Replacing ServiceNow in 45 Days: How AI Could Erode the SaaS Moat
Serval says it swapped out Lucid Motors' ServiceNow system in 45 days. One case settles nothing, but it questions the idea that enterprise software is too hard to leave.
1. What if switching really takes 45 days?
Companies rarely change enterprise software once it is installed. Employee accounts, permissions, approval workflows and records are all tied together. Switching costs money, and the whole organization has to adapt.
This structure, which makes it hard for customers to move to another product, is known as switching costs. It is a major competitive advantage for enterprise software such as ServiceNow.
Now comes a report that electric-vehicle maker Lucid Motors replaced its existing ServiceNow system with Serval in 45 days. The account comes from Serval CEO Jake Stauch, who described the customer case to Reuters.
One case does not mean switching has become easy for every large company. But it suggests that a replacement once dismissed as impossible because it took too long could become a real purchasing option.
2. Why investors valued Serval at $3 billion
On Oct. 29, Serval announced it had raised $150 million in a round led by Sequoia Capital. The private valuation was $3 billion, triple the level of December 2025.
- $150M: new funding raised
- $3B: private valuation
- 3x: valuation change versus December 2025
The market is not looking only at the ability to build a better chatbot. Serval uses AI agents to automate repetitive work inside companies, such as handling IT requests, managing employee account permissions and managing software assets.
The strategy is aggressive. Rather than bolting AI features onto an existing system, Serval aims to replace work platforms like ServiceNow altogether.
3. The question in software competition is changing
Established SaaS companies have grown by giving customers more features. The more features, the more deeply a company depended on the system.
AI agents pose a different question. Suppose an employee asks, "Set up the account permissions a new hire needs." The test is not how many screens that takes to click through. It is whether the work finishes, through approval and permission grants.
| How many features does it have? | How well does it finish the job? |
|---|---|
| Work screens, admin tools, links to existing systems, long contract ties | Resolving requests, handling permissions, automating approvals, saving real employee time |
That does not erase the value of work records and customer data. For AI to handle a task, it needs to know who can approve what and which procedures to follow. Incumbents have built up that information over many years.
In the end, the contest may turn not on whether a company has AI features, but on how safely and quickly it can produce results using existing work data and processes.
4. Signals to watch
- Platform replacements: Do replacement cases by newcomers like Serval spread to other large companies?
- Real task performance: How much do AI agents cut human involvement in resolving requests and handling account permissions?
- Retention and renewals: Do existing SaaS customers renew because of AI features, or move to competitors?
- AI monetization and cost savings: As AI feature sales grow, do customers also see lower operating costs?
Even if share prices move first, there is little reason to rush to a conclusion about a company's competitiveness unless these indicators change.
Insight Times Editorial Desk
