Skip to main content
Welcome to Deloitte

If we have selected the wrong experience for you, please change it above.

AI disruption in software markets

What drives defensibility of software businesses in the face of disruption from AI

For decades, the software investment thesis was relatively straightforward. Strong growth, recurring revenue, and high margins supported premium valuations, while systems of record, workflow enablement, and switching costs underpinned strategic value and transaction multiples. AI has broken that script, much like prior technology transitions such as the internet and cloud, but at a faster pace and with broader impact across software workflows and value creation.

Key takeaways

  • Software market recovery: Market has recovered significantly from ‘SaaS-pocalypse’ lows as investors have come to appreciate that moat exists for some established software providers.
  • AI risk to software providers varies: Not all software businesses face the same level of disruption from AI.
  • The AI defensibility framework: Evaluating three critical lenses—the software’s functionality, control points, and commercial model resilience—can assess defensibility against AI uncertainty.
  • Strategic responses: Software businesses can respond to the threat of AI encroachment by establishing entrenched workflows, embedding AI, expanding proprietary data/context, and developing resilient commercial models.

 

How is AI disrupting software defensibility and M&A strategy?

Many investors and strategic buyers are reassessing which products represent true control points, which moats will hold, and how quickly disruption could pressure revenue, valuations, and exit outcomes. What once looked like a steady growth sector now demands sharper differentiation in both investment and M&A strategy. 

Evaluating AI disruption requires considering management capabilities, technology infrastructure, market dynamics, customer adoption, and other factors. At the center of the analysis, however, is the software itself: where it creates value, what control points it owns, and how resiliently it captures that value. 

This point of view offers considerations for assessing where software businesses are most exposed, what remains defensible, and how the evolution of AI from assistants to autonomous agents could impact established software providers. It also provides a perspective on how corporate leaders, investment sponsors, and acquirers can reposition portfolios, deploy capital, and pursue M&A in response.

How is AI reshaping software valuations and the investor durability standard?

AI is reshaping how investors value software, pressuring multiples as the market debates which revenue streams will endure. Many investors are applying a new durability standard that favors AI-resilient monetization models (e.g., usage or outcome-based), defensible and entrenched solutions, and sustained workflow relevance over feature-level AI enhancements alone.

Several trends are driving this reset:

  • Declining software creation costs are reducing the intrinsic value of code. AI is reducing the time and cost required to build software, making code, features, and roadmap velocity less inherently differentiated as more of the application layer becomes replicable.
  • AI-native entrants are increasing competitive intensity. Lower development barriers allow startups and internal teams to launch faster, reopening categories once viewed as stable.
  • The moats that matter are being re-sorted. Investors are placing greater value on proprietary data and context, trust, systems of record, workflow ownership, and execution control.
  • Enterprise buyers are reevaluating build versus buy. As AI makes it easier to build or extend workflows internally, buyers are scrutinizing renewal value and vendor stickiness, particularly in narrower or non-core use cases.
  • Investors are repricing software against a new durability standard. Investors are compressing multiples and reassessing which software businesses can sustain their competitive position as AI lowers barriers to entry and accelerates product development. As a result, greater value is being placed on sticky customer relationships, proprietary data/context, and other unique advantages that are difficult to replicate.
     

How is AI uncertainty reshaping software M&A valuation and deal activity?

Software deal activity has slowed sharply as AI-driven uncertainty reshapes valuation frameworks and widens bid-ask spreads. Buyers are being forced to price based on a wider range of potential disruptive outcomes, including product displacement, variability in growth outlooks, and pricing pressure. Sellers, meanwhile, are often reluctant to accept discounts tied to AI exposure, creating a pricing standoff while both sides wait for clearer evidence of resilience.

At the same time, transactions are becoming more capability-driven than transformational. Acquihires and small capability tuck-ins can close AI talent or product gaps faster and with less risk than internal development, particularly where speed-to-market matters. Larger deals remain constrained by overall risk and uncertainty as the market rapidly evolves as well as general hesitancy associated with integration complexity. 

Point solutions, especially those without strong data or workflow ownership, are also being squeezed by platform convergence and consolidation. Private equity is using roll-ups to build scale, broaden scope, and strengthen defensibility. At the same time, investors are placing increased emphasis on using due diligence to assess the potential market and business disruption, opportunities for value creation through deployment of AI use cases, and business readiness for AI adoption.

Establishing a clear framework to assess software defensibility in the face of AI uncertainty, therefore, is critical to building conviction around valuation and investment theses.

Assessing AI defensibility: 10 dimensions for SaaS M&A and strategy

Deloitte has developed a framework to help organizations assess the impact of agentic AI on software-as-a-service (SaaS) companies and define potential implications for M&A and strategy. The framework moves beyond broad “SaaS-pocalypse” thinking by evaluating three lenses: the software’s functionality, the control points that make it difficult to displace, and the resilience of its commercial model. Within these three lenses, AI defensibility can be evaluated across 10 dimensions:

  1. Platform scale and scope – Broader platforms are generally more defensible because they sit across multiple workflows and integrations, making them harder for AI-native point solutions to displace. Greater scope also creates more opportunities to embed AI across the stack and capture value from adjacent use cases.
  2. System of record – Software that serves as the authoritative repository for auditable business records is one of the strongest forms of defensibility, since that software acts as a source of truth that other systems and decisions depend on. Even if AI displaces the interface, systems that hold these records often remain central.
  3. Decision criticality – Products that drive high-consequence decisions or actions are more defensible as they require accuracy, reliability, and accountability. AI may assist workflows, but replacement is harder when customers require determinism, precision, human-in-the-loop safeguards, or vendor assurance and support.
  4. Human interaction and collaboration – Software is more defensible when it facilitates human collaboration or near-real-time human interaction (e.g., communication or collaboration software) that AI can support but not fully replace. AI can often automate surrounding tasks, but the human-interaction layer remains defensible.

5. Proprietary data and unique context – Proprietary data and verticalized domain intelligence provide contextual advantages that AI models cannot reconstruct. This content moat holds regardless of architectural position if the product focuses on niche data sets, industry knowledge, or accumulated customer context.

6. Regulatory requirements and compliance – Software subject to regulatory mandates, audit frameworks, and compliance requirements (e.g., HIPAA, SOX, PCI-DSS, FedRAMP) is harder to displace, as switching can trigger recertification, revalidation, or regulatory risk. Compliance-driven switching barriers persist even when AI can replicate functionality.

7. Technical and process embeddedness – Software deeply embedded in customer systems and processes is harder to replace, as switching requires process change, technical reintegration, and operational disruption. Even when AI can replicate features, displacement is harder when the product is heavily integrated into processes and technology.

8. Breadth of ecosystem – A broad ecosystem creates defensibility when a product connects to a large network of customers, partners, suppliers, or stakeholders whose activity reinforces adoption. Ecosystems create network effects, integration depth, and adoption barriers that make displacement harder.

9. Revenue model resilience – A company’s commercial model shapes how quickly AI disruption translates into economic pressure, especially when pricing is exposed to seat compression, usage shifts, or changing buyer expectations. More resilient models align to delivered value, making them easier to defend as economics evolve.

10. Revenue diversification – Diversified revenue streams are more defensible because they reduce reliance on any single feature, workflow, or product that AI may commoditize. Companies with multiple monetization engines have more flexibility to absorb disruption in one area while preserving value elsewhere. 

Strategic response archetypes for software portfolio companies

AI response strategies should vary by disruption severity, the strength of current positioning, and time to impact. Companies with strong data, workflow, or execution moats can fortify and expand; those with a defensible base but clear exposure should adapt and defend. Businesses with an exposed core but residual strengths may need to transform and rebuild, while those with weak moats and easily replicated outputs may need to reposition and harvest. Across these archetypes, the agenda ranges from deepening workflow leadership and owning orchestration to shifting monetization, building proprietary context, and reallocating capital toward more durable control points.

How agentic AI and autopilot models shift SaaS defensibility

AI is advancing from co-pilots that assist people to coworkers that complete bounded tasks, and potentially to an autopilot model in which teams of agents plan, coordinate, and execute repeatable knowledge work under human review. As adoption progresses, defensibility will increasingly depend on where value sits across context, trust, execution, and control. 

In the co-pilot phase, AI improves productivity while core systems and data remain largely intact. As AI begins doing routine work, value shifts toward workflow orchestration, proprietary context, and trusted systems that connect insight to execution. In a more autonomous end state, more of the application layer may be commoditized as agents take over more workflows, concentrating durable value in systems of record, execution rails, governance, compliance, permissions, and accountability.

Strategic takeaways for software investors and C-suite leaders

Before the next strategic move or investment, leaders should ask:

  • How defensible is the current business based on the software’s functionality, control points, and commercial model against an accelerated AI adoption scenario?
  • Are we allocating capital toward assets/products that are defensible and provide an opportunity to deliver AI-enabled value?
  • How replaceable are the business offerings, and what can further enhance stickiness?
  • How can we adapt the offerings to build unique data, context, or embeddedness that leverages AI to enhance value and retention?
  • Are the monetization model and supporting operational capabilities modernized to focus on value delivered, or exposed to seat compression and pricing reset?
  • To what extent do the business’s systems and technology enable accelerated adoption of AI?
  • How can AI improve productivity and increase operating leverage while also enhancing customer experience?

If the market is no longer underwriting software to perpetual premium growth, the relevant question is no longer whether software is “good” or “bad,” but where value sits, how durable the control points are, and how quickly disruption translates into economic pressure. 

The leaders in this transition may not be the companies with the loudest AI narrative, but the ones that can convert AI into durable workflow ownership, trusted execution, and resilient monetization faster than the market expects.

Endnotes:

1Share price is per LSEG database, rebased to $100, and depicts 10/1/25 through 9/21/26; ProShares S&P 500 Ex-Technology ETF (SPXT) used as proxy for S&P 500 excluding technology industry; iShares Expanded Tech-Software Sector ETF (IGV) used as proxy for broader software subsector
2 Based on S&P Capital Intelligence database of all US public companies with >$1B in revenue, ‘Software Sub-sector’ is a subset of the scrape whose primary industry is ‘Application Software’
3 Transaction data per S&P Capital Intelligence screen for all US ‘Application Software’ M&A whole or partial acquisitions since 2022 through 2026Q2
4 X.AI acquisition by SpaceX on February 2, 2026 excluded from deal value visual given skewing of data ($250B)

Did you find this useful?

Thanks for your feedback