Amazon Web Services (AWS) has unveiled new artificial intelligence capabilities for its Amazon Connect cloud contact centre platform, wagering that the technology can make call agents more efficient and improve customer service.
The enhancements draw on large language models and other AI systems available via Amazon Bedrock, an AWS interface allowing companies to integrate AWS offerings with advanced machine learning.
The launch comes as more organisations field customer inquiries through call centres and online chat platforms. Cloud-based tools like Amazon Connect, analysts say, allow smaller businesses to launch contact centres swiftly instead of investing in legacy on-premises systems.
AWS, the cloud unit of e-commerce group Amazon, contends that its latest AI features will help users handle customer issues faster, capture vital details from interactions, and build automated chatbots requiring less manual programming.
“The contact centre industry is poised to be fundamentally transformed by generative AI, offering customer service agents, contact centre supervisors, and contact centre administrators new ways to deliver personalised customer experiences even more effectively,” said Pasquale DeMaio, vice president of Amazon Connect at AWS.
The most transformative addition could be Amazon Q in Connect, according to technology experts. The tool supplies call agents with suggested responses to customer inquiries by analysing intents and scanning for relevant data sources.
For a rental car company, Amazon Q might detect a caller wants to modify their booking and then automatically generate a reply about applicable change fees, while walking the agent through steps to alter the reservation.
The feature builds on Amazon Connect Wisdom, a repository of questions, answers and processes that helps agents address common issues. Amazon Q takes real-time guidance further by responding to unique cases using language models.
Analysts say squeezing more productivity from agents could enable contact centres to scale operations without ballooning headcount – an urgent priority given today’s tight labour market.
“This should allow agents to handle more calls per day and reduce handle times,” said Ryan Daily, CEO of CallMiner, which develops speech analytics software.
Amazon Connect has also introduced AI-generated summaries of calls through a feature called Contact Lens. Instead of poring over full transcripts, supervisors can quickly review summaries pinpointing topics, sentiment and follow-up actions.
Meanwhile, the company is touting new natural language abilities enabling administrators to easily build chatbots for websites by describing their desired bot in plain terms. Amazon Lex will then generate suitable responses for unusual customer inputs that might otherwise require handing off conversations to human agents.
Finally, machine learning techniques will automatically stitch together unified customer profiles from data stored across assorted applications – key for delivering personalised service.
The launch comes as Microsoft and other tech giants similarly talk up AI’s potential in customer engagement and sales scenarios. Continued advances in generative language models that can generate text, such as Google Brain’s PaLM, have fuelled commercial interest.
AWS executives believe melding such models with Amazon Connect’s contact centre scaffolding provides a potent combination. Early adopters of Amazon Q in industries like healthcare and financial services are already reporting positive results, Mr DeMaio said.
Still, effectively implementing these tools typically requires tweaking pretrained models using company-specific data so responses seem natural and helpful, not stilted and confusing. AWS will likely lean on its professional services organisation to aid customers incorporating the new AI capabilities, experts said.
The question now is whether the additions, anchored by Amazon Q in Connect, prove smart enough to transform call centre operations.
“This is a bold move by AWS,” Mr Daily said. “The next year will show if contact centres adopt these new innovations at scale.”