Referenzwert amberGPT

Few technologies have generated as much hype in recent years as ChatGPT from OpenAI.

We have also noticed this and know that many companies are asking themselves how they can also profit from GPT technology. But what many realise relatively quickly: ChatGPT has its limits and is only trained with public data until 2021.

But: this technology can also be used for internal company use cases. This is exactly the feature we have developed with amberAI – and this is what it looks like:

Use Cases

The most important question to ask first is what you can use amberAI for. Some use cases are already worked out – but there is so much more! amberAI is a new module of amberSearch.

Let’s start with some already known use cases:

1. Let amberSearch write an e-mail for you

Example sales email generated with the following prompt: “Write me a cold call email with 3 USP’s to sales managers of medium-sized companies”.

No matter if you have to answer an enquiry in customer service with your company’s know-how, write an invitation to your team or a sales email with your USP’s: amberSearch first finds the most relevant hits for the respective topic and converts them with amberAI into meaningful sentences that you can use afterwards.

2. Summarise projects and write project reports with amberAI

How nice would it be if you could just quickly summarise projects? This is now possible with amberSearch. The advantage of amberAI is that – if available – it does not only use information from one document, but interprets the know-how of different documents together. The more content there is and the more specific your prompt is, the better the output of amberAI can be.

What an example project summary might look like

3. Find the right contact person

The right contact person is just a request away. Based on the information we can extract from metadata and documents, we can tell you who deals with which topic in your organisation. Thus, amberSearch massively simplifies the search for experts in your company.

4. Find reference projects

Are you looking for a reference project for one of your customers or are you trying to find out how you solved a case a few years ago? With amberSearch this is very easy. We analyse the different documents and show them to you accordingly. This way you can focus on the eyes that really add value.

5. Find internal information on policies and guidelines

A big pain point for many employees is finding policies or the information in the policies. With amberSearch, staff can now find all existing policies and have them summarised as well.

How does amberAI work?

Instead of training AI models with our clients’ data, we use an AI model that has already been pre-trained for this use case. For our clients, this means we don’t need data to re-train AI models.

Furthermore, because the search process is live, we can take all access rights into account and users only find the information they really have access to.

In a first step, our software amberSearch – taking into account access rights – finds the most relevant documents for the respective query. It does this from the user’s point of view with the information from all connected data sources (i.e. not only documents but also e-mails, team messages, intranet, …). Only then is a selection of the most suitable hits passed on to the text-generating AI model and the answer generated. Afterwards, amberSearch references where the information comes from.

Further use cases

The end “of the line” is far from being reached here, various other use cases are conceivable and possible.

Here we are looking forward to suggestions from innovators, in order to be able to work on further possible use cases together.