We get asked this all the time. The answer is simple:
KnowledgeFlow is focused on solving your specific business problems. Copilot is configured for generalised use.
Our priority is the success of the teams using KnowledgeFlow, not get as many licences as possible across an organisation. We differentiate ourselves in three key areas:
Sector specific skills and experience
- We’ve invested heavily in the last two years learning how to develop and fine-tune Retrieval Augmented Generative (RAG) AI tools specifically to solve sector specific knowledge and workflow problems. That ‘bottom-up’ approach means we have a real-world impact very quickly. Copilot by its very nature is ‘top-down’ and has to take a ‘lowest common denominator’ approach.
- We believe AI should be ‘business-led’ not ‘technology-led’. Our sector experts have years of experience in applying RAG AI to solve sector specific problems. Too often we’ve seen Copilot implemented by techies who don’t understand the use case (or even worse, don’t have a use case and are trying to invent one to justify the spend).
- We’re LLM agnostic. We help people select the best models for the problem they have. Yes, we work with Microsoft Azure and Chat GPT all the time. They are great. But we can also build in AWS Bedrock if that’s what’s best for the customer.
Ongoing roduct development on your solutions
- We’re constantly updating tools as new models are released, and old models are retired. Knowing how to manage and ‘refine’ the models to produce accurate and explainable results is hugely important for building users’ trust in AI tools. Copilot has a more limited ability to quickly and easily adjust important settings and data handling techniques because changes can affect the whole organisation.
- We’re working closely with every one of our customers to get their input to developing new solutions, features, and enhancements. This means we’re building what our customers want and need in real time.
- We can very quickly tailor the three main components of the product: 1. UI/UX; 2. System prompt; and 3. the training Document/Index. With an enterprise-wide solution like Copilot, there (rightly) needs to be limits on activities like advanced prompt engineering, strictness calibration, and model parameter exposure to maintain accessibility for a wider user base.
Dedicated support
- We pride ourselves on how quickly we respond to issues, questions, and enhancement requests.
- We understand that you need to make a business case for AI. Two years ago, next to no one had a budget line labelled “AI”, so this is new for everyone. The KnowledgeFlow Admin Portal gives you instant access to user information and stats o you can see what’s going on, understand the questions that staff are asking and therefore identify problem areas in your business. That anables you to create real-world impact reports like this one for North Yorkshire Council.
