This part of a blog series examines AI Voice Agents’ impact on the CCaaS market. You can find part 1 from last week here – https://www.cavell.com/what-is-happening-with-the-voice-ai-agent-market-understanding-ai-voice-in-ccaas-part-1/

Part 1 addressed the voice agent industry landscape, looking at the different companies developing in the space, the funding they had received, and who was building the ‘right to win’.

This piece takes that analysis and tries to understand the impact of Voice AI startups on the broader contact center landscape.

Why may the contact center vendors win?

It’s essential to stop and acknowledge that startups aren’t just working with Voice AI technology. Many large companies within the CCaaS space (and beyond) are spending a lot of money (spent $955m on Cognigy) on building full-stack Conversational AI technology.

There is a view of the future where these enterprise-scale technologies scale down to become accessible to the SMB market. Then the existing channel providers of these CCaaS vendors begin selling these new tools at volume, cornering the market with experience that these new companies cannot match.

Companies like 8×8, RingCentral, and AWS have already made substantial moves to make CX tools available to the SMB. This will shortly be (and already has been) followed by Voice AI and AI Receptionist use cases, as these companies seek to tackle the same use case gaps the startups have identified.

I don’t think it is particularly likely that these companies will choose to buy a vertical Voice AI startup unless they build a significant piece of technology that justifies the spend. The reality is that the majority of the identified use cases that will sell at scale for these companies can be fairly rapidly constructed—with a bit of a lag time to integrate with industry-specific tools.

My prediction in this space is fairly straightforward. If there is a profitable off-the-shelf use case, these companies (or another CX player) will build it. Then their partners with regional presence, vertical specialisations and a strong go-to-market approach will take that technology and begin to drive volume through it.

Will Voice AI destroy the contact center players?

To return to the title question, after examining and explaining the space, its easy to see why the answer is no. Some of these companies will carve out vertical or regional niches. Some will be acquired for their capabilities. Others will offer their technology for integration into the contact center stack.

Contact center players have an established market presence and ‘right-to-win’ in this area. They will be the companies selling the majority of voice AI to tackle the identified key use cases.

However, there is a broader question of the impact of voice AI on human agent figures, so let’s address that now.

Will humans be driven out of the contact center by Voice AI?

A question that I constantly get asked right now is whether technologies like Voice AI will push humans out of contact centers?

The answer is yes, but… not at the rates that the Voice AI startup companies would like.

I think before diving into this question, it’s important to address the different sizes of the market based on the use cases identified before, and the potential impact of the technology on each of those.

• SMB
  • Many voice AI startups we have identified are selling AI voice-powered use cases into specific vertical SMBs. For example, restaurant booking for three restaurant chains, or appointment booking for tradespeople.
  • When reading through the use cases on all these websites, you’ll see they are cautious with their wording, emphasising out-of-hours or 24/7 support. In the case of restaurants, they talk about reducing the amount of staff time spent answering the phone but not eliminating it.
  • In the short term, these companies expect their Voice AI tools to work alongside the existing people who answer phones at these companies, not replace them.
  • These companies also have a vested interest in fostering sentiments of local presence, which come with a personal touch.
  • Prediction if Voice AI continues to develop: These solutions help humans do their jobs in SMBs but never cost the job of the person who used to answer the phone. They enable that person to do other things and occasionally answer the phone because it works for the business to have a direct human touch.
• Mid-Market
  • While many of the same use cases are found in the mid-market as in the SMB, they gain complexity, often integrating with other systems or providing some form of custom automation.
  • These companies are also more likely to have roles solely dedicated to answering the phone.
  • Prediction if Voice AI continues to develop: Some workers will be replaced as Voice AI handles more basic queries. However, companies will keep humans on to handle more complex issues and monitor what the AI is doing for quality control.
• Large Companies
  • Most companies at this level use Voice AI for various custom and off-the-shelf use cases.
  • At this level, data sharing and cross-system integration are vital pieces of deployments.
  • Often, the Voice AI they purchase is a feature within a larger platform.
  • Accuracy and trust at scale is the biggest test of frontline AI in these companies as a % error turns into large amounts of frustrated customers.
  • Prediction if Voice AI continues to develop:
    • There is a key watershed moment, which is when Voice AI begins to be trusted with large scale front-line customer communications.
    • Now the adoption is slowed as it is being used in a constrained manner that uses LLMs to provide conversational understanding, intent and sentiment mining and broader conversational data.
    • This increased knowledge is used to better determine and customise workflows but not used in a freeform fully agentic manner.
    • For the immediate future this represents a gradual replacement of agents with Automation, but this falloff will ramp up as the technology gains more trust.
    • However, there are early indications that customer contact volume may also increase as the quality of customer service improves or becomes more digital. So the exact figure of agents that will be replaced is still to be determined.

When we discuss the displacement of humans by automation, one other important thing to remember is that many AI projects still fail. Many companies are finding challenges with data readiness, scaling capabilities and costs, so until these issues are resolved and deployments become more standardised, adoption rates will still be lower than the capabilities of the technologies imply.

For all of these reasons, Cavell’s global forecast predicts a demand for 16.9 million agent roles globally by 2029, and 1.9 million of that demand will be met by automated interactions and agents becoming more efficient due to agent assistance.

After this point offset will likely accelerate, though to what extent will depend on consumer attitudes towards automation and AI, which are currently more negative but may shift.

