Why We’re Building the AI Integration Layer
3 minute read
This is Part Six of Making Every Conversation Count, an ongoing video series featuring 8x8 CEO Sam Wilson and Liz Miller, VP and Principal Analyst at Constellation Research. Miller covers customer engagement, CX strategy, and AI-driven business outcomes.
When Liz Miller, VP and Principal Analyst at Constellation Research, and I discussed the topic of competing in the AI space, I went back in time for a fitting analogy: Imagine it’s 1997. Why would anyone in their right mind try to go head-to-head with Netscape? That would be like picking a fight with a mountain. You’d lose, and the mountain wouldn’t care.
Similarly, today I see our competitors chasing a minuscule edge in LLM capabilities while facing Microsoft and other giants. It’s unrealistic, unachievable, and frankly a waste of time.
At 8x8, our bet is on partnering with the folks already doing LLMs the right way. On top of that, we’re building a layer that improves what the best models out there can actually do for your CX.
Many companies fall into the trap of assuming that orchestration is adequately handled by the model they’re implementing. LLMs are powerful, it’s true, but they’re not that great at execution in isolation.
Language models are still largely text- and language-based; our world runs on voice, chat, SMS, meetings, and more. We already know that Claude, GPT, and the like can’t directly manage critical CX operations, such as making a voice call, managing an SMS thread, or enforcing a routing policy. But if they’re plugged into the right orchestration layer, they become capable of that and more.
We’re looking for a solution that helps manage dependencies, enforce policy constraints, and even catch execution failures.
As Michael Tessler, Co-founder and former CEO of BroadSoft, put it: "Voice, messaging, and meetings have reached maturity. The innovation now sits in the intelligence layer on top."
True innovation is centered on reliably deploying AI to production while building in both orchestration and mitigation. A well-crafted integration layer:
If you’ve been keeping up to date with tech news, you know that AI deployments have been going over-budget left and right, with realizations about cost overruns and token usage coming far, far too late. According to a CX today study, 73% of agentic AI implementations went over budget; some by more than 240%
Why too late? Because there was very little governance and minimal management built into many of these implementations. That’s just one more reason that the platform you choose for your business communications matters; it’s 2026, and it’s very likely that your organization is already working with AI across your customer experience functions. For your own sake (and the sake of your bottom line) the integration layer for AI is as important as the LLM itself.
When I think about AI in CX, I think about ensuring that it doesn’t break, blow the budget, or chase away customers. When it comes to CX enterprise solutions, the LLM is only as good as how well it's connected to real workflows, channels, context, and customers.
We’re not trying to out-model the model companies. Instead, we're focusing on the supporting infrastructure. You can stay flexible while connecting to only the best that the market has to offer.
Stay current on what matters in CX and IT. Subscribe to our LinkedIn newsletter for regular analysis on the decisions shaping customer experience, technology, and AI. Clarity you can act on.
