Data & Transparency At The Heart Of Everything We Do
Setting Chattie Up For Enterprise Use
Step 1: Provide Documentation About Your Business
In order to provide highly contextual answers, we need to learn about your business. You can provide us with relevant company documentation in any format. We use our proprietary data transformation process to prepare your data for AI consumption.
Step 2: Select LLM & Create Reference Library
Select which Large Language Model (LLM) to utilise for your use case. Clients designate a unique key, segregating them from other client instances. We create a reference library of your business information, stored in ISO27001 database in vector format, ready for the AI to retrieve at point of engagement.
Step 3: Specify Behaviour & Use Case
Characteristics, behaviour and business tonality are specified at the outset, along with the use case and any additional instructions that are expected of the AI. Our approach minimises AI hallucinations and ensures your business retains control of responses and the brand imagine you are looking to maintain.
Step 4: Deploy & Test
Chattie can be deployed internally or externally, on public facing websites or applications. Before launch, we undertake a rigorous testing process to capture any anomalies in behaviour or response. The user acceptance testing is also utilised to identify any gaps in the reference library content.
Step 5: Audit, Analyse & Improve
Once deployed into production, we audit every single interaction between your consumers and the AI responses. Not only does this allow us to produce new, unique insights about interactions with your business, but it also fosters continuous improvement and puts you in control of the AI.
