What No One Is Telling You About the Future of IT Services Firms Facing SaaS and AI in 2026

In 2026, the IT services SaaS reseller model is condemned by a structural double threat: vendors are building out their direct channels and squeezing resale margins from below, while autonomous AI agents challenge the per-seat licensing model from above. The future of SaaS clients for IT services firms is a topic everyone discusses at sales kickoffs... and nobody truly confronts. In 2025, the French IT services market shrank 1.8% to 34.3 billion euros, a first in fifteen years, according to Numeum data relayed by Karanext. We optimize sales dashboards, track renewal rates, chase MRR. Meanwhile, two bombs are being assembled in silence. On one side, an aging application estate crumbling inside your clients' IT systems, blocking any AI integration. On the other, autonomous AI agents that are starting to navigate your SaaS solutions in place of human users, calling into question the very licensing model your resale margin depends on.

The SaaS reseller model: a shrinking role

The VAR/SaaS reseller model is commoditizing because vendors have built out their direct channels, and clients now compare, purchase, and deploy on their own, without an intermediary. For ten years, IT services firms built a solid business on reselling SaaS solutions. Security, collaboration, cloud ERP, ITSM: margins were decent, client relationships relatively stable, and the sales cycle well-oiled.

That era is running out. Major vendors are expanding their direct channels, resale prices are eroding, and the role of simple reseller is commoditizing at speed. A client can compare, buy, and deploy solo in a few clicks. What was worth gold five years ago is becoming an interchangeable service where price is the only variable. Meanwhile, 77% of new IT projects were delivered as SaaS in the second half of 2025, up from 53% in 2021, and software vendor revenue grew by +8.2% over the same period. The money is shifting. It is not disappearing.

Cris Réseaux saw it coming. A cybersecurity specialist positioned as a VAR, the company publicly announced its pivot toward an MSSD (Managed Security Service Distributor) model, capable of operating across the entire value chain: pre-sales, deployment, training, support, and managed services. The stated objective is clear: give their partners the means to act with greater control and confidence in the face of their clients' cyber challenges.

This is not a sector anecdote. It is the signal of a repositioning every IT services firm will have to negotiate.

What makes this pivot essential is not just competitive pressure. It is what is actually happening inside your clients' IT environments.

The silent bomb inside your clients' IT systems

Most of your clients' critical applications are over ten years old. Without exposed APIs or modular architectures, they block any AI integration, making your SaaS sales cycles structurally incomplete because the IT environment is not ready to absorb what you are selling.

Here is what nobody says clearly during SaaS sales cycles: your clients have an aging application estate that makes a significant share of integrations either impossible or far too expensive.

The invisible technical debt: 10% of the IT budget spent for nothing

Isimédia, a French IT services firm specializing in business applications since 1996, documents this reality with precise figures. According to a Gartner benchmark cited by Isimédia's leadership in their repositioning interview, 10% of IT department budgets goes toward maintaining legacy applications, and within those maintenance costs, 20 to 40% corresponds solely to managing technical debt: the layers of code accumulated without ever being refactored, expensive to maintain and blocking all evolution.

The structural problem runs even deeper. A critical business application stays in production 15 to 20 years on average, according to field data from application modernization specialists. A CIO stays in post 4 to 5 years. A developer, 2 to 3 years. The result: the people who built those applications left long ago. The knowledge is lost, the code has become a black box, and every change is a calculated risk.

Digital sovereignty goes beyond hosting

On top of this comes the question of digital sovereignty, which Jean-François Caplat and Jean-Marc Durand, co-directors of Isimédia, address head-on: hosting data in France is not enough. Sovereignty means being in control of your digital assets. An application whose source code nobody understands anymore, whose knowledge is no longer documented, whose maintainer left for a competitor: that organization does not have sovereignty over its information system, regardless of which servers it uses.

This is the context in which your clients call you to "add some SaaS." The real question is not which solution to buy. The real question is: can their IT environment absorb it without breaking everything?

AI: both the aggravating factor and the solution

AI is simultaneously the revealer of the legacy problem and the lever for solving it: it makes monolithic IT systems even more of a blocker for new integrations, while making their modernization more accessible, and finally cost-effective for mid-market companies. The arrival of AI in enterprises is changing the tempo of this crisis. Executive teams want use cases, CIOs are under pressure to deliver. But a monolithic, closed information system built on 2000s-era architectures cannot accommodate these new use cases.

Integrations remain impossible, APIs do not exist, data is trapped in proprietary formats. AI amplifies the legacy problem: where the aging application was merely expensive to maintain, it is now a strategic blocker. This is exactly what I observe in enterprise AI integration projects: the main obstacle is not the model, it is the IT system that is supposed to host it.

