Is the Conversational AI Market growing now?
In our conversational AI market deck, you will find everything you need to understand the market
SUMMARY
Yes. The Conversational AI market is growing quickly now, but most of the growth has shifted away from traditional chatbots toward customer agents and voice AI that can actually complete tasks.
The category is becoming harder to track because its fastest-growing products are increasingly sold as customer agents, CX AI, voice AI or agentic AI. Older Conversational AI market estimates can therefore miss a meaningful part of what companies are buying today.
Established contact-center vendors provide some of the clearest commercial evidence. NiCE, Five9 and Salesforce are reporting AI-related growth of 52%, 78% and more than 200%, far ahead of the surrounding cloud or subscription businesses.
AI-native companies are no longer tiny experiments either. Sierra, Cresta, Decagon and Parloa alone now represent more than $400 million in disclosed or estimated annualized revenue, despite several of them being only a few years old.
Voice AI may be the most visibly accelerating part of the market. Using the bottom of Vapi's reported daily range, Vapi, Bland and Retell AI together are already handling at least roughly 115 million calls per month.
Enterprise adoption is running ahead of enterprise confidence. Most Conversational AI leaders surveyed by Rasa were expanding their programs even though confidence in handling complex conversations remained moderate and concerns had shifted toward transparency and compliance.
Traditional company chatbots tell a very different story. Consumer usage has barely moved since 2022 while the use of general-purpose GenAI for customer-service problems has nearly doubled, suggesting that customers want better AI rather than simply more chatbots.
Investment is booming but unusually concentrated. Five major customer-agent and voice-AI rounds totaled about $1.65 billion, with Sierra, Parloa and Decagon taking roughly 94% of that capital.
Strategic buyers are putting billions behind the same shift. Salesforce's proposed acquisition of Fin and NiCE's acquisition of Cognigy represent about $4.56 billion of value directed toward specialized conversational and agentic customer-service technology.
The weak spot is ROI. Spending and deployment are rising much faster than the number of companies able to prove financial returns, and AI-driven customer-service layoffs are still much less widespread than the headlines imply. Conversational AI is clearly in a growth phase, but the market is scaling before many buyers have fully worked out the economics.
This market map, featured in our conversational AI market deck, highlights top companies and startups in the conversational AI market
Is the Conversational AI Market Growing Now?
What does Conversational AI actually mean today?
Conversational AI today means AI that talks with customers or employees in natural language and increasingly completes the task inside that conversation.
That definition covers customer-service agents such as Sierra, Decagon, Fin and Parloa, voice-agent platforms such as Vapi, Retell AI and Bland, and the conversational automation built into contact-center platforms from NiCE, Genesys and Five9. We also include products such as Salesforce Agentforce when they are being used to handle customer conversations.
We would leave coding agents, research agents and autonomous back-office software outside the category, even when people interact with them through chat. Otherwise almost every AI agent becomes “Conversational AI” and the market stops being useful to measure.
The naming has shifted quickly lately. Vendors increasingly describe the same underlying work as customer agents, voice AI, CX AI or agentic AI. That makes old market-size estimates less useful, because a growing share of Conversational AI revenue is now reported under newer labels.
Is Conversational AI really growing, or are companies just renaming chatbots “AI agents”?
Conversational AI is genuinely growing now. The new “AI agent” language makes the market harder to track, but the commercial growth underneath it is real.
Genesys gives us a good example. Genesys Cloud reached $2.8 billion in ARR in its latest reported quarter, up nearly 35% year over year. Earlier disclosures showed that more than 70% of Genesys Cloud customers were already using its AI products and that AI accounted for 20% of new-business annual contract value during the fiscal year.
The product itself has changed. A traditional chatbot mainly retrieved information or followed predefined paths. Current customer agents can identify intent, pull information from company systems, update accounts, process requests and decide when a human should take over.
That explains why today's fastest-growing vendors rarely present themselves as chatbot companies. The market has moved toward software that tries to resolve the conversation rather than simply answer it.
