Cochlear is integrating conversational voice-to-text functionality into its CRM that could eliminate the need for its “sales road warriors” to type lead management notes into the system after meetings.
The system would make use of a number of AWS services, but principally Amazon Lex as the conversational interface and also a recently-launched “AI walled garden” that exists inside of Cochlear.
The internal AI system is based on Amazon Bedrock and its agentic extension, Bedrock AgentCore.
Cochlear IT networks and telecommunications director Andy McLaughlin told the attendees at the AWS re:Invent summit in Las Vegas that the walled garden was only released internally “about a month or two ago”.
Cochlear has affectionately named the walled AI system 'Clark' as an homage to pioneer of cochlear implant technology, Dr Graeme Clark.
“Imagine [as a sales road warrior] you're out there, you've seen a prospect, you've come out of that meeting, and you need to update that lead in Salesforce,” McLaughlin said.
“What we're doing is you call into a number, it recognises who the salesperson is that's called in, and they can enter a PIN as a bit of added security.
“They sit there and talk about the customer they were dealing with. Salesforce knows they were seeing that customer, so it's quite easy to link that to the right case.
“They [describe] the meeting and talk about the next call objectives.
“We then throw all that, using lambda functions, to Bedrock, which packages that whole conversation up into all of the right [data] fields that you can, then input straight into the CRM.
“And your salesperson would be very happy that they don't have to go and log into Salesforce to do this – they can do this straight after the meeting while it's all top of mind.”
McLaughlin said that Cochlear’s technology teams were still “playing with [this] at the moment, but the intent is to get it into production use.
“We're hoping to roll this out to our sales teams later in the [financial] year,” he said.

iTnews Executive Retreat - Security Leaders Edition