Second scissor effect: autonomous agents are attacking the licensing model. While legacy IT blocks integrations from below, AI agents are attacking the resale model from above. An agent can now navigate a CRM, ITSM, or collaboration interface and perform the tasks of multiple human users. The result: your clients are starting to wonder how many seats they are actually funding. The collective wake-up call was brutal: in February 2026, $285 billion in SaaS market cap was wiped out in 48 hours when Wall Street priced in that AI agents mechanically reduce seat counts. Atlassian, whose model relies entirely on seat expansion, reported its first absolute decline in enterprise seats in February 2026 (-36% in one month). Monday.com officially replaced 100 SDR salespeople with AI agents. In twelve months, the per-seat model dropped from 21% to 15% of enterprise contracts, while 40% of enterprise SaaS contracts now include outcome-based pricing elements (up from 15% two years earlier). According to a Channel News survey relayed by Karanext, 81% of IT services firms identify AI as their primary growth opportunity, and 64% say they already use it in at least part of their operations, but few have grasped that it is simultaneously eroding their license resale base. The paradox is brutal: vendor members of Numeum report +12.5% productivity gains in their technical teams in 2025 thanks to generative AI, while IT services firms posted a -1.8% decline. That is the most telling fracture in the sector, and the disruption that consulting firms are beginning to document: AI does not just threaten your ability to deliver, it undermines the perceived value of the product you resell. This shift, which IT analysts have dubbed the "SaaSpocalypse," the destruction of the per-seat model by autonomous agents, started weighing on the stock valuations of major vendors and integrators from early 2026.

Facing this erosion, the most responsive IT services firms are exploring a pricing model shift: moving from per-seat to value-based billing, whether a flat fee per business domain or indexing on measurable gains (productivity, reduced maintenance costs). This pivot structurally resists seat erosion because it captures the value created, not the number of users.

Isimédia, however, flips this reasoning. Their leadership states it directly: AI now makes it possible to carry out an application transformation with shorter timelines and more reasonable costs than before. Their internal R&D produced a specific methodology: extracting the knowledge buried in legacy applications (documenting the black boxes, mapping what the code actually does), then leveraging generative AI to produce clean code on a modern architecture.

AI is not just a goal for your clients. It is also the lever that finally makes their application transformation affordable.

The condition: a progressive transformation, never a "big bang." No service interruption, no business disruption. And an approach that starts with business needs, not migration for migration's sake.

VAR, MSSD, transformation partner: three paths, only one is durable

Three trajectories are open to IT services firms today, and only one builds a client relationship that lasts beyond 24 months. The market is segmenting clearly around three positioning strategies, with very different consequences for value creation and client relationship duration:

Cris Réseaux's pivot toward MSSD illustrates the second path. Three structured service tiers: cyber-diagnostic (assessing cyber maturity and identifying priority actions), cyber-deployment (implementing protective measures), and cyber-support (tailored ongoing guidance with advice and customization). These three offerings are organized around five cyber fundamentals: identify, protect, detect, respond, recover. The goal is to shift from license supplier to long-term security operator.

Isimédia illustrates the third path, a more ambitious one. Their transformation offering does not start with technical migration. It starts with identifying what the application actually needs to do, what to keep, what to add, what to remove. The technical work follows, in support of a strategic alignment built through co-design with the client. The average duration of a transformation project: three to six months, progressing in stages, never disrupting service.

The strength of this positioning rests on a genuine scarcity: mastering both the legacy architectures (in Isimédia's case, PCSoft tools), the target cloud and web architectures, and the change management process. Few players can assemble that combination.

Verdict: what this changes for your IT services firm starting in 2026

I have long believed that the same shift hitting traditional developers will hit IT services firms. Developers who simply write code will be overtaken by those who orchestrate tools, AI agents, and workflows. IT services firms that simply resell SaaS licenses will follow the same trajectory.

The real question is not whether your clients want "AI." What they want is to save time, reduce maintenance costs, and stop depending on code nobody understands anymore. AI is the means. Application transformation is the project. And business process automation cannot start on an IT system held together with duct tape.

The path is clear: either keep selling SaaS building blocks into IT systems that cannot digest them and watch margins compress year after year. Or become the partner that secures the ground first before building on it, and build a relationship that lasts years, not quarters.

A complementary signal confirms this shift: while IT services firms declined by -1.8%, Gartner forecasts +11% growth in European IT spending for 2026. The market is not collapsing, it is redistributing. Numeum is more cautious on IT services firms themselves: its December 2025 projections anticipate only +1.4% growth for French IT services firms in 2026, to 35 billion euros. The recovery will be real but uneven: firms that have pivoted toward transformation and managed services will capture it; the rest will continue to absorb the pressure on resale margins. And with it, billing models are shifting too: value-based pricing (productivity gains, reduced maintenance costs, shorter timelines) is beginning to supplant per-seat or per-day billing in commercial discussions. This is a pricing pivot as much as a strategic one, and it structurally favors the player that controls the IT system end to end.

