and keep your library current as your documents change.
You receiveA live assistant and ongoing updates
The model supports your people rather than replacing them. Every answer is drawn from your documents and gives a starting point that a person can check, use and act on.
A private assistant that answers questions from your own documents instead of searching the web.
In Compliance Logger, incidents and reports are checked against the policies you upload rather than a guessed policy.
Connect your own applications to our private LLM through HLBNGA's APIs, so your systems can use AI on your data without public AI services.
Long policies, contracts and reports condensed into clear summaries drawn from the document itself.
Quick answers about your own AML/CTF program and procedures, drawn from the program itself, so staff apply the rules you have actually written.
New starters get answers about internal processes from your own handbooks and procedures, not from the internet.
Good answers start with well-organised documents. We take your policies, procedures and reference material, organise them into a private library and prepare them so the model can find the right passage quickly.
When documents change, the library is updated, so answers reflect your current versions rather than an old copy someone saved years ago.
| Document | Version | Status |
|---|---|---|
| AML/CTF Program Part A | v7 | LIVE |
| ECDD Procedure | v3.2 | INDEXINGLIVE |
| ECDD Procedure | v3.1 | LIVEREPLACED |
| Whistleblower Policy | v2 | LIVE |
| Code of Conduct | v5 | LIVE |
| Staff Onboarding Handbook | v4 | LIVE |
Public AI chatbots are built to answer anything, using general knowledge from the internet. That makes them handy for everyday tasks, but risky for questions about your own business: they cannot see your documents unless you paste them in, and once you do, that information has left your organisation.
A private LLM connected to your documents works the other way round. It answers from your data and keeps your information away from public AI services.
| Public AI chatbot | HLBNGA private LLM | |
|---|---|---|
| Your data goes to a public AI provider | Yes | No |
| Answers drawn from | General internet knowledge | Your own documents |
| Knows your internal policies | Only if you paste them in | Yes |
| Set up and kept current for you | No | Yes |
| Risk of made-up answers about your policies | Higher | Reduced |
| Works inside your systems | A separate tool | Chatbot, APIs and products |
No. Your documents and questions are processed by HLBNGA's own private LLM, on infrastructure we own and operate. They are not sent to ChatGPT or any other public AI service, and the model does not search the open web for answers.
Public chatbots answer from general knowledge gathered from the internet. Our private LLM answers from your own documents, so you are not relying on a guess about your own policies.
No. Your documents are used to find answers to your organisation's questions. They are not shared with public AI providers or used to train public AI models.
In South Africa, on infrastructure HLBNGA owns and operates, where it is protected under South Africa's Protection of Personal Information Act (POPIA). It is never sent to public AI services in any country. If your organisation is covered by Australia's Privacy Act, talk to us about what this means for your obligations when personal information is processed overseas.
Yes. The same private LLM powers HLBNGA products such as Compliance Logger, and it is available through our APIs so your own applications can use it.
Your library is updated with the new versions and outdated ones are removed, so answers reflect the documents you currently use.
Talk to us about a private assistant built on your own documents, and see the difference when answers come from your data instead of the internet.