Conversational AI platforms: Navigating continuous change
Conversational AI platforms provide today's benchmark for self-service and AI-driven customer engagement. While not the only option for delivering conversational experiences — I introduced a framework a year ago to evaluate the different options across sales and service — these platforms have set the standard for what’s possible in the space.
Large analyst firms like Gartner, Forrester, and IDC provide comprehensive evaluations of these platforms through their respective frameworks: Magic Quadrant, Wave, and MarketScape.
The core capabilities of these platforms span four critical areas:
Integrations with back-end systems, communication channels (contact centers, web chat, messaging), and knowledge sources.
AI technologies for speech and natural language processing, understanding, and generation (NLP/NLU/NLG).
No-code conversation design environment to build applications.
Toolsets for defining, testing, and refining intents and entities.
In just 18 months, GenAI has reshaped the conversational AI market. Platforms have undergone two rounds of evolution—sometimes requiring a complete rebuild of functions—and must keep pace with relentless innovation. A new generation of platforms is emerging, driven by several trends and evolving needs.
Model choice - Built-in proprietary NLP/U technology may have been a key differentiator a few years ago, but today, what matters is the ability to leverage the right, constantly evolving technologies. Platforms must not only provide access to best-of-breed AI models but also orchestrate their use effectively, selecting the most appropriate technology for each interaction.
Model combination - While GenAI expands natural language capabilities with minimal learning, traditional NLP remains essential for its proven reliability and cost-effective scaling. The ability to support both on a single platform is critical. More broadly, platforms must support the growing trend of combining multiple specialized models for optimal performance.
Model optimization - GenAI has transformed intent management, enabling more intuitive definition and refinement. However, new toolsets are required to customize the various AI technologies. This goes beyond just prompting and using RAG techniques to involve chaining and combining models, as well as tuning models and optimizing inferences.
Voice AI - The ability to provide access to best-of-breed AI technologies extends to voice AI. Latency and scaling challenges continue to drive enterprises toward best-in-class speech-to-text solutions. Additionally, the detrimental impact of misunderstandings on customer experience highlights the critical need for top-tier technology. Furthermore, the frontier in natural voice continues to advance, with text-to-speech technologies achieving near-human quality and speech-to-speech models just around the corner.
Transactional and informational interactions - Historically, Conversational AI platforms focused on transactional interactions, while informational ones were managed by other solutions using techniques like question-answering or semantic search. GenAI is driving the convergence of these approaches, with platforms now required to support both types of interactions and create blended experiences. For example, you can design an experience that first uncovers specific intents and if none are matched, resorts to open-ended questions before escalating to a human. Alternatively, open-ended dialogue can be used to gather information for a specific intent in a more fluid, less scripted manner.
Agentic - Agentic AI is reshaping the landscape once again. While I believe process-defined, deterministic workflows will continue to dominate CX and sales use cases in the short term, the promise of autonomous agents will transform how applications are developed in the long term.
Integrations - Many platforms have solid integration capabilities and will naturally be compelled to turn them into agents. This evolution will require robust orchestration and governance. As agents become standard components in most applications, platforms must also be able to connect to and leverage them effectively.
Context - Platforms must manage context over longer periods, evolving from point interactions to extended conversations. Since context is crucial for the overall experience and its personalization, it must be shared across applications that manage the entire customer journey.
Orchestration - Orchestration goes beyond interactions and AI—it must enable sophisticated AI-human collaboration. Emerging needs include allowing experts to share their tacit knowledge, as highlighted by Christina McAllister and Kate Leggett, or enabling proactive expert pairing, exemplified by companies like SentioCX.
Scaling through iterations - Successful projects often plateau after deployment, struggling to expand use cases. Platforms must enable rapid iteration and continuous improvement.
The tension between disaggregating functions for independent evolution and assembling an expanding set of technologies makes it difficult to predict what platforms will look like in a few years. Not all providers will successfully transition—some, burdened by technical debt, will be forced to pivot toward specialized solutions.
When evaluating platforms, the key is to define the flexibility you truly need and make tradeoffs accordingly. A purpose-built solution may be a better fit than a broad platform, allowing you to leverage the vendor’s deep domain expertise. However, that doesn’t eliminate the need for rigorous validation of their technology stack and architecture.
Given that 'platform' is a catch-all term in vendor messaging, it’s essential to cut through the noise and classify offerings accurately.
As conversational AI evolves toward the orchestration of conversations, technologies, and human-AI collaboration, use these trends as strategic lenses to guide your decisions. Above all, prioritize openness to navigate this evolving landscape.