Why AI just doesn't seem to work for your sales team

While the AI hype remains strong, and tools such as ChatGPT and Copilot are rapidly gaining ground in sales teams, the question still remains is how and whether to deploy these tools in a business context.

One of the biggest mistakes a sales leader can make, is embracing AI without a clear strategy. They often don’t know how it can actually improve business processes. Instead of ensuring a lasting impact, these tools are often rolled out as ‘empty shells’. I can ensure you, this is a one-way ticket to failure.

For individual use, tools like ChatGPT and Copilot are a good fit. They can help with everyday tasks such as tweaking an e-mail, structuring a presentation, or quickly gathering information. Above all, these types of applications increase personal productivity: they enable sales reps to work faster and perform tasks more effectively.

"We're relying too much on individual knowledge"

But here lies the immediate limitation: the added value and impact of those tools become largely dependant on the skill of the user. A sales rep who knows how to formulate or ‘prompt’ a question well, will get better results than someone who is less skilled.

In a business context, this becomes a problem. When such tools are applied at scale, the output remains dependent on individual expertise, producing inconsistent and unreliable results. While just too many companies assume that the added value of tools like ChatGPT and Copilot for personal use bring, also apply when deployed company-wide. That's a fallacy.‍

Hallucinations galore

Take your sales department for example. It's tempting to use ChatGPT for questions such as: “What's going on in customer X's industry?” The answer generated is often useful, but insufficient to provide the in-depth insights you need as a salesperson.

This leaves us with three major problems. First, the question itself is inconsistent: salesperson A formulates his question differently than salesperson B, leading to divergent results. This lack of uniformity undermines the scalability of the process.

^{"When there's no uniformity, it won't be scaleable"}

Second, how do you know if the information generated is accurate and up-to-date? Inaccurate or outdated data can lead to hallucinations - incorrect or irrelevant information, which can hurt your business. Without direct access to correct data, this remains uncertain.

Third, these interactions are often stand-alone queries, without any integration with broader business processes. Only with information about which offerings are important and which products the customer has worked with in the past will you solve a use case.‍

Specialized models for specific processes

Companies looking to address complex business challenges should look for end-to-end AI solutions that go beyond simple chat interfaces. An effective AI tool in a business context should do more than just providing answers: it should help automate entire processes. This is why uman provides that end-to-end solution, and is more than just an AI assistant. Those often will only solve part of the problem. 

The reasoning capabilities of modern AI models have become significantly more powerful over the past year. These models can process data in multiple steps and draw logical conclusions in a coherent manner, reducing the risk of hallucinations - as long as the input data is correct and complete.‍

Data quality is key

Here lies the crux of the problem for many companies: the quality of their data. The idea that AI models need to be trained with massive amounts of data is now outdated. Modern AI models are so sophisticated that they can work effectively with minimal amounts of data. A small data set, such as a brief description of a product or service, can be enough to generate valuable insights.

But if we want those models to succeed, the inputted data must be accurate and current. Companies should provide a robust data governance system to ensure that changes, such as price adjustments, are made correctly and in a timely manner.

Only by having these fundamentals in place can AI be more than just hype and provide sustainable impact. Only then, it will provide tangible results for your sales team.

‍

This article was first published in DataNews.