Develop faster, automate more, break down silos and give technology teams greater autonomy: long before the arrival of generative artificial intelligence, these goals were already at the heart of DevOps culture.
AI is dramatically expanding what is possible. From modernizing legacy systems and automating security controls to deploying AI agents and enabling AI-assisted development, technology teams now have capabilities that would have been difficult to imagine only a few years ago.
To understand what is truly changing on the ground, we spoke with Benjamin Lallement and Nicolas Duperré, two of the five co-founders of Montréal-based company Gologic.
Together, they reflect on Gologic’s evolution and share their vision for the future of DevOps, one in which humans and AI move forward together.
Can you tell us about Gologic’s story and the vision that inspired its creation?
Benjamin: Gologic was founded on a simple belief: technology teams can deliver better products when they have the right practices, the right tools and, above all, the ability to become more autonomous.
Since the company was founded in Montréal in 2011, our mission has been to share software delivery best practices so that IT teams can collaborate more effectively, innovate and create greater value.
Nicolas: This vision is not simply about implementing new tools. It also requires supporting teams, transferring knowledge and enabling people to understand the technologies they use.
That philosophy remains at the heart of Gologic today. Our experts do more than recommend solutions: they design, integrate and implement them alongside the teams that will ultimately use them.
How would you define the “DevOps culture”?
Nicolas: DevOps is first and foremost a culture focused on breaking down silos. For a long time, software development followed a much more sequential model. An organization would identify a need, prepare a set of requirements, secure a budget, assemble a team to complete the project and then hand the result over to other teams responsible for infrastructure or operations.
Once the project was completed, the team could be dissolved, and much of the knowledge it had accumulated would disappear with it.
DevOps proposes almost the opposite approach: creating autonomous teams capable of analyzing, developing, deploying and operating a product over the long term. The goal is to remove barriers and bring teams together around shared processes, tools and objectives.
DevOps is therefore a culture designed to maximize synergy and collaboration among people, processes and tools.
Gologic now uses the term “DevOpsᴬᴵ.” What does AI concretely bring to DevOps culture?
Benjamin: DevOps was already focused on automating whatever could be automated. AI takes that logic even further by making it possible to incorporate tasks that require analysis and reasoning into the process.
One particularly telling example involves the validation steps that often slow down software delivery within large organizations.
Before putting an application into production, a team may need approval from architects, security specialists or various compliance leads. These controls are essential, but their accumulation can also slow the process considerably.
AI can now perform part of this verification work. A security expert can, for example, define a set of requirements that are then built directly into an automated process, allowing certain validations to be performed systematically.
Humans are not disappearing. Experts still define the rules and give final approval. Automation instead allows their expertise to be applied more quickly and at a greater scale, making human reviews faster, more effective and higher in quality.
AI has evolved very quickly over the past few years. How has your own vision evolved?
Nicolas: The arrival of AI does not challenge the philosophy on which Gologic was built. Instead, it gives that philosophy a new dimension.
Over the past three years, we have been exploring how AI can address some of the most difficult challenges in software development. Technical debt is a good example.
As systems age, technologies become obsolete and dependencies accumulate. An organization may eventually reach a point where modernizing a particular application would involve significant costs, several years of work and a high risk that the project will not succeed.
That is where AI truly changes the equation. Modernization work that was previously extremely time-consuming, sometimes virtually impossible and often seen as unrewarding by developers can now be accelerated with AI.
The goal is not to replace human expertise. It is to amplify it through AI.
What is the most common mistake you see when an organization begins integrating AI into its technology practices?
Benjamin: One of the most common mistakes is surprisingly simple: giving employees a tool and allowing everyone to use it independently.
One developer uses an assistant, another experiments with a different approach, and each person develops their own prompts and methods.
At a small scale, that freedom encourages experimentation, of course. But in an organization with dozens or hundreds of developers, it can quickly become difficult to manage.
At Gologic, we therefore place significant emphasis on guidance and training. For the past few years, the company has organized hands-on workshops in which groups of approximately 12 developers work directly with the tools to learn how to write better prompts, manage costs and understand security considerations.
Nicolas: Another mistake we observed at the beginning of the generative AI wave was the belief that it would be enough to replace the professionals already working within a company.
Some organizations that reduced their teams at the beginning of the generative AI wave are now rehiring because they have realized that they also lost part of their ability to understand and control what they produce.
AI must first be understood as a tool for amplification, not as a substitute for expertise.
How is AI adoption affecting decisions related to architecture, cloud computing and infrastructure?
Benjamin: It first requires organizations to ask a fundamental question: where is their data going?
To perform effectively, AI tools need context. This may mean giving them access to code, documentation or various forms of internal data. But a company cannot simply “open up” all its data without understanding where it will be processed and how it will be used.
Data sovereignty is therefore becoming a much more concrete issue. This concern, already very prominent in Europe, is now gaining importance in Canada and rightly so.
For us at Gologic, one of the next major areas of focus will be helping organizations map their data, understand what is being sent to different tools and determine what should be processed locally, particularly on Canadian infrastructure.
