Frequently Asked Questions

What Large Language Model does Enki use?

Enki uses various models in the software, depending on the use case, from OpenAI’s GPT-4 to more proprietary models trained from scratch. Enki’s GPT-4 also added improvements to the default GPT-4 model with features such as traceable sources (you can click URLs to see where the results are coming from), up-to-date data (we can extract recent insights), and hallucination minimization. Connect with us to see in a real life demo as well.

Is the model trained with open source data only, or also energy specific sources that require license?

The model is trained with both open-source data and energy-specific data that Enki scrapes from the web. For sources that require a license, Enki uses what is publicly available. For example, reading the paper with ScienceDirect might require a subscription, but the abstract is available publicly.

How often is Enki data updated with the latest information?

The data can be updated at any interval that makes sense for Enki’s users. When Enki runs a report, the model is updated with data up to the current point in time, and it can be refreshed at regular intervals (daily, monthly, quarterly, etc.).

Does Enki use customers’ data to train and improve the model?

The user data is not used to train the model. Enki does not and will not use any personal, private or confidential data in model training.

Does Enki keep any confidential information?

Enki does not store any confidential and private information about its clients and users.

Do I need to install Enki on my machine?

Enki is a web application and does not require any installations. Enki runs in Azure cloud.

Is it possible to weight the rating of the sources? E.g. we want focus on what we consider as the most relevant sources.

Yes, users can provide us with a list of sources that you find most valuable to your team, and we will rank it higher. Enki also has a built-in list of high-credibility sources and will allow you to filter high-credibility sources from low-credibility sources.

Do you have guidelines of how to secure as efficient usage as possible?

Yes, we can go over best practices as part of onboarding.

Do we have to create a login for each dashboard, how can we share the dashboards internally?

Enki protects all the dashboard with a password. Sometimes our clients make a request create a special share link to allow ease of use within their teams. To accommodate the client’s requests, we create a shared complex key that cannot be decoded.

How does Enki Research differ from other solutions like Chat GPT 4, Co-Pilot, Bing or Bard?

Enki Research is a tailored AI research agent for the energy transition sector, delivering precise, fresh, web-sourced market intelligence through an intuitive chat interface.

Unlike other solutions like GPT4, Enki differentiates itself with the following:

  • Rank data based on what matters most to the energy transition domain
  • Deliver sector-specific insights with analyst-like precision and provide sources – this minimizes hallucinations
  • Foster strategic thinking through guided questions and action suggestions
  • Write market intelligence reports with up-to-date information and transparency to save you time and impress your leadership team

Can you customize Enki dashboards with our team’s logo like Salesforce does?

Yes, we can customize your dashboard with your logo and team name. We can even customize your URL such as if requested, with password protection.