AI is reshaping enterprise SaaS, but service still drives value

At a glance

What this is: C1.ai argues that enterprise SaaS is being reshaped by AI, but the durable value still comes from service, platform depth, expertise, and machine-ready integration paths such as MCP.

Why it matters: This matters because as AI agents interact more directly with enterprise software, IAM and PAM teams have to govern authorization, access scope, and trust boundaries for both people and non-human actors.

👉 Read C1.ai's analysis of how AI is reshaping enterprise SaaS value

Context

Enterprise SaaS is under pressure from AI because code generation is getting cheaper, but code alone does not deliver a running service. The real governance gap is that enterprises still need software to be operated, secured, supported, integrated, and authorized correctly across human and machine users.

For identity teams, the interesting shift is not the death of SaaS but the change in how software is consumed. As MCP-style machine interfaces grow, the control problem moves from screen-based user access toward agent access, delegated authority, and the boundaries between domain logic and authorization.

That means the core question is no longer whether AI can produce an application. It is whether the surrounding control plane can safely manage who and what is allowed to act inside it, which is why this theme is relevant to both IAM and non-human identity governance.

Key questions

Q: How should teams govern SaaS access when bots and AI agents are also active?

A: Treat bots and AI agents as governed identities, not exceptions inside the SaaS stack. Assign ownership, lifecycle rules, and revocation paths for each non-human actor, then make sure review workflows can surface them separately from employee access. If the platform cannot distinguish those actors, governance is incomplete.

Q: Why does MCP create new authorization risk for enterprise software?

A: Because MCP turns tool access into machine-readable delegation, which can expand the number of systems an agent can reach without a person actively navigating each step. If those permissions are broad or inherited, the agent becomes a powerful execution path. Security teams need to design for action scope, not only for authentication events.

Q: What do security teams get wrong about using AI for SaaS security analysis?

A: A common mistake is assuming AI output can be trusted without context. In practice, security teams still need controls that reduce hallucination risk, such as narrow use cases, current source data, and human review. AI is most useful when it explains and organises known configuration evidence rather than inventing conclusions or replacing judgement.

Q: What is the difference between human and machine access governance?

A: Human governance relies on periodic judgment by managers and administrators. Machine governance requires telemetry, ownership mapping, and automated enforcement because access changes too fast for manual review. The key difference is that machine access must be governed continuously, not episodically.

Technical breakdown

Why software generation does not equal service delivery

AI can produce application code quickly, but enterprise SaaS value depends on the operating layer around the code. That includes deployment, patching, support, scaling, observability, integration management, and business logic that reflects how a domain actually works. In practice, the control problem is not the existence of software. It is whether the software can be governed across its lifecycle once it is in production and connected to business systems. This is where platform depth matters more than a one-off feature build.

Practical implication: treat AI-built functionality as a lifecycle governance problem, not just a development output.

What MCP changes for machine access and authorization

Model Context Protocol creates a structured way for AI agents to interact with tools and data sources. That is useful, but it also means the product surface is no longer just a human interface. Every agent-to-tool interaction becomes an authorization decision, with scope, delegation, and session boundaries that need to be explicit. If MCP is implemented without tight access modeling, the risk shifts from user misuse to machine misuse, where an agent can reach more systems than a human operator would ever navigate manually.

Practical implication: map each MCP-connected action to a defined entitlement, not to a broad user role.

Why platformization raises the stakes for non-human identity governance

Platformization concentrates data, business logic, and automation into a smaller number of products. That can improve value creation, but it also concentrates identity risk because one system may now host multiple agentic workflows, service identities, and integration credentials. In an AI-heavy SaaS model, the trust boundary is not just authentication at login. It is whether each non-human actor has a constrained purpose, a narrow permission scope, and a clear offboarding path when the workflow changes. That is a classic NHI governance challenge, just expressed through AI-native software.

Practical implication: inventory non-human actors inside platforms before their permissions become embedded in business workflows.

NHI Mgmt Group analysis

Platformization is becoming an identity control problem, not just a product strategy. As enterprise SaaS concentrates more business logic and automation, the number of identities that matter inside the platform expands. That includes users, service accounts, integrations, and AI agents. The practical implication is that teams need governance over the full access graph, not only over human login events.

Machine-facing UX creates a new authorization surface. Native MCP connectivity is not just a usability feature for agents. It is a transport for delegated action, which means the product boundary now includes machine-readable permissions, not only screen-based workflows. The field needs to treat agent access as first-class identity design, or AI will inherit permissions that were never intended for autonomous or semi-autonomous use.

