Skip to content

Everyone Wants an AI Tool – But Do They Know How to Train It?

2 minute read

With Artificial Intelligence being such a hot topic today, many organisations are keen to adopt modern platforms, tools, and solutions. However, once they do so, they discover that they cannot achieve the desired outcomes.

The reason for this?

AI is smart, but it needs training and customising.


To teach an AI tool to work correctly, you need to feed it data. More importantly, you need to teach it how to interpret this data. Otherwise, it may draw incorrect conclusions.

This occurred a few years ago when Amazon developed an algorithm to automate its hiring process. Soon after, it was discovered that the automated tool was discriminating against female applications. This discrimination was not surprising, considering that it was the result of the data it was fed – that is, the resumes of Amazon’s software engineers, the majority of who were male.

It follows then, that your AI tool requires more than just data. It requires further fine-tuning and customisation to ensure that it does not draw incorrect conclusions.

Teaching an AI tool how to properly interpret data can be a long, challenging, and burdensome process. For this reason, one can use ‘pre-trained models’ to save time. While these models still require some level of training since they tend to be too generic, it is much simpler to fine-tune an AI tool with a smaller dataset than to start from scratch.


Second Nature offers an innovative AI sales coach that is fast and easy to train. Named Jenny, this AI coach can help salespeople practice conversations with prospects to help them reach their best performance levels. Jenny’s conversations are based on talk tracks and sales playbooks. More specifically, they are based on your own fine-tuning.

Instead of undergoing a long training process, all you would need to fine-tune this AI tool is to ask 3-5 of your best salespeople to have a sales conversation with it. When you first set up a simulation, the AI coach enters ‘learning mode’. During this stage, it asks questions, makes observations, and raises objections to learn from the salesperson’s answers. In addition, it is important to note that this tool must have conversations with several different people to understand that there are different kinds of ‘good sales conversations’ featuring different wording, speech patterns and styles.

To further facilitate the training process, Second Nature’s AI tool has a Learning Mode Visibility feature. This allows users to see the progress of the training and how much longer the AI tool needs before it is finished. Once the training is completed, the AI Sales Coach can begin having live conversations with the organisation’s sales reps. Although it will not be in learning mode once its training is finished, the AI Sales Coach will continue to learn from every conversation…

Read more about training AI here.

Keep up to date with the latest events, resources and articles.

Sign-up for the Engage Sales Newsletter