Netomi CEO Says AI Agents Could Push Customer Experience Market Toward $5 Trillion

Netomi founder and CEO Puneet Mehta expects the customer experience market to expand from about $500 billion today to $5 trillion by 2030, arguing that artificial intelligence will move the sector far beyond traditional customer support.

Mehta said companies currently spend roughly $500 billion a year on customer experience-related knowledge work. That includes support, service, internal knowledge systems and related workflows. As AI systems become more capable, he expects that spending category to grow tenfold as companies use autonomous agents for sales, conversion, upselling and cross-selling.

“Customer experience today is structured as a silo,” Mehta said. “That layer of technology and people does not fully talk to every system and every process autonomously in the company. Once that starts to happen, it unlocks a much bigger category.”

Mehta argues that AI will not simply replace support teams or automate customer service tickets. Instead, he sees AI becoming a connective layer across enterprise operations, allowing companies to link customer interactions with sales systems, product data, billing, compliance and internal workflows.

That shift, he said, could turn customer experience from a cost center into a broader revenue and automation layer.

Netomi recently raised $110 million in a Series C round backed by Accenture Ventures and Adobe Ventures, bringing its total funding to $168 million. Mehta declined to disclose the company’s valuation but said Netomi is nearing unicorn status.

The company works with major enterprise clients including Delta, United Airlines, MetLife, ESPN and ATB Financial. Mehta said Netomi is building a unified AI platform rather than a collection of disconnected tools.

“Most companies are building point solutions,” Mehta said. “They’re solving one problem at a time. We believe the future is a connected enterprise where AI systems aren’t operating in silos but working together across the entire organization.”

Mehta also rejected the argument that AI is drawing capital and attention away from crypto. He said the two sectors should be viewed as complementary because autonomous AI systems will need payment infrastructure that can operate continuously and settle transactions in real time.

“The idea that AI is simply sucking capital away from crypto is a fundamental misunderstanding of where technology is heading,” Mehta said. “We are not in a zero-sum battle for venture dollars.”

According to Mehta, the next phase of enterprise software will involve AI agents handling increasingly complex business functions, including financial transactions. That creates a problem for legacy banking systems, which are not designed for autonomous software operating around the clock.

“AI agents are moving money and assets faster than legacy enterprises can follow,” he said. “An autonomous agent cannot rely on traditional banking systems that take days to settle transactions via manual paperwork.”

Mehta said fully automated software systems require two core components: AI capable of decision-making and blockchain-based payment infrastructure capable of moving money instantly.

“To achieve true end-to-end automation, these software systems require always-on capital rails that operate 24/7,” he said.

That view aligns with a growing argument among crypto executives that AI agents could become a major source of stablecoin demand. Stablecoins and blockchain-based settlement networks are increasingly being positioned as infrastructure for real-time payments, cross-border treasury movement and machine-driven commerce.

Still, enterprise adoption remains uncertain. Many software companies continue to rely on traditional banking networks and payment providers, and it remains unclear how quickly blockchain-based settlement will become a standard part of AI-driven commerce.

For Netomi, the larger bet is that enterprise AI will move from isolated automation toward connected systems that can understand, act and transact across the business.

AI Agents Will Not Kill Crypto — They May Give Stablecoins Their Cleanest Use Case Yet

The lazy take is that AI stole crypto’s oxygen.

I get why people say it. Venture money chased AI. Founders rebranded overnight. Every pitch deck suddenly had agents, copilots and automation layers slapped on top. Crypto felt cold for a while unless you were in stablecoins, Bitcoin ETFs or infrastructure.

But Mehta’s point is sharper than that.

AI and crypto are not fighting for the same chair.

They may end up needing each other.

Because if agents actually become economic actors — not chatbots, not support macros, actual software that buys, sells, pays, books, renews, refunds and settles — then traditional payment rails start looking clunky fast.

Banks sleep.
Agents don’t.

That is the whole tension.

An AI agent handling customer experience across a global enterprise cannot wait around for manual approvals, batch processing, local banking hours and settlement delays if the business logic says money needs to move now.

That is where stablecoins start to make sense.

Not as hype tokens.
Not as yield farms.
Not as another “future of finance” slogan.

As boring payment rails for software.

And boring is good here.

The $5 trillion customer experience forecast sounds aggressive. Very aggressive. A tenfold jump from $500 billion by 2030 is not some tiny TAM expansion. It assumes customer experience stops being a department and becomes a layer across the company.

Support becomes sales.
Sales becomes retention.
Retention becomes billing.
Billing becomes workflow.
Workflow becomes automated decision-making.

Messy, but believable.

Because the old customer experience model is broken. Anyone who has dealt with enterprise support knows it. One team has the ticket. Another team has the billing data. Another has the product logs. Another has account history. Nobody sees the full picture.

AI agents are supposed to fix that.

But once they start fixing it, they run into money.

