From self-service to agent assistance, routing, and workforce scheduling, AI has permeated every single building block of the customer experience stack. Yet AI is often deployed as a collection of add-on applications working on a subset of interactions or a segment of the customer journey; it’s not integral to the very composition of the stack. Additionally, the coming of age of large language models (LLMs) has highlighted the critical role that comprehensive datasets play in the quality and effectiveness of AI predictions. We find ourselves at a crossroads: CX software built historically around the queuing, routing, and handling of interactions will evolve to place an open data platform at its core. It’s time to shift this perspective and consolidate CX data to enable AI as a set of capabilities that can be activated across the entire stack.
Consider a simple customer-interaction scenario starting with an intelligent voice agent (IVA) and escalating to a human agent. AI can be used to summarize the IVA phase, categorize it for routing, provide real-time guidance during the live interaction, retrieve relevant knowledge articles, and later analyze the conversation for quality management. Each of these AI models was trained on interaction histories pulled out of existing systems with no cross-pollination. Providers must turn their troves of recordings, transcripts, and detailed interaction records into data platforms that can be expanded and shared across applications.
At the Enterprise Connect event in March, industry leaders from Five9, Genesys, Google, NICE, Salesforce, and Verint reached a rare consensus during Sheila McGee Smith's panel discussion on what comes after CCaaS. All vendors teased the idea that their platforms should become data platforms and orchestrate customer experiences across the enterprise.
This transformation requires a substantial overhaul of current product architectures, akin to the changes brought by the migration to the cloud. Let's keep in mind that no on-premises software was able to make the transition and that today's dominant contact center solutions were all created in the last 15 years.
Now, let's delve into the key components of this transition into data-centric CX software:
Unlocking the Potential of CX Data
Today's CCaaS platforms collect interaction details and record the actual conversations, whether voice or digital. While this interaction data holds invaluable insights, it lacks context and outcomes residing in other systems. These vital elements must be integrated and combined into the CX data platform so it can deliver its full potential. Furthermore, it should be possible to ingest data from touchpoints managed by different software to provide a holistic view of customer interactions.
Unifying Datasets
To integrate data from other systems, CCaaS platforms need to add to their interaction data stores three core components: a well-structured data model, interaction stitching capabilities, and identity correlation. The data model is pivotal, as consolidating customer data is no small task — past attempts by CRM solutions serve as a testament to this challenge. The data model should define the most valuable data elements, enabling the building of integrations to populate them. Interaction stitching allows linking together related interactions that are part of the same customer journey. Identity correlation allows connecting interactions with the same customer who may have been using different phone numbers, email addresses, or social handles. These three capabilities come together to create a rich and comprehensive CX dataset.
Embracing a Multi-model Future
We are entering an era where numerous AI models will coexist. This requires making data accessible to any of them, whether they are commercial applications or developed in-house. CCaaS platforms must open up access to their repositories and make them extensible so that as many SaaS applications as possible can leverage them without having to recreate their own mini-data repositories. Such an approach is critical with SaaS driving the proliferation of applications.
Integrating with Enterprise Data Architecture
CX data plays a significant role in the broader data architecture of enterprises. Thus, a CX data platform must seamlessly integrate into the enterprise data stack. Given the sheer volume of data involved, CCaaS providers need to turn to new integration techniques like streaming or zero-copy access to minimize data movement and meet the real-time requirements of contact centers.
Data Governance
Data governance is of utmost importance, encompassing regulatory compliance, the ability to anchor models to specific datasets, protecting proprietary information, and ensuring ethical data use. CCaaS vendors need to develop robust toolsets for its enforcement.
Game On
So far, CCaaS providers have been focused on adding AI capabilities to their suites. Interestingly, the new entrants in the contact center arena are the ones most active in building data platforms:
- Amazon Connect has enjoyed significant traction with enterprises building their data platforms on top of AWS cloud services: they can layer contact center software and AI features on top of their data infrastructure.
- Salesforce has anchored its contact center solution to its data platform. Recent additions include a new metadata framework to create a unified view of the data across multivendor applications and Einstein GPT Trust Layer.
- Twilio, capitalizing on its Segment — customer data platform (CDP) — acquisition, just announced CustomerAI capabilities.
- Verint, which aims to disrupt the CCaaS market with its Open CCaaS Platform, is basing it on its Open Engagement Data Hub.
- Qualtrics is building its foray into the contact center space on Experience iD, its unified repository of interactions and experiences.
I expect CCaaS companies to step up their efforts quickly. AI's prowess is no longer solely reliant on models but on the breadth and quality of datasets it can use. The future of CX hinges on transforming CCaaS into a data platform that enables multiple AI applications and orchestrates customer experiences across the enterprise. It will be a game-changer, and the game is on.
