From Snowflake to Sierra, every enterprise software company is selling the same AI agent.

The enterprise software industry is embroiled in an unprecedented mêlée, with traditional market boundaries being thoroughly disrupted by artificial intelligence.

From database giant Snowflake to CRM leader Salesforce, nearly all the major tech companies are racing to launch similar general-purpose AI agents, pushing former partners onto the stage of direct competition. This has created enormous growth opportunities, but also life-or-death challenges.

The latest development in this trend is that companies with previously distinct business domains are now penetrating each other's core markets. Salesforce, a company mainly focused on customer relationship software, has recently launched an AI agent to solve IT help desk issues. Meanwhile, IT service management software company ServiceNow has introduced an AI agent for sales personnel. Last month, database service provider Snowflake also launched its own AI agent product, claiming it can handle tasks ranging from sales to finance and other professional roles.

This "arms race" of AI agents has triggered significant confusion among enterprise clients. Due to high overlap in product functionalities, corporate buyers face "extremely difficult" choices, with some companies postponing large-scale purchasing decisions. This chaos reflects the high stakes, as Snowflake CEO Sridhar Ramaswamy said: for software companies in this transformation, "it's either a path to a trillion-dollar market cap, or zero."

This phenomenon also directly echoes the recent deep-seated market anxiety that "all software will be replaced by AI." However, a closer look at market dynamics reveals that traditional software giants are building fortifications based on their vast customer bases and accumulated data, indicating that the future industry landscape will be far more complex than a simple "disrupted-disruptor" narrative.

Market Overlap, Blurred Boundaries

The rise of AI agents is thoroughly rewriting the competitive landscape of enterprise software. At least seven major technology companies are currently engaged in head-to-head battles in eight different functional domains, selling automation AI agents for roles in engineering, analytics, finance, marketing, sales, and customer service, among others.

A partial reason for this homogenous competition is that many startups and enterprise software giants rely on foundational AI models from companies like OpenAI and Anthropic to power their agent products. Matt Luizzi, Director of Analytics at Whoop, said, "Every piece of software we use has launched its own AI agent solution." From Slack to Snowflake to Google Workspace, they all promise to handle similar tasks like mining data, predicting sales, and even communicating with customers.

This trend means that established database and data streaming companies like Snowflake and Confluent are now competing with emerging AI application startups like Sierra and Decagon in areas like sales and customer support agents. "It is an exciting yet chaotic time," commented Jay Kreps, CEO of data management company Confluent.

Software Giants' Advantage: Data Convenience and Hybrid AI Model Strategy

Faced with an influx of similar AI products, enterprise buyers generally feel at a loss. Ryan Teeples, Chief Strategy Officer at accounting company 1-800Accountant, said that the selection process has become unusually difficult due to a large amount of overlap between tools, even though his company has started paying for some of Salesforce's agent tools.

Amidst this chaos, existing software giants are leveraging their core advantage—data. They argue that agents capable of extracting data directly from their core software products (such as CRM apps or data warehouses) are more convenient for customers. This strategy appears to be helping them take the lead in winning customers for initial trials. Snowflake revealed last week that 1,000 customers have already used its agent product, Snowflake Intelligence, creating 12,000 agents.

Convenience has become key to enterprise decision-making. Pierre Matuchet, Senior VP of IT at European HR company Adecco, said their selection logic is very simple: "If the data is stored in Salesforce, we use Salesforce. If the data is outside Salesforce, then we consider other suppliers." Similarly, Peter Stoltz, CIO of tablet manufacturer reMarkable, said they chose Salesforce's Slack agent because most employees already use Slack.

Although OpenAI CEO Sam Altman has suggested that AI agents may eventually fully replace work software like Slack, analyses from Wall Street and the strategies of industry giants paint a different picture. Goldman Sachs noted in a report that the current phase resembles the transition from on-premise to cloud in the software industry—AI is more likely to be a “force multiplier” for industry leaders rather than a pure disruptor.

To build moats, enterprise software giants commonly adopt a hybrid AI model strategy. They combine domain-specific models trained on their own proprietary data (such as Snowflake’s Arctic model) with external cutting-edge large language models. This strategy locks customers into the ecosystems they are familiar with and deeply integrated, while retaining flexibility.

Moreover, the “mission-critical” nature of enterprise software creates a natural barrier. Goldman Sachs analysts note that AI model “hallucinations” can have serious consequences in enterprise settings, so customers are extremely cautious when migrating core business processes. This means that even if native AI products are technologically superior, it will be difficult for them to gain widespread customer adoption in the short term.

Slow Enterprise Adoption, Management and Coordination of AI Agents in Focus

Despite its vast prospects, the commercialization path of AI agents is far from smooth. Currently, this new technology has not produced significant revenue growth for companies like Salesforce, ServiceNow, and Microsoft. Salesforce CEO Marc Benioff has recently toned down claims about the ease of agent deployment, reflecting the industry consensus that enterprise adoption of AI is slow.

There are many reasons for slow adoption. Firstly, configuring agents may require substantial manual help, prompting companies like Amazon and Salesforce to dedicate extra staff for support. Secondly, some companies think vendor-supplied AI is not mature enough. For example, Alex Devkar, SVP of Engineering at used car dealer Carvana, said they chose to develop their own AI chatbot because it performed better than existing products on the market.

On the business model side, most vendors charge by usage after a free trial period, typically at a cost of 20 to 30 cents per task. But for companies like Snowflake, the current priority is to capture users, not immediate profitability.

As the number of AI agents used inside enterprises booms, a new challenge faces all CIOs ."Everyone must decide whether a centralized agent platform is needed to coordinate all these agents scattered across different software," said Eswar Veluri, CTO of fitness company Equinox. "We haven’t crossed that threshold yet." This issue suggests the next competitive focus in enterprise software may revolve around the management and coordination of AI agents.

Risk Warning and DisclaimerThe market carries risks; investments should be made cautiously. This article does not constitute personal investment advice, nor does it consider the specific investment objectives, financial circumstances, or needs of individual users. Users should consider whether any opinions, views, or conclusions within this article suit their specific circumstances. Act on this information at your own risk.