Refund the customer.
Upgrade the account.
Offer a discount.
Pay a vendor.
Trigger a credit.
Settle an invoice.
Move treasury between markets.
Buy capacity.
Book logistics.

That is where the agent needs a wallet, a payment credential or some kind of scoped financial authority.

And that is where the rails matter.

I do not think every AI agent will run on stablecoins. That is nonsense. Cards, bank transfers and existing payment processors will stay deeply embedded, especially in regulated enterprise environments.

But for 24/7, cross-border, software-native movement of value, stablecoins have a cleaner pitch than almost anything else in crypto.

They already work the way agents need money to work.

Always on.
Programmable.
Global.
Fast.
Readable by software.
Not tied to branch hours.

That is why Mehta’s argument lands.

The strongest crypto use cases are usually the ones that do not require people to believe in crypto ideology. Stablecoins won because people wanted digital dollars that move easily. Tokenized treasuries grew because people wanted yield-bearing dollar assets onchain. Bitcoin ETFs worked because investors wanted exposure through familiar wrappers.

AI-agent payments could follow the same pattern.

Not “buy this token because decentralization.”
More like: “This agent needs to settle value instantly across borders, and stablecoins are the least awkward tool.”

That is a much better pitch.

Netomi’s own story also matters here. This is not a tiny lab talking its book. The company just raised $110 million, has $168 million in total funding, and works with names like Delta, United Airlines, MetLife, ESPN and ATB Financial.

That gives Mehta’s view more weight.

He is not describing a crypto-native fantasy world where every enterprise suddenly becomes a DAO. He is describing enterprise software getting more autonomous, then hitting a payments bottleneck.

That bottleneck is real.

But I would still be careful.

Enterprise AI moves slower than Twitter thinks. Big companies do not just hand agents the keys to financial workflows because a demo looked slick. They will test, sandbox, restrict, audit, delay and lawyer everything.

Good.

They should.

Autonomous agents moving money can become a disaster if the permission layer is weak. One bad instruction chain, one compromised workflow, one hallucinated approval, one vendor spoof, and suddenly the “agentic economy” becomes a support nightmare.

So the real opportunity is not just stablecoins.

It is controlled stablecoin usage.

Limits.
Approvals.
Audit trails.
Identity.
Policy engines.
Fraud monitoring.
Reversibility where needed.
Compliance wrappers.
Human override.

That is not as sexy as “AI agents will transact globally.”

But it is what enterprises will actually buy.

Mehta’s customer experience framing is interesting because CX is where the first serious agent workflows may show up. Not because customer support is glamorous. It is not. But because it sits at the intersection of customer intent, company data and revenue leakage.

A support agent that can answer a question is useful.

A support agent that can solve the issue, issue the credit, update the subscription, trigger fulfillment and prevent churn is much more valuable.

That is the jump from cost center to revenue layer.

And if that happens across large enterprises, the market size can expand quickly.

Still, $5 trillion by 2030 is a big number. I would not treat it as destiny. I would treat it as a direction-of-travel claim: customer experience is becoming broader, more automated and more tied to revenue.

The crypto angle depends on whether agents are allowed to transact outside closed payment systems.

That is the fight.

If AI agents mostly operate inside existing SaaS billing stacks and card networks, stablecoins get some edge use cases but not the main flow.

If agents start coordinating across companies, regions, platforms and marketplaces, stablecoins become much more interesting.

Because then you need neutral settlement rails.

That is where banks are weakest.

This is also why the “AI is draining crypto” narrative feels shallow. Capital may rotate between sectors in the short term. Fine. Traders care about that. Funds care about that. But structurally, AI creates more need for automated settlement, not less.

More agents means more machine-initiated transactions.

More machine-initiated transactions means more demand for payment systems that software can use cleanly.

Crypto has been waiting for a user experience problem big enough to justify its rails.

AI agents might be one.

Might.

Not guaranteed.

The winners will probably not be random agent tokens or “AI plus blockchain” wrappers with no reason to exist. We have seen that movie. Most of it is garbage.

The real winners are more likely to be stablecoin issuers, wallet infrastructure, compliance layers, payment orchestration platforms and companies that make agent permissions safe enough for enterprises.

Less moonshot.
More plumbing.

That is where the money usually ends up anyway.

My read: Mehta is right on the direction, but the timeline may be messier than the headline suggests.

The customer experience market can expand massively if AI moves from answering tickets to executing workflows. And if agents execute workflows, they will eventually need payment rails that match software speed.

Stablecoins fit that job better than most crypto products.

But enterprise adoption will not be a straight line. It will be slow, restricted and full of guardrails until someone proves the fraud and compliance risks are manageable.

The only move that makes sense is to watch where agents get permission to move money first.

Small refunds.
Usage credits.
Subscription changes.
Vendor payments.
Cross-border treasury flows.
Marketplace purchases.
Software-to-software transactions.

That is where this becomes real.

Not in the pitch deck.

In the payment logs.