Ensuring operational efficiency and making robust technology decisions requires us to closely follow the changes reshaping our ecosystem. In a recent article published in the Journal du Net (JDN), entitled “Service as a Software: Is AI Driving the Comeback of Services Companies?”, I explored this historic shift.
Fifteen years ago, investor Marc Andreessen explained why “software is eating the world”. In 2026, agentic AI is beginning to shift the balance, bringing service companies back to the forefront.
This article takes that analysis further, exploring how this hybrid model is changing the way companies manage skills, data and product development.
From Day-Rate factories to “Umbrella Consultants”: A model running out of steam
At the beginning of my engineering career, I saw consulting as a collective of professionals capable of solving complex architectural challenges.
The operational reality evolved differently.
Over time, the industry increasingly became driven by daily rates and billable capacity, with the volume of code produced often taking precedence over its actual impact on product quality. For years, this model reduced IT services companies and software development agencies to little more than pools of additional resources brought in to compensate for project delays.
This shift also gave rise to what could be described as the “umbrella consultant”.
In this model, external consultants are brought in not primarily for their expertise, but to validate decisions that have already been made by management or to provide a layer of protection if a project fails. Calling on a third party simply for reassurance can often reveal a deeper lack of internal confidence.
The industrialisation of AI is now making this model increasingly obsolete.
The fundamental change is that IT services companies are no longer limited to selling human hours. They can now provide augmented, industrialised delivery capacity.
From SaaS to Service as a Software
Recent developments in the market reinforce the declining relevance of simply supplying additional technical resources.
Economic analysis published by Les Echos, alongside Sequoia Capital’s “Services: The New Software” thesis, points towards the same shift: generative AI is changing the fundamentals of software development.
Over the past decade, Software as a Service (SaaS) became the dominant model. Companies bought software licences and then trained internal teams to extract value from those tools — often while accumulating significant technical debt.
Service as a Software changes this dynamic.
Technology leaders are no longer simply buying a tool that their teams must configure. Increasingly, they expect a functional outcome.
Agentic AI is redefining technology projects.
Discover how our collaborative engineering model has been put into practice with more than 80 French technology companies.
→ Explore Novity client case studies
The Rise of the hybrid model: Execution vs Judgement
In practice, value no longer comes solely from software on one side or human expertise on the other. It comes from combining the two, supported by business data and insights from real customer usage.
One sign of this convergence is that OpenAI and Anthropic are moving into consulting, as highlighted by LeMagIT.
Julien Bek, a partner at Sequoia Capital since 2023, summarises this dynamic with the idea that the next trillion-dollar company will be a software company that looks like a services business.
Automation does not eliminate the need for engineers. Instead, it changes where their value lies.
Technical work is increasingly divided into two distinct areas:
- Technical execution: Writing standard code, generating unit tests and fixing common issues can increasingly be handled by autonomous agents and AI-assisted development tools. The engineer is no longer simply writing code; they are also directing and supervising these digital counterparts.
- Strategic judgement: This is where services and software engineering converge. Companies that use AI not merely as a productivity tool, but as a way to rethink their value proposition, can fundamentally transform their business model. Engineering effort increasingly shifts towards orchestration, architecture consulting, AI governance, data management, cybersecurity and business integration.
🔎 Checklist: Structuring your tech team around “Service as a Software”
- Augmented capacity: Are your current technology partners billing you primarily for time spent, or providing industrialised delivery capacity?
- Business and data alignment: Are your external developers working with the business data and real customer usage patterns needed to understand your product?
- Governance and security: Does your technology partner have the expertise to orchestrate autonomous agents while protecting your infrastructure?
The Novity approach: Collaborative engineering in practice
We created Novity in 2019 specifically to address the limitations of on-demand outsourcing.
Our approach is built around a simple principle: technical collaboration should never amount to delegating work to interchangeable resources.
Instead, our methodology focuses on shared responsibility:
- Challenge the requirement: Every project begins with an understanding of your environment and the areas where acceleration can create value, allowing us to identify the right expertise for the challenge.
- Combine engineering and AI capabilities: Our engineers integrate AI tools into their workflows to automate production tasks and spend more of their time on improvement, problem-solving and architectural decisions.
- Build for long-term scalability: The objective is to make delivery more reliable while documenting and transferring knowledge to create a sustainable technical foundation.
Sources
- Les Echos, Charlie Perreau - « Pour une fois, on est sexy » : l'IA ou la revanche des entreprises de services informatiques
- Silicon.fr - Les services, nouvel eldorado de l’IA
- LeMagIT - OpenAI se lance dans le consulting, Anthropic aussi
- JDN, Tribune de Xavier Lernould - Service as a Software : l'IA signe-t-elle la revanche des sociétés de services ?