Can you tell us about the governance challenges associated with AI? For example, how can organizations maintain control and traceability without creating so many constraints that they slow innovation?
Nicolas: This is one of the major dilemmas we are seeing among our clients.
DevOps is designed to make teams autonomous. But if every team begins developing its own agents, prompts and methods for using AI, an organization can quickly end up with a million different practices.
Organizations must therefore decide where to draw the line. Should each team be free to experiment, or should a central team be responsible for the tools, rules and certain shared prompts?
This dilemma is currently at the heart of discussions within many large companies. In just a few years, some have gone from having one person tasked with “studying AI” to establishing entire teams or departments dedicated to its integration.
The next step is to determine who sets the rules, who controls the infrastructure and who makes architectural decisions. Gologic supports organizations throughout this process using a proven methodology.
How can organizations find the right balance between delivery speed, quality, security and reliability as they automate more tasks?
Benjamin: Let’s use the phenomenon of “vibe coding” as an example. Thanks to AI, someone with limited development experience can now create a complete application in very little time. For a startup or a prototype, this capability is remarkable and can be transformative.
The problem arises when it is time to put that application into production. How will it actually be deployed? Where will the data be stored? How will access be secured? What happens if the application needs to serve one million users? How will it respond under heavy loads?
Today, AI is not yet able to commit to assuming all these responsibilities. It cannot be held accountable in the same way a human can.
Nicolas: If we compare this to manufacturing a car, it would be like building a vehicle without considering what comes next. It may appear to work perfectly in the garage, yet be unable to handle the highway or withstand a Canadian winter.
That is precisely where DevOps practices remain essential. The easier it becomes to create software, the more important it becomes to have the expertise required to ensure that it is reliable, secure and truly ready for production.
With the emergence of AI agents capable of performing multiple tasks autonomously, are we moving toward processes that can analyze themselves, detect problems and eventually propose or apply their own fixes?
Benjamin: The level of automation will continue to increase, but we remain cautious about the prospect of complete autonomy.
Nicolas: If we assign a series of tasks to 100 agents and return ten days later to review their work, we open the door to significant errors and inefficiencies. It would be like doing the same thing with 100 employees. They may have completed an enormous amount of work, but there is no guarantee that they are moving in the right direction or are aligned with your business objectives.
The value of agents will therefore also depend on an organization’s ability to define its objectives, rules, control mechanisms and the points at which human intervention remains necessary.
After more than 500 engagements and over 100 clients, what changes are you currently seeing within organizations that offer the clearest indication of the next stage of DevOps and AI?
Benjamin: The next stage will not be the disappearance of DevOps, but a profound evolution in how it operates. AI will likely make it possible to standardize and automate certain best practices much more broadly. Security controls, processes and agents could be deployed more consistently throughout the different stages of the software delivery lifecycle. But humans will remain central.
Nicolas: Every day, we meet professionals with 20 or 25 years of experience who wonder whether their knowledge still holds the same value when compared with younger developers who have mastered AI tools.
At Gologic, we have reached the opposite conclusion. When an expert learns to use AI effectively, their years of experience become a considerable advantage. They know how to ask the right questions, recognize a poor answer and understand the consequences of a decision.
Humans and AI make a powerful team.
Looking ahead, what is your hope for Canada’s AI ecosystem?
Benjamin: One of the major challenges in the coming years will be developing much stronger Canadian capabilities in infrastructure and data processing.
Canadian organizations remain highly dependent on foreign solutions. As AI gains access to increasingly sensitive and strategic data, the question of where that information is processed takes on new importance.
Canada must therefore further develop its own capabilities so that companies can process certain data locally and maintain greater control over their infrastructure.
The ecosystem must also make this transition by investing in people. The goal is not to convince developers that their expertise remains important, but to show them how AI can help them apply it in new ways.
You will be participating in ALL IN 2026. Why is it important for you to take part in the event?
Nicolas: Participating in ALL IN is a natural part of Gologic’s evolution.
For approximately three years, our teams have been working at the intersection of DevOps and artificial intelligence. ALL IN gives us an opportunity to connect directly with organizations that are now facing the same questions.
Our experience at the previous edition was particularly valuable. Gologic had the opportunity to engage with major Canadian organizations connections that helped propel the company forward.
Meet the Gologic team at ALL IN 2026
In 2026, Benjamin and Nicolas want to continue these conversations around issues that have become increasingly concrete: governance, AI agents, security, technical debt, costs, team training and data sovereignty.
At the event, participants will be able to visit Gologic at its dedicated booth, speak directly with the team about these issues and discover how the company supports organizations throughout their transformation.
In addition to Benjamin and Nicolas, Gologic’s other co-founders André Saint-Germain, Haidar Dahnoun and Denis Dallaire will also be attending.
Gologic will also host a workshop, with further details to be announced soon in the ALL IN app.
The team aims to demonstrate concretely how it supports organizations at every stage from initial planning and team training to implementation and the integration of DevOps practices adapted to the age of AI.
