AWS is “investing heavily” in large language models (LLMs) for key contact center use cases.
The vendor made the claim in a YouTube video posted earlier this month, which introduces how it has already augmented its Amazon Connect platform with generative AI.
Indeed, Dave Lemons, Builder Solutions Architect at AWS, shared three existing use cases: expanding agent-assist, enhancing manager-assist, and improving customer self-service experiences.
Before introducing these, Lemons stressed:
[We’ll be] ensuring the right level of security, accuracy, observability, and removing bias to make sure generative AI outcomes can be both trusted and valuable.
To maximize these outcomes, AWS builds on the AI and machine learning capabilities already embedded into its CCaaS solution.
Use Case 1 – Expand Agent Assist Capabilities
Amazon Connect Wisdom – the CCaaS platform’s agent-assist offering – predates the explosion of LLMs onto the enterprise tech scene.
By analyzing customer interactions, Wisdom bids to understand customer intent and deliver agents helpful information to resolve a customer’s query in real-time.
It does so by digging into the knowledge management system, internal support documents, CRM, and other customer databases.
GenAI now takes this further. It sieves through all that knowledge to recommend customer responses that agents can review, edit, and send.
As such, the agent doesn’t have to craft their replies from scratch. That means agents can respond to customers swiftly. Moreover, with a human-in-the-loop, contact centers reduce the risk of sending out incorrect information.
“The responses and solutions can also continually improve based on what previously worked well,” added Lemons.
Finally, as Wisdom dips into various databases and extracts information from customer profiles, it may personalize its recommended responses to further bolster contact center conversations.
Use Case 2 – Enhance Manager Assist
GenAI assists managers with improved real-time and post-contact analytics. It does so through Amazon Connect Contact Lens.
The conversational intelligence platform already categorized customer issues, assessed customer sentiment, transcribed conversations, and automated the scoring of agent performance.
AWS has expanded these use cases with GenAI. Lemons explained:
There’s an opportunity for LLMs to produce concise summarizations of long conversations and capture the most important information.
The existing transcription capability makes this possible, which the LLMs work their magic on to capture the most crucial takeaways from the conversations. That includes the contact reason and outcome of the interaction.
That capability reduces the time agents and managers spend taking and reviewing notes.
Moreover, the summaries prove helpful for context sharing when transferring contacts between agents and supervisors in the case of an escalation.
GenAI also supports managers as they review contacts, supplementing the existing automated call scoring evaluation form.
Indeed, it can provide answers – and a rationale – to questions such as: Did the customer leave the call satisfied?
Meanwhile, it spotlights agent coaching opportunities – based on the transcript – to save managers time. An example of this is in the screenshot below.
Use Case 3 – Improve Customer Self-Service Experiences
Learnings from agent- and manager-assist use cases will help businesses improve customer self-service experiences.
For example, businesses may understand their most prominent, frustrating contact reasons and apply self-service to support customers in overcoming them.
Moreover, the intelligence will inform new business processes and workflow configurations.
Yet, GenAI has also enhanced self-service solutions in their own right, including AWS’s conversational AI platform, Amazon Lex, which sits inside of Connect.
Indeed, GenAI can increase the scope of such applications, so they automate interactions without developers specifying intents and building out conversational flows.
As such, virtual agents can solve queries they’ve received no training to handle.
After all, with augmented GenAI, they can understand intent, scour safe knowledge sources – FAQs, agent support content, web links – and provide answers to various contact reasons.
In doing so, the virtual agent may cite the source it used to provide the response, giving the customer another well of information if they wish to dive deeper.
More to Come from AWS
“By learning what drives great agent outcomes and layering in analytics for managers, knowledge content can be improved,” summarized Lemons. “This can drive enhanced self-service and agent-assisted interactions.”
Yet, there are many opportunities that AWS can snatch to extend the use of GenAI across its Amazon Connect platform.
For instance, it could follow the lead of Salesforce and release a feature that automates the creation of knowledge articles for emerging contact reasons.
Alternatively, it may consider a tool that pinpoints opportunities for customer service automation – as Five9 has.
Yet, there are many paths it could take as CCaaS vendors battle to use GenAI as a differentiator and further their standing in a busy market.
To gain a better idea of what these paths might look like, check out our article: 7 Generative AI Uses Cases for Contact Centers
