Morgan Stanley Says AI Agents Could Boost Call-Center Firms, Not Threaten Them
Artificial intelligence agents are increasingly being viewed as a threat to traditional call-center businesses, but Morgan Stanley sees a different possibility: AI could increase the number of interactions businesses handle and create new revenue opportunities for communications and customer-service software providers.
In a September 23 research note, Morgan Stanley analyst Elizabeth Porter argued that consumer AI agents such as Meta's Muse and Instinct could benefit companies that have been considered vulnerable to automation. The reasoning is that consumer AI may make it easier for people to contact businesses, while enterprise AI makes it cheaper for companies to handle those interactions.
That distinction could reshape the investment case for companies including Twilio, NICE, Five9 and RingCentral, which provide communications, contact-center or customer-engagement technology.
Why AI agents were initially seen as a threat
The traditional concern is straightforward.
If an AI assistant can call a restaurant, resolve a billing issue, schedule an appointment or complete a customer-service request without a person becoming involved, fewer human call-center agents may be required.
That creates a potential challenge for companies whose revenue has historically been linked to the number of human seats or agents using their platforms.
Morgan Stanley's latest analysis does not dismiss that risk. Instead, it suggests that the volume of interactions could rise enough to offset some of the reduction in human-agent demand.
The distinction is important because AI can change both sides of the customer-service equation.
Consumer AI could create more business interactions
Morgan Stanley's central argument is that consumer AI agents reduce the effort required to interact with businesses.
For example, a person who might previously have avoided making a phone call because of waiting times could delegate the task to an AI agent.
The agent could make the call, authenticate the user, ask questions, book an appointment, confirm the reservation and send a follow-up message.
That creates multiple digital interactions even if the human customer never speaks directly with a company employee.
Porter summarized the idea as: consumer AI makes it easier to ask, while enterprise AI makes it cheaper to answer.
For communications platforms, that could be significant because many of these services are connected to the underlying calls, messages, authentication and software workflows required to complete the task.
Twilio could benefit from higher interaction volumes
Morgan Stanley identified Twilio as having particularly broad exposure to this potential shift.
The company provides communications infrastructure that businesses use for messaging, voice calls and related customer-engagement services.
If an AI agent initiates more calls and messages on behalf of consumers, those interactions can create additional usage for communications infrastructure.
The examples cited by Morgan Stanley include authentication texts, reservation confirmations and follow-up communications.
This is different from a traditional seat-based model.
Instead of generating revenue primarily from the number of human employees using a customer-service platform, a usage-based model can potentially benefit from more transactions, regardless of whether a human or AI agent initiates them.
That makes the growth of AI agents potentially complementary to communications infrastructure rather than purely disruptive.
Contact-center software faces a more complicated transition
Companies such as NICE, Five9 and RingCentral occupy a different position.
Their customers can use AI to automate tasks that previously required human workers, potentially reducing the number of seats they need.
A "seat" generally refers to a licensed user or employee on a software platform. If fewer human agents are required, seat-based revenue can come under pressure.
Morgan Stanley argues that higher interaction volumes could partly offset that effect.
As AI handles more conversations, contact-center providers could increasingly charge for automated interactions or other AI-powered capabilities rather than relying only on human-agent seats.
That would represent a shift in the economics of the industry.
Instead of asking how many people work in a call center, investors may increasingly need to consider how many customer interactions the platform processes and how much revenue it generates from each interaction.
The opportunity extends beyond call centers
Morgan Stanley's thesis also reaches companies involved in websites, commerce and digital business infrastructure.
AI agents need to discover businesses, find information, make bookings and complete transactions.
That creates potential demand for systems that allow businesses to be found and interacted with programmatically.
Morgan Stanley sees Wix as better positioned than GoDaddy in this particular context, while identifying Shopify as the most direct beneficiary among e-commerce software providers.
The logic is that websites could evolve from being destinations where humans browse information into infrastructure that AI agents use to complete tasks.
That could change the value of digital storefronts.
