At the end of all calls that become sales, salespeople need to inform the customer, all the benefits that are included in the service they are hiring with Claro's sales operation.
The quality assurance team created a web page for each type of product, with a checklist about what should be said to the customer, as there were many possibilities to be memorized.
Sellers open the page of the products they sold and start reading each topic, for each type of product. But there was such a wide range of products, each with its own specific features and benefits.
Salespeople would end up talking to the customer about benefits that their package didn't have or even forgot to report everything that their package had, including important contract information and payment method for the service.
After the calls, a company qualified by Claro, monitored the calls and if they found any errors at this stage, the sellers who had made that sale were penalized, receiving a decrease in their final commission.
In addition, Escale received a fine if the number of erroneous calls exceeded a certain percentage.Despite all the efforts of the quality and training team, these errors kept happening more and more often, which was already becoming unsustainable, since the value of the fine was already too high.
It was then that they sought help from the product team, to understand how to solve this problem that both business and internal users were experiencing.
My role in the resolution was initially to have a deeper understanding of the problem. To do this, I performed a series of activities. I sat next to the attendants and followed the entire sales process, to understand what were the biggest difficulties and pains in the checklist process and also understand what was the reaction of the end customer to the situation.
I also interviewed some stakeholders to map the process in a more macro way, also understanding what were the main business pains. Once I mapped the whole process and the pains, I shared everything with the team so that a feeling of empathy was generated.
Next, I included the stakeholders and users in the process and suggested an ideation activity so that we could all suggest different ways to solve the problem.
The solution we came up with was to automate this part of the process, as it involved a lot of logic within Claro's diverse product catalog. We also identified that as there were constant changes in the catalog, the most scalable way to do this would be if the quality team could have the autonomy to change the script and the logic behind it when needed.
From that moment, I outlined some wireframes of how this process would work and validated technically with the development team, once validated, was the time to validate the idea with stakeholders and the users. Only after both validations did we start developing a lean version of the product for testing.
We created an interface where the quality team could define the texts and logic behind the products. The sellers, on the other hand, would make the selling process they were used to, with the difference that now, when selecting the products, it would automatically generate the correct text that they should speak to the customer, without causing them to worry or have to think much in this process.
We did the first test with a small team of salespeople and identified a much higher assertiveness in the checklists. But there were still many mistakes. Some did not read correctly or willingly changed the reading order. Which made us think of evolving the product to the next level through text-to-speech technology, where audios were generated over the texts that defined by quality team.
This made us clear the amount of errors that happened in the checklist, so that the user did not have to worry about this step, they had to simply warn the customer about the checklist and play the audio. They reported that this made them feel less frustrated and worried about whether they would be punished or not.
