Laurent Giraid is a technologist whose work sits at the cutting edge of Artificial Intelligence, specifically focusing on how machine learning and ethical design influence our daily digital interactions. With a deep background in natural language processing, Laurent has spent years dissecting the subtle ways algorithms shape social behavior and content consumption. Today, he helps us understand the radical transformation of Instagram, a platform where every pixel and interaction is now filtered through an AI lens. This conversation provides a rare look at how the machinery behind our screens is evolving to prioritize the one thing it cannot actually produce: genuine human connection.
The following discussion explores the tension between algorithmic efficiency and authentic creativity, highlighting how Instagram utilizes distinct systems for its various features like Reels and the Feed. We delve into the shifting metrics of success, where the speed of initial engagement and the trust inherent in a private share outweigh traditional likes. The conversation also covers the practical side of automation, showing how AI handles the heavy lifting of content production and customer service, while ultimately leaving the most complex emotional work to human experts.
Instagram’s ranking systems for Reels and the Explore page are now entirely driven by AI—how does this shift affect what users actually see on their screens every day?
It is a massive technical undertaking because we are talking about a platform used by over three billion people, where every single suggestion is now tailored by a sophisticated AI system. This means the app no longer operates on a single, monolithic algorithm; instead, it utilizes a specialized system for the Feed, another for Reