Contact us
Sales Email Analyzer helps employees to process responses to automated sales inquires
Sales Email Analyzer helps employees to process responses to automated sales inquires
Client:
Region:
CEE
Industry:
Hospitality and tourism
About the Customer
Screbo Innovation Ltd. is a medium-sized company operating in the vibrant hospitality and tourism industry in the Central and Eastern European region, renowned for its innovative and creative solutions that make the daily operations of units in the hospitality sector a lot more efficient and streamlined, thus enhancing their overall productivity and profitability.
They have a heavily utilized automatic sales pipeline that reaches out to potential partners every day of the week.
Customer Challenge
Screbo Innovation Ltd.'s sales representatives and assistants have a lot of replies from their Automated Sales Pipeline.
The automation reaches out, but to process the results they still use manual labor, which is preferable, since in sales it's better to have a hands-on approach.
Manually prioritizing these can take up a lot of time, which could be spent with potential clients.
However, emails will have a number of differing responses, and obviously, these can be categorized as uninterested, or interested. There are plenty of cases when you are not having a business relation with the responder, since it’s not relevant to them, maybe they have already solved their business problem, or sometimes the system can misfire and reach out to a person who finds the opportunity unappealing.
In these cases, it is a waste of time (in a statistics sense), to manually process the responses, when you could be dealing with someone actually interested, which is unproductive in a large-scale operation.
Solution
To address this challenge, Screbo Innovation Ltd. partnered with SnapSoft. Using AWS Lambda and Amazon Comprehend, SnapSoft created a cloud-based Artificial Intelligence service to predict client interest. It also integrates with the Gmail domain so that the automated sales email responses can be preprocessed for Sales.
The first step is to log in to SEA (Sales Email Analyzer), so the users of the sales email accounts may give permission to our service to process the emails contained within the used account. This can be done with our provided API Gateway URL, which will open the authentication window using the Login AWS Lambda.
After logging in and giving the necessary permissions, the user will be informed if the registration for the service was successful.
The application by default will schedule an hourly processing rate of emails whose value is chosen based on real-life statistics and cost analysis of AI service inference.
This hourly rate can be changed on demand of course since we provide our deployment with a CloudFormation Stack compatible declarative configuration that contains customization variables and enables the use of expressions that can set up the custom needs of customers.
There is nothing to be done by the user from this point on; whenever a new email is found on the registered account, the Sales Email Analyzer will query the content and process it using our custom-made Amazon Comprehend classification model. This model is trained on real-life data gathered by us, which was made with a large number of anonymized sales reach outs and their responses, and was optimized with AI techniques. We use batch asynchronous data processing with Amazon Comprehend Jobs, which can process these gathered data from multiple accounts.
When the classification model is finished with labeling the email contents with the proper classes - which could currently be positive or negative - it will create labels using Gmail and assign them based on the classification job to their proper message threads. These labels can be used to filter, and categorize the emails that are relevant to our user.
Results and Benefits
Using the Sales Email Analyzer cloud service Screbo Innovation Ltd. was able to reduce time spent on manually prioritizing their sales responses; it makes it easier to detect unusable negative message threads and people who do not wish to be contacted by them. This makes the sales pipeline significantly more efficient since they can filter out these contacts. Eliminated bad sales leads provide a way to spend most of the time on potential clients and customers. When a lead is found this way, there is a middle layer between the automation and the uniquely tailored hands-on customer-focused flow that an experienced representative can provide. This layer can be used as a preprocessing step for sales.
This way, it's easier to make educated choices based on the level of positivity in the messages threads, making new sales strategies much easier to implement, test and optimize the process as a whole.