MyCatalyst: FAQs

See Frequently Asked Questions referring to implementation and costings of MyCatalyst.

Implementation Information

Can MyCatalyst be implemented on my existing portal?

Yes, the AI-chat tool will be implemented into your existing interface by the

Content Catalyst team.

 

Will MyCatalyst adhere to our brand guidelines on our portal?

To maintain consistency with your existing portal, the MyCatalyst interface

can be customised to align with your company’s brand identity.

 

What support do you offer during implementation?

The Content Catalyst team provides comprehensive support during the

implementation of MyCatalyst:

Training: We will offer thorough training to ensure your team is equipped to

maximise the value of the AI chat discovery capabilities.

Onboarding: We will work closely with you over the first eight weeks following

implementation. During this time, we will monitor the quality of answers

provided and provide feedback to our development team to drive continuous

improvements.

 

What publication formats are supported?

MyCatalyst will generate answers from content originated in PDF, PowerPoint

and Word documents.

Currently (April 2024) only text in documents is readable by GPT, which means

that images in documents and data in tables cannot be extracted into

answers. We are working to introduce this feature.

Currently DataViewer, PowerBI Visualiser and plug-in products are excluded

from MyCatalyst answers.

 

Is MyCatalyst integrated with current user workflow tools?

No, the standard workflow tools (clippings / share etc.) are not available for

MyCatalyst generated answers. We plan to build this functionality over the

next few months.

 

Which version of GPT is used?

GPT-4 Turbo has been designated as the default model within MyCatalyst. Our

rigorous testing has demonstrated that answers provided by the GPT-4

model are substantially more accurate than those of preceding versions.

(Please be aware that there is a considerable variance in the pricing structure

across different iterations of OpenAI’s GPT models).

 

How can I set-up account in OpenAI to work with MyCatalyst?

Your account manager will guide you through setting up an OpenAI account

and obtaining an API key to connect with MyCatalyst. They will help you create

an OpenAI account to generate the necessary API key.

 

What plans have you to further develop MyCatalyst?

We believe generative AI will play a significant role in delivering analyst

research over the next few years. We intend to invest heavily and develop a

fully personalised user experience

 

Costs Information

 

Why is Content Catalyst charging for MyCatalyst?

Content Catalyst is charging for the MyCatalyst feature to help fund the

ongoing development and improvement of the platform. Maintaining and

enhancing the software requires significant resources. By charging for new

features like MyCatalyst, we can continue to evolve the platform and provide

robust support to users. This upgrade is intended to enhance the overall

experience for Content Catalyst clients.

 

How do usage fees work?

OpenAI charge for their services based on usage. The API key you connect in

your workspace setup links your API usage in Team-GPT directly to your OpenAI

account, ensuring usage costs are accurately billed to you.

You do not need a ChatGPT plus subscription to use Team-GPT. It only works

within OpenAI’s platform and doesn’t include access to GPT-4. Team-GPT

works on the basis of Open AI’s API, which requires pay-per-use billing. Usage

billing happens automatically at the end of each month and it is handled

directly by OpenAI. This not a flat recurring fee like a ChatGPT Plus Subscription.

If you don’t use the API, there is no charge. See Open AI's usage policy here. 

 

Note: OpenAI API Usage billing is handled entirely by OpenAI, and Content

Catalyst is not in any way involved in the transaction.

See Open AI Pricing here. 

 

What will OpenAI’s fees be for my instance of MyCatalyst?

The pricing for OpenAI's services is based on the usage of 'tokens', which

correspond to the length of the chat interactions. However, there are several

variables that can impact the actual costs, including; the specific version of

GPT used, the number of users on the instance, the number of prompts per user and

the length of the prompts.

Given the multiple factors involved, it is difficult to provide a precise cost

estimate upfront. Your account manager will work with you to forecast the

likely costs based on your anticipated usage patterns.

After a few months of implementation, we hope to have ‘real world’ data from

client sites and will be more accurately be able to forecast costs.

 

What is a ‘Conversation’ for billing purposes?

For billing purposes, a 'Conversation' is defined as a series of up to 5 questions

within the same thread. This includes:

- The initial question

- Up to 4 follow-up questions

The cost of the Conversation increases as more follow-up questions are

asked, as the system needs to process the previous questions, answers, and

relevant content snippets along with the new question.

 

How do I set up an OpenAI account for billing?

1. Go to your OpenAI account

2. Navigate to Billing settings

3. Go to Payment methods: Click “Add a payment method” and add the

necessary details.

4. Set usage limits: Navigate to Usage limits and set limits your comfortable

with. Do this to prevent unexpected charges and get usage alerts.

 

Can we limit OpenAI’s fees?

Payment is automatically taken by OpenAI at the end of each month but you can

also set a usage limit. When your usage limit is met, questions will no longer be

able to be asked using MyCatalyst until the usage limit is increased or reset at

the start of a month.

 

Will my content (intellectual property) be used to train OpenAI’s LLM?

No, MyCatalyst uses Retrieval Augmented Generation, which means that small

snippets of relevant content are found separately before being sent as part of

a prompt. Open AI’s terms state that prompts used in their APIs are not for

training future models. Review OpenAI’s privacy policy here. 

 

Can my customer access content they are not licensed to use?

MyCatalyst chat has a unique feature that respects user licensing. The

generated answers only use content and data that the user (your customer /

subscriber) has permission to access. This ensures that the answers provided

are drawn from the user's licensed products.

In the future, MyCatalyst plans to highlight any unlicensed content to chat users.

This will create an opportunity to upsell additional products or services.

 

What is Content Catalyst’s relationship with OpenAI?

We are using OpenAI’s API infrastructure to power AI interactions through

MyCatalyst. We have no other relationship.

 

What risks are associated with implementing MyCatalyst?

With such a new and powerful technology there are risks associated with

implementing AI. Using artificial intelligence to generate answers from your

analyst content does present some risks. To mitigate these hazards, we have

consulted an experienced intellectual property lawyer to identify risks and

potential liabilities associated with deploying MyCatalyst. Your account

manager will discuss these with you and deal with any concerns.

 

What happens if an answer is incorrect or misleading?

MyCatalyst is designed to be a helpful assistant, but it may occasionally

produce responses that are incorrect or misleading. This is known as

"hallucinating" information, and it’s a byproduct of some of the current

limitations of frontier Generative AI models.

Users should not rely on MyCatalyst as a singular source of truth and should

carefully scrutinise any important advisory answers served by MyCatalyst.

Users can use the thumbs down button to let us know if a particular response

was unhelpful.

 

Is Content Catalyst liable for ‘mistakes’ in answers served by MyCatalyst?

We are not liable for any inaccurate, misleading, incorrect or unexpected

outputs from MyCatalyst. We will provide a full set of terms and conditions

before implementation that covers warranties and disclaimers.

 

Do you recommend adding a user warning?

We recommend adding ‘warnings’ for MyCatalyst users to ensure they

understand that answers are generated by AI. We will advise on specific

language.