A third signal is accelerating this urgency: according to Turnover IT, drawing on PAC-Numeum data from late 2025, large enterprises are actively reducing the number of their IT providers, consolidating their portfolio around a handful of strategic partners. For an IT services firm that has not yet made this pivot, being excluded from that "core panel" can erase years of commercial relationship in a single tender.

The future of SaaS clients for IT services firms is the future of IT services firms themselves. The companies that will have understood this in 2026 are the ones people will cite as examples five years from now.

Frequently asked questions

Why is the VAR/SaaS reseller model for IT services firms in danger in 2026?

The VAR model is commoditizing because major vendors are building out their direct channels, compressing resale margins. Clients can now compare, purchase, and deploy on their own in a few clicks. The result: the only differentiating variable becomes price, a losing spiral in the long run. The -1.8% contraction of the IT services market in 2025 is partly readable through this lens.

What is an MSSD and why is it a better path than a simple VAR?

An MSSD (Managed Security Service Distributor) goes beyond license resale to include deployment, management, monitoring, and continuous advisory. Client engagement shifts from transactional (one-time sale) to operational (12 to 36 months of recurring contract). The pivot is structured in three tiers: diagnostic, deployment, and tailored ongoing support, as formalized by Cris Réseaux in the cyber segment.

How exactly does the legacy application estate block AI integration?

Applications over 15 years old do not expose APIs, data is trapped in proprietary formats, and monolithic architectures cannot support the integration of external services. Without IT modernization, every AI project remains a pilot that never scales. This is the obstacle I encounter systematically in enterprise AI integration projects: it is not the model that blocks, it is the IT system that is supposed to host it.

How long does an AI-assisted application transformation take?

Following Isimédia's approach (extracting the business knowledge buried in the code, then regenerating on a modern architecture with generative AI) a transformation project runs 3 to 6 months, progressing in stages without service interruption. AI has significantly reduced timelines and costs compared to traditional migrations, making the project finally feasible for mid-market companies.

Do AI agents directly threaten the SaaS licensing model I resell?

This is the most underestimated risk in the sector right now. An autonomous agent can today execute tasks in a CRM, ITSM, or collaboration tool in place of human users. If your clients consolidate their seats because an agent handles part of the work, your resold license volume drops mechanically, without anyone notifying you. The answer is not to resist but to anticipate: integration, configuration, and supervision services for AI agents are precisely where margin is not indexed to seat count.

What is the real difference between SaaS reseller, MSSD, and transformation partner?

A SaaS reseller sells licenses (transactional, short-term relationship). An MSSD operationally manages the solution over time (12 to 36 months). An application transformation partner rebuilds the IT system so it can accommodate any future solution, a strategic relationship of 3 to 5 years with far higher perceived value and strong competitive differentiation. These are three different markets, not three rungs on the same ladder.

How should I adapt my billing model when AI agents are reducing the number of licenses I resell?

The answer emerging among the most advanced IT services firms is the shift from per-seat to value-based billing: a flat fee per business domain or indexing on measurable gains (reduced maintenance costs, productivity gains in equivalent hours). This model resists seat erosion because it captures the value created, not the number of users. Integration, supervision, and agent configuration services then become the new margin core, independent of resold license volume.

What is the "SaaSpocalypse" and why does it directly concern IT services resellers?

The SaaSpocalypse refers to the destruction of the per-seat pricing model by autonomous AI agents. In February 2026, the wake-up call was brutal on financial markets: $285 billion in SaaS market cap was wiped out in 48 hours when Wall Street grasped that an AI agent can do the work of multiple human users without consuming additional seats. Atlassian, whose entire business model rests on seat expansion, reported its first absolute decline in enterprise seats. For an IT services firm reselling licenses indexed to user count, this is not a distant stock market crisis: it is the signal that your clients have begun consolidating their seats, and that your resale volumes will decline mechanically, even without anyone notifying you.

Vidéos YouTube

• From VAR to MSSD: The Future of Cris Réseaux · Cris Réseaux
• Isimedia 2.0: Interview with Jean-François Caplat and Jean-Marc Durand on the Brand Relaunch · Groupe ISIA

Articles & ressources

• IT Services in 2026: Measured Recovery, AI at the Heart of Strategy · Karanext
• France's Digital Market: 2025 Review and 2026 Outlook · Global Security Mag
• Is AI Disrupting IT Services Business Models? · OCTO Talks
• The IT Services Market in 2026: Fragile Recovery, Pressure on Models · Turnover IT (PAC-Numeum data)
• The SaaSpocalypse: $285B Wiped, AI Agents Rising · Taskade Research
• Digital Market Trends 2025, 2026 Outlook · Numeum