CallSine has emerged from stealth to train large language models (LLMs) using data about products, services and key marketing messages in a way that creates personalization messages for specific outbound sales engagements.
The company is making available both an application and an application programming interface (API) that integrates with customer relationship management (CRM) applications from Microsoft, Salesforce, HubSpot, Salesloft, Oracle, SugarCRM, Zoho and monday.com.
CallSine makes use of vector embeddings to analyze data and then employs retrieval augmented generation (RAG) techniques to retrieve the most relevant information or content pieces for each industry segment of specific customer profiles using data pulled from LinkedIn profiles and company Web sites.
That approach empowers organizations to merge the knowledge they have about products and services with a curated list of sales targets and other information to create more targeted outreach campaigns at scale, says CallSine president Logan Kelly.
For example, a salesperson could merge PDF documents, PowerPoint slides and HTML files pulled from their Web site to create a message personalized for a specific individual, he adds. “Personalization is what wins in sales,” says Kelly.
Arguably, the biggest sales challenge most organizations have is that individual sales representatives are only able to simultaneously work on a limited number of sales proposals. Many salespeople are also not naturally gifted writers, so the amount of time required for them to hone messages can be substantial. CallSine can automate tasks in seconds that previously would have required hours to complete.
In addition, it often takes months to effectively train new salespeople. CallSine provides an application that reduces the amount of time required to learn product, services and use cases that are relevant to customers in a way that makes it easier to onboard new members to a sales team, says Kelly.
That’s critical because finding and retaining sales representatives is always a challenge for most organizations. Every time a salesperson leaves an organization it can take six months or more to train another one, notes Kelly.
While just about every sales platform is adding generative AI capabilities, CallSine is making a case for a platform that doesn’t require much AI expertise to use. In effect, CallSine is automating the data science workflows that would otherwise be required to customize an LLM.
Of course, many salespeople are already using generative AI tools to craft messages, but the productivity gain enabled by using general purpose LLMs that have not been customized for sales is limited. In addition, given the tendency of LLMs to occasionally hallucinate, the more data they are exposed to the more likely the output generated will be relevant.
Finally, organizations should also make sure that sensitive data isn’t being shared with an LLM in a way that one day might be used to train the next iteration of the LLM. That could result in that data later becoming generally available to anyone using that platform.
Regardless of approach, there is no doubt AI is going to have a major impact on sales. The issue now is determining how soon and to what extent.