Data gravity and domain expertise now determine security resilience. The more value a platform holds in business logic and customer context, the more damaging mis-scoped access becomes. This is especially important for NHI governance because access drift in integrated systems is easier to hide than user-role drift. Practitioners should assume that the control plane, not the interface, is where SaaS differentiation increasingly turns into security exposure.

The assumption that human review catches every meaningful access decision is weakening. AI-assisted software and agentic interfaces reduce how often a person directly touches the application. That does not eliminate governance, but it does reduce the effectiveness of reviews that depend on visible human interaction. Teams should shift attention toward entitlement design, delegated authority, and lifecycle controls that bind access to purpose rather than to convenience.

Enterprise SaaS will be judged by how safely it handles non-human work. The next differentiation layer is not only feature velocity but whether the vendor can support machine access without expanding blast radius. That makes NHI controls, scoped integration permissions, and clear separation between human and agent actions central to procurement and architecture decisions.

From our research library:

• The average enterprise SaaS platform connects to 42 or more third-party applications through OAuth tokens, API keys, webhooks and automation platforms.

What this signals

Machine-facing SaaS is pushing identity governance toward the control plane. When software is built and operated with AI in the loop, the main question becomes which identities can invoke which business functions. That is an IAM and NHI issue as much as a product-design issue, because access scope becomes part of the product architecture.

Platformization concentrates both value and risk. A SaaS platform that combines data gravity, domain logic, and automation can reduce tool sprawl, but it also creates a larger blast radius if entitlements are too broad. Security teams should expect procurement, architecture, and identity governance to converge around the same control questions.

MCP-style interfaces make agent governance unavoidable. The moment an AI agent can call tools directly, teams need clear rules for delegation, revocation, logging, and separation of duties. If those controls are missing, the organisation is effectively treating a machine actor like an ungoverned admin session.

For practitioners

• Define MCP entitlement boundaries Map every agent-to-tool path to a discrete entitlement, with least-privilege scopes for read, write, and execute actions. Do not let agent connectivity inherit broad application roles by default.
• Inventory non-human identities inside SaaS platforms Identify service accounts, automation tokens, API keys, and agent identities that operate inside core SaaS platforms. Record ownership, purpose, and offboarding triggers before access becomes embedded in workflows.
• Separate human and machine actions in control design Treat human login access and machine execution access as different control problems. Use distinct approval paths, logging, and authorization logic for AI agents and service identities.
• Review platformization for access sprawl Check whether product expansion is creating hidden privilege growth across integrations, add-ons, and internal automations. Prioritise the areas where one platform now concentrates multiple business processes and identities.

Key takeaways

• AI is not eliminating enterprise SaaS value, but it is shifting where value is created and where risk concentrates, especially around service delivery, platform depth, and domain expertise.
• The governance challenge moves toward authorization boundaries for machine interactions, because AI agents and integrations can act inside software in ways human users never do.
• For practitioners, the priority is to inventory non-human identities, tighten delegated access, and separate human approval from machine execution before AI-native workflows become entrenched.

Key terms

• Model Context Protocol: Model Context Protocol is an open protocol that lets AI agents connect to tools and data sources. It expands what an agent can reach, so governance has to cover not only the model and its prompts, but also every system that can receive or return agent-driven data.
• Non-Human Identity (NHI): A digital identity assigned to a non-human entity such as a software application, service account, API key, bot, machine, or AI agent that enables it to authenticate and interact with systems without direct human involvement. NHIs now outnumber human identities in most enterprises by 25 to 50 times.
• Platformization: The process of expanding a product into a broader integrated platform with shared data, logic, and workflows. In security terms, platformization concentrates authority and integrations, which can improve visibility but also increase the blast radius of a compromised credential or over-scoped role.
• Agentic Interface: An agentic interface is a product interaction model in which software agents can call tools and perform tasks directly rather than only presenting information to a human. The security challenge is to define what the agent may do, under which conditions, and with what revocation path when its task or trust context changes.

What's in the full article

C1.ai's full blog post covers the operational detail this post intentionally leaves for the source:

• The author’s deeper discussion of why service, support, and domain expertise still determine enterprise SaaS value.
• More detail on platformization and why compound products outlast point solutions in AI-heavy markets.
• The article’s own framing of MCP as the machine-facing UX layer for AI agents.
• Additional commentary on how AI changes feature velocity, customer experience, and product strategy.

👉 C1.ai's full post expands on platformization, MCP-driven UX, and the service layer behind SaaS value.

Deepen your knowledge

NHI Foundation Level course, the industry's only accredited NHI security programme, covers NHI governance, machine identity security, and secrets management. It helps practitioners design controls for the access models that modern platforms and AI-driven workflows depend on.