If you want more recent data on this point, please see our latest conversational AI market report.
As this chart shows, and as featured in our conversational AI market deck, search interest in conversational AI has increased sharply
Are traditional customer-service chatbots still growing?
Traditional company-owned customer-service chatbots are barely growing in consumer usage, even while people are using conversational AI much more often elsewhere.
Gartner surveyed 3,566 consumers earlier this year and found that people were about three times more likely to use third-party generative-AI tools such as ChatGPT, Gemini or Copilot for a service problem than a chatbot supplied by the company they were dealing with. Use of third-party GenAI for customer service had nearly doubled over the previous year, while use of company-provided chatbots had remained statistically unchanged since 2022.
The difference gets clearer when we look at what people expect AI to do. Among customers already using GenAI, 58% had asked it to complete a task on their behalf. In B2B interactions, that figure reached 74%. Booking something, changing an account or submitting information is becoming part of the expected experience.
So the growth today is coming mainly from more capable conversational agents rather than another wave of FAQ bots. Consumers have had years to try the old version, and the usage data shows little appetite for more of the same.
Is voice AI the fastest-growing part of Conversational AI right now?
Voice AI looks like one of the fastest-moving parts of Conversational AI right now, with several young platforms already handling tens of millions of real calls every month.
Vapi says it has supported more than one billion calls and currently processes roughly one million to five million calls per day. Amazon Ring tested more than 40 voice-AI vendors before selecting Vapi and now routes all of its inbound calls through the platform.
Bland reports more than 55 million calls per month and more than 619 million calls handled to date. The company also raised another $50 million this year and says it now handles complex conversations for customers including Samsara, Kin Insurance and CNO Financial Group.
Stripe's Retell AI case study provides a third data point. Retell went from roughly $1 million to more than $10 million in annualized revenue within a year while reaching more than 30 million calls per month. Retell now says its annualized revenue has moved much higher, with a very small team.
Even using the bottom of Vapi's stated daily range, Vapi, Bland and Retell together account for at least roughly 115 million calls per month. The figures are company-reported and cover slightly different periods, but that order of magnitude is already far beyond a demo market.
| Voice AI company | Current scale we can verify | What stands out |
|---|---|---|
| Vapi | 1M-5M calls per day; 1B+ supported | Amazon Ring moved 100% of inbound calls onto the platform |
| Bland | 55M+ calls per month | More than 619M calls handled to date |
| Retell AI | 30M+ calls per month | Revenue grew from about $1M to $10M+ annualized within a year |
If you want more recent data on this point, please see our latest conversational AI market report.
This chart, included in our conversational AI market deck, shows annual VC investment in conversational AI startups
Are enterprises actually scaling Conversational AI today?
Enterprises are scaling Conversational AI now even though many of the people running these systems still do not fully trust them.
Rasa surveyed 30 enterprise Conversational AI leaders across industries including finance, healthcare, retail, government and telecom. Sixty-seven percent said they were expanding or scaling their programs. Yet their average confidence in AI's ability to handle complex conversations was only 4.37 out of 7.
That is the interesting bit: companies are scaling systems they still only partly trust.
Their concerns have also moved. Sixty percent of the Rasa respondents ranked black-box behavior or compliance as their biggest challenge, ahead of integration and deployment complexity. Ninety-three percent described transparency as very important or critical.
A larger study from TELUS Digital and Ryan Strategic Advisory surveyed 815 enterprise customer-experience leaders and found that human agents assisted by AI were already the leading delivery model in six of the seven customer-experience functions studied. Fully autonomous agents are growing as well, but the typical enterprise deployment these days combines AI with existing human operations.
Deployment is moving faster than confidence. That creates some messy implementations, but it is still real adoption.
Are companies spending more on Conversational AI now?
Companies are spending much more on customer-service AI today, and the AI parts of several established software businesses are growing several times faster than their underlying platforms.
NiCE's latest results put CX AI and self-service ARR at $362 million, up 52% year over year. AI now represents 15% of its cloud revenue. NiCE's total cloud revenue grew 12.6% over the same period, so its AI business grew about four times faster.
Five9 shows an even wider gap. Its latest quarterly revenue reached $312.4 million, with subscription revenue up 14%. AI revenue reached roughly $39 million for the quarter and grew 78% year over year, more than five times the growth rate of subscription revenue overall. AI now accounts for around 15% of Five9 subscription revenue, compared with about 9% a year earlier.
Salesforce sits at a much larger scale, although Agentforce covers more than customer conversations. Agentforce ARR reached $1.2 billion in its latest reported quarter, up 205% year over year. Salesforce also reported 3.8 billion agentic work units delivered across Agentforce and Slack.
We should not add these figures together and call the result a Conversational AI market size because the products and accounting definitions differ. The cleaner comparison is the growth rate: AI-related revenue is expanding much faster than the surrounding software businesses.
| Company | Latest AI commercial metric | Current growth |
|---|---|---|
| NiCE | $362M AI ARR | +52% YoY |
| Five9 | ~$39M quarterly AI revenue | +78% YoY |
| Salesforce | $1.2B Agentforce ARR | +205% YoY |
If you want more recent data on this point, please see our latest conversational AI market report.
This chart, included in our conversational AI market deck, breaks down Cognigy's playbook in conversational AI
Are Conversational AI startups making real money yet?
Several Conversational AI startups are already building businesses at a scale that would have looked exceptional for this category two years ago.
Sierra reported passing $150 million in ARR after reaching $100 million only a few months earlier. Cresta has crossed $100 million ARR. Sacra estimates Decagon reached around $100 million in annualized revenue after ending 2025 at roughly $44 million. Parloa says it surpassed $50 million ARR and reported 150% net revenue retention, which means its existing customer base was still expanding spending strongly after churn and contraction.
Using only those four companies, the latest disclosed or estimated run rates add up to more than $400 million of annualized revenue. We should treat Decagon's figure as an estimate and the others as company disclosures, but the aggregate still changes the picture considerably.
The speed is the striking part. These companies were not slowly building $50 million businesses over a decade. Sierra launched in 2023. Decagon was founded in 2023. Cresta's current growth accelerated as generative AI moved into contact centers. Parloa tripled its private valuation in eight months after its revenue passed the $50 million level.
Conversational AI now has several independent companies producing tens or hundreds of millions of dollars in recurring or annualized revenue. Funding announcements alone could not tell us that.
Is venture money spreading across Conversational AI, or piling into a few winners?
Investors are putting huge amounts of money into Conversational AI, but most of that capital is currently piling into a very small group of companies.
We checked five large customer-agent and voice-AI rounds announced this year. Sierra raised $950 million, Parloa $350 million, Decagon $250 million, Vapi $50 million and Bland $50 million. Together, that comes to $1.65 billion.
Sierra, Parloa and Decagon account for $1.55 billion of the total, or about 94%. Vapi and Bland represent only around 6% combined.
The valuation jumps are similarly concentrated. Sierra moved above $15 billion. Decagon reached $4.5 billion, three times its previous valuation in roughly six months. Parloa reached $3 billion after being valued at $1 billion eight months earlier.
This is a fast-growing market with an increasingly narrow group of perceived leaders. Funding conditions for the top companies are extraordinary, but they tell us much less about what the average Conversational AI startup can raise.
| Company | Latest round this year | Valuation |
|---|---|---|
| Sierra | $950M | Above $15B |
| Parloa | $350M | $3B |
| Decagon | $250M | $4.5B |
| Vapi | $50M | About $500M |
| Bland | $50M | Not disclosed |
This chart, included in our conversational AI market deck, shows annual funding in conversational AI startups
Why are Salesforce and NiCE spending billions to buy Conversational AI companies?
Salesforce and NiCE are paying billions for Conversational AI because buying proven technology and customers has become strategically important even for companies that already build AI themselves.