Meta's Muse shows how the model is developing
The analysis comes as Meta expands the capabilities of its consumer AI assistant, Muse.
Reuters reported that Meta is testing a "human concierge" feature for Muse, with human contractors handling some calls made through the AI assistant. Muse can place calls to businesses, interact with people who answer and generate transcripts and summaries.
The development illustrates the direction of the technology.
Instead of simply answering questions, AI assistants are increasingly being designed to take actions on behalf of users.
That is the core development behind Morgan Stanley's thesis.
If AI agents become widespread, businesses may receive a larger number of machine-initiated interactions even as the role of human customer-service representatives changes.
This is a medium-term thesis, not an immediate earnings catalyst
Morgan Stanley itself is cautious about the timing.
Porter described the argument as a stronger medium-term bullish case, rather than a near-term catalyst, and warned against extrapolating the thesis directly into third-quarter results.
That qualification matters for investors.
AI adoption can take time because businesses need to integrate agents with customer databases, payment systems, authentication tools and internal workflows.
There are also questions around reliability, privacy, security and the ability of AI agents to complete complex tasks without human intervention.
The technology therefore needs to move from impressive demonstrations to dependable commercial usage before the potential increase in interaction volumes becomes a meaningful revenue driver.
AI could change the unit economics of customer service
The broader investment question is not simply whether AI will replace call-center employees.
It is whether AI changes the economics of customer interaction.
Suppose a company previously had to employ hundreds of workers to answer customer questions. If AI can handle routine requests at a lower cost, the company can potentially serve more customers without increasing staffing at the same rate.
At the same time, consumers may become more willing to contact businesses because their AI assistants can handle the work.
That combination could produce a larger overall number of interactions.
For software providers, the opportunity would therefore come from enabling those interactions, rather than necessarily preserving the existing number of human call-center jobs.
Risks to the AI-agent thesis
There are several uncertainties investors should consider.
Adoption speed: Consumers may be slow to delegate sensitive tasks to autonomous agents.
Reliability: AI agents still need to handle errors, ambiguous instructions and unusual customer-service situations.
Security: Agents accessing accounts and making transactions create new identity and cybersecurity risks. Morgan Stanley has separately argued that the rise of agentic AI could substantially expand demand for identity-security products because companies need to control what non-human agents can access and do.
Pricing pressure: If AI dramatically lowers the cost of communications and customer service, vendors could face competition that pushes prices down.
Seat cannibalization: Higher interaction volumes may not fully compensate contact-center platforms if human-agent seats decline faster than usage-based AI revenue grows.
These factors mean the thesis is a possible structural opportunity, not a guaranteed outcome.
What investors should watch
Investors tracking AI and call-center stocks should focus on several indicators over the next few quarters.
First, watch whether companies report rising AI-related usage revenue rather than simply announcing new AI features.
Second, monitor whether customer-service platforms successfully transition from seat-based pricing toward usage- or interaction-based models.
Third, look for evidence that consumer AI agents are actually completing commercial tasks at scale.
Finally, watch companies' margins. If AI lowers the cost of serving customers, the long-term benefit will depend partly on whether vendors can retain some of those efficiency gains rather than passing all of them through to customers.
Bottom line
Morgan Stanley's latest view challenges the simple idea that AI agents will automatically destroy the call-center software industry.
The bank argues that consumer AI could make it easier for people and businesses to interact, while enterprise AI makes those interactions cheaper to handle. That combination could increase overall communication volumes and create opportunities for companies such as Twilio, NICE, Five9 and RingCentral.
The thesis remains a medium-term scenario rather than an immediate earnings catalyst. The critical evidence will come from actual AI-agent adoption, interaction volumes, pricing models and revenue growth.
For investors, the key question is gradually shifting from “How many call-center jobs will AI replace?” to “How much economic activity will AI-generated interactions create, and which companies will capture that value?”
Follow the blog for more updates on AI, technology stocks, global markets and the businesses being reshaped by artificial intelligence.
This article is for informational and educational purposes only and should not be considered investment advice.