CranemoorAI

Make your investment in AI really count

Get more valuable results from your interactions with AI by using an AI/RAG product that provides real control.

Build and submit prompts with a process where all prompts and responses are stored for later reference and which enables submission of your own PDF data files.

Grasp the opportunity to gain the important benefits of a fully structured and traceable approach that can contribute to ISO9001 certification.

Keep a human in the loop so you don't need to rely on agentic AI or to allow direct access to your core systems

The Product

Cranemoor AI's QueryFrame software provides facilities to maintain details of AI providers and the LLM models they offer.

It allows the user to store standard prompts, data files associated with those prompts, and submit them to different models.

The response text from the AI is recorded in a SQL database together with the date and time of the submission, the response latency and the token cost of the submission.

A basic analysis of responses can be implemented with PowerBI to be viewed either within client software or made more widely available via Intranet.

Select the Demo page to see the software in action.

Prompt Engineering

Prompt engineering is the art and science of defining the prompts (queries) put to AI LLMs (large language models) in order to generate results that are accurate and useful.

They are the link by which humans interact with AI models. The prompt must define clearly the task, scope, boundaries and context for the AI to use and can include criteria by which the quality of the output can be judged.

The output must be both precise and concise, relevant, accurate and safe. It should be free of ambiguity, hallucinations (i.e. output not based on sufficient evidence) and bias.

Mastery of prompt engineering can help to unlock the full potential of AI allowing it to become an invaluable research assistant in analysing large datasets and solving complex problems.

QueryFrame is designed to facilitate good prompt engineering by creating a library of prompts that can be analysed and improved over time. It allows files to be associated with prompts such that they can be reused or omitted as the prompt evolves.

The prompt itself can be built and rebuilt from a set of more basic clauses that provide a framework by the which prompt can be structured.

Select Task Provider edit Model edit Prompt edit Submit prompt Review submission (for the best viewing experience set player to full-screen)

AI in Bournemouth

3 Sided Cube - AI Workflows

Bournemouth Unversity - MSc Human Centred Artificial Intelligence

If you would like to discuss how QueryFrame might be of value to your organisation please contact Cranemoor AI either by email:

 info@cranemoor-ai.net

or telephone:

 020 3051 3765

or by post:

 Wilson House,
 2, Lorne Park Road,
 Bournemouth BH1 1JN