Salesforce agreed to acquire Fin, formerly Intercom, for approximately $3.6 billion. Fin's AI agent handles conversations across live chat, email, WhatsApp, SMS, phone and Slack. Salesforce already had a rapidly growing Agentforce business when it made the offer.
NiCE made a similar move with Cognigy, one of the better-known enterprise Conversational AI platforms. The transaction valued Cognigy at about $955 million. At the time of the deal, NiCE said Cognigy was expected to grow ARR by roughly 80% in 2026 and already served more than 1,000 brands.
Those two transactions alone represent about $4.56 billion of strategic value directed toward specialized Conversational AI businesses.
Both buyers already had large customer bases, AI teams and distribution. They still decided that production-ready conversational technology, integrations and enterprise deployments were valuable enough to buy rather than spend several years reproducing internally.
Does Conversational AI actually save companies money?
Conversational AI can save companies a lot of money in the right deployment, but the average enterprise still has a serious ROI problem.
Gartner surveyed 1,303 senior leaders and found that service and support organizations had allocated a median 12% of their 2025 budgets to AI, the highest share among the ten business functions studied. Only 24% of service and support leaders, however, could demonstrate positive financial returns across their AI use cases.
Spending is already large while proven returns remain much less common.
At the same time, we can find production examples that explain why companies keep trying. Salesforce says Fin agents resolve an average of 76% of support volume end to end. Amazon Ring said customer satisfaction improved after it moved all inbound calls onto Vapi. Large contact-center vendors also keep reporting faster AI bookings and earlier customer deployments.
The economic case depends heavily on execution. A company that automates high-volume, repetitive conversations and connects the agent properly to its internal systems can remove a meaningful amount of manual work. A badly integrated agent can simply add model costs, implementation work and another customer-service channel to supervise.
That is roughly where the market is today: spending can rise very quickly even while only a minority of companies have fully proven the ROI.
If you want more recent data on this point, please see our latest conversational AI market report.
This chart, included in our conversational AI market deck, compares the main business model options for conversational AI enterprise platforms
Are AI agents really replacing customer-service workers now?
Conversational AI is already reducing some frontline customer-service work, while broad replacement of human agents remains relatively rare today.
Gartner surveyed 321 customer-service leaders and found that only 20% of organizations had actually reduced agent headcount because of AI. Thirty-one percent had implemented or were planning AI-driven frontline layoffs, but 85% were expanding the responsibilities of human agents.
Many companies are taking the slower route. Sixty-three percent of the surveyed service leaders said they were reducing frontline headcount gradually through attrition while moving remaining employees toward more complex work.
Customers are also putting a limit on full automation. In Gartner's more recent consumer research, 87% said access to a human agent was essential when a company used generative AI for customer service. Half said AI had made service interactions easier, so consumers are clearly accepting the technology, but they still want an escape route when the conversation gets difficult.
The workforce effect is already real, especially for repetitive tasks and contact volume. The current evidence points much more clearly toward smaller or differently structured support teams than toward customer-service departments disappearing.
Is the Conversational AI market growing now?
Yes. The Conversational AI market is growing quickly now, with the strongest growth coming from AI customer agents and voice agents rather than traditional chatbots.
The latest commercial numbers are difficult to dismiss. Established platforms are reporting AI growth rates of 52%, 78% and more than 200% in businesses where the surrounding cloud or subscription revenue is growing much more slowly. Several AI-native customer-service companies have already reached roughly $50 million to $150 million or more in annualized revenue. Three voice platforms we examined are collectively handling at least about 115 million calls per month using the lower end of Vapi's stated range.
Capital is moving in the same direction. Five major financings we checked add up to $1.65 billion this year, while Salesforce and NiCE have committed roughly $4.56 billion combined to acquire Fin and Cognigy. Those amounts are too large to explain away as a few chatbot experiments.
There are still two clear brakes on the market. Traditional company-owned chatbot usage has barely moved since 2022, and Gartner found that only 24% of service leaders could demonstrate positive financial returns across their AI use cases. Enterprises are spending and deploying faster than they are proving the economics.