Routes to Market

Routes to market will be a key differentiator between different Voice AI players. We have already established that the companies with more long-term viability are those building broad use cases applicable to multiple industries. However, these companies still need vertical specialisation, which does not have to be developed in-house.

Companies that will be more likely to grow and thrive will build a channel with vertical specialisations who can take their broad technology and specialise. These will compete directly with the single vertical companies attempting to sell direct, but bring more established processes, broader skills and experiences from other industries to help refine the product/use case.

So, what does the landscape of the Voice AI players that we have reviewed look like right now?

Of the companies we reviewed,64% were selling direct only, 24% were selling with a mix of direct and partners. As expected, companies offering Customer Service AI or Horizontal AI platforms were more likely to sell via partners, as this underscores those platforms are taking a broader approach and relying on partners to customise the solution for their customer base.

In Cavell’s Channel Insights database, where we track 43k different channel partners globally, only 14.8k (roughly 34%) offer CCaaS services. 18.5k claim to offer AI, though as we have already determined, there are many poor solutions mislabelled as AI, so that number is likely inflated.

This tells you that the IT channel has not fully embraced providing CCaaS or AI services. This means there is still plenty of room for the provision of both services to grow.

Can Voice AI open up new revenue streams for service providers?

This is a top-of-mind question for many channel partners in the telecoms space. They have seen ARPU fall on UC and Telecoms-only use cases, they have considered (and many have) adding CCaaS to their portfolio, and now there is the potential disruption in the CCaaS space as well.

However, the answer isn’t that straightforward as it depends on market segment, skills, and internal processes/training. I imagine many companies will find themselves fairly easily able to pivot into selling automated answering and the simpler off-the-shelf use cases, as they are deliberately easy to deploy, and the use cases are easy for sales teams to understand.

However, when you take the step up to more complex Voice AI deployments, many service providers still lack the skills and education to effectively sell these solutions. These companies need internal AI experts, deployment specialists, and a sales and pre-sales team that is properly trained in the platform’s capabilities. Cavell has identified through its conversations with service providers that there is still a skills gap in the channel, which, coupled with the lack of knowledge of the capabilities on the buyer side, is leading to slower adoption than expected.

The opportunity is there; does the telecoms channel make sense to sell these new capabilities? But are they ready? Some are, and others need to upskill quickly or miss out.

But I fundamentally believe that the company selling voice lines, collaboration, networks, and phones is the natural fit to also sell voice AI and CCaaS services. At the low-to-mid level of complexity, the market does not need a separate range of providers just to provide straightforward voice-based services. The channel is already there for that.

The only real challenge to that is whether the use cases and deployment become so easy that deployment doesn’t need channel assistance, but most companies will want the reassurance of a channel provider, just like they do now for communication systems.

What Will Voice AI Really Do to CCaaS?
Most CCaaS vendors already have Voice AI projects in motion but two commercial realities could hit harder than they expect.

1. Pricing Pressure Will Escalate
   The per-seat license model has underpinned CCaaS economics for decades. Voice AI disrupts that in two ways:

• • Agent seat erosion – Vendors immediately lose license revenue if automation reduces headcount. Shifting to consumption pricing plugs that gap in theory, but in practice, it’s a thinner-margin, harder-to-mark-up proposition.
  • Utilisation dilution—An agent who spends 40% of their time on calls instead of 90% is no longer perceived as a “full seat” by the buyer. This prompt demands for hybrid human/AI licenses or unbundled features, which risk shrinking total contract value.
  • For vendors, the question isn’t if pricing will change, but whether they can design models that preserve margin while reflecting AI’s impact on labour.

2. Channel Readiness Will Slow Rollout
   Selling AI isn’t the same as selling voice-of-sale seats. While partners will adapt, Cavell’s data shows that only a minority are currently equipped to position and support advanced contact center solutions, let alone Voice AI. Vendors must carry more of the sales cycle until the channel catches up.

Bottom Line: Voice AI will land first as a pricing problem, not a technology problem. Vendors that fail to realign their commercial models before the adoption curve steepens will find their Voice AI wins offset by revenue leakage.

As mentioned above, not every channel partner the CCaaS vendors have is equipped with the skills to sell Voice AI. Vendors already must take up more of the burden of selling AI and need their channels to upskill. Voice AI is just another thing the channel needs to upskill to understand.

Conclusions

Voice AI will not kill the contact center, but it is making some large changes. The market is filling with different types of vendors but they aren’t all playing the same game. In last weeks post we broke down exactly what this vendor landscape is becoming and who is building the ‘right to win’.

For CCaaS incumbents, the advantage lies in scale and distribution. Once they solve pricing and shift away from pure per-seat models to hybrid or usage-based structures that preserve margin. They can deliver the most profitable Voice AI use cases to their existing base at volume.

The channel’s opportunity is real but uneven. Partners able to sell and integrate advanced Voice AI will capture strategic deals; those who stick to basic deployments will be commoditised. But in order to take advantage of any of these, the channel needs to build some real skills, not just in deploying and productising Voice AI but also in the sales team’s ability to recognise new Voice AI opportunities.

Thanks for reading.

This analysis is just one slice of Cavell’s CX intelligence. The 2025 CCaaS Market Evolution Report gives you the market forecasts, vendor insights, and strategic guidance you need to win in the Voice AI era globally and in 24 individual markets. Let’s talk! – Finbarr.begley@cavell.com