Conversational AI is already in a real growth phase today, but the growth has moved away from the old chatbot market. The money, usage and enterprise activity are concentrating around agents that can understand a conversation, take action and increasingly handle voice. That narrower version of the market is growing very fast.
If you want more recent data on this point, please see our latest conversational AI market report.
This chart, featured in our conversational AI market deck, illustrates revenue distribution by customer segment in the conversational AI market
OUR METHODOLOGY
This analysis tests whether the Conversational AI market is genuinely growing today by looking across real-world usage, enterprise deployment, customer spending, company revenue, startup traction, venture investment, strategic acquisitions, demonstrated ROI and workforce impact. We use these dimensions together rather than relying on a single market-size forecast.
We define Conversational AI as AI used to conduct natural-language interactions with customers or employees and, increasingly, complete tasks inside those interactions. That includes customer-service agents, voice agents and conversational automation inside contact-center platforms. We exclude coding agents, research agents and autonomous back-office software when conversation is only the interface rather than the core product.
That boundary matters more than it used to because the terminology is changing quickly. Products that previously sat comfortably inside Conversational AI are now frequently described as customer agents, CX AI, voice AI or agentic AI. We therefore follow the underlying use case rather than relying only on the label a vendor currently uses.
We prioritized recent operating and commercial evidence over broad market forecasts. Revenue, ARR, customer adoption, production call volumes, deployed usage and enterprise spending carry more weight in the analysis than funding, valuations or company positioning. Funding and acquisitions are used mainly to show where investors and strategic buyers are committing capital.
We do not add vendor AI revenue figures together to create a market-size estimate. NiCE, Five9, Salesforce, Genesys and AI-native vendors define AI revenue and activity differently, and several products extend beyond Conversational AI. Their growth rates are more useful here than an artificial combined total.
Voice-AI scale is treated conservatively. For the comparison between Vapi, Bland and Retell AI, we use the bottom of Vapi's reported one-million-to-five-million-call daily range and combine it with the companies' reported monthly volumes. These figures are company-reported and refer to slightly different periods, so they are used as an order-of-magnitude measure rather than a precise market total.
For venture activity, we aggregated the five large rounds examined in the article: Sierra, Parloa, Decagon, Vapi and Bland. Their latest rounds total approximately $1.65 billion, while Sierra, Parloa and Decagon account for about 94% of that amount. We use the concentration as part of the finding rather than treating total funding alone as proof that the whole startup market is equally strong.
We also looked deliberately for evidence that cuts against the growth story. Gartner's consumer data on stagnant company-chatbot usage, the limited share of service organizations demonstrating positive AI ROI, continued demand for access to human agents, and relatively modest AI-driven headcount reduction are important counterweights to the revenue, usage and investment data.
Key sources used for this analysis include: Genesys on Genesys Cloud ARR and AI adoption, Gartner on third-party GenAI, company chatbots and customer behavior, Rasa's 2026 State of Enterprise Conversational AI, TELUS Digital and Ryan Strategic Advisory's enterprise CX AI survey, NiCE's Q2 2026 results, Five9's Q2 2026 results, Salesforce's fiscal 2027 first-quarter results, Sierra's $150 million ARR milestone, Cresta's $100 million ARR milestone, Parloa's revenue and retention disclosure, Vapi's production-scale disclosures, Bland AI's call-volume disclosures, and Stripe's Retell AI case study.
For financing and strategic transactions, key sources include: Sierra's latest financing announcement, Parloa's $350 million Series D, Decagon's Series D announcement, Vapi's Series B, Bland AI's Series C, Salesforce's agreement to acquire Fin, and NiCE's Cognigy acquisition announcement.
For the workforce analysis, we also use Gartner's survey on AI-driven customer-service headcount reduction and Gartner's research on changing human-agent responsibilities and attrition. The final conclusion comes from the consistency of these commercial, operating, adoption and counter-signals rather than from any one statistic.
This chart, included in our conversational AI market deck, shows how AI chatbot platform technology has evolved over time
Related blog posts
- How large is the conversational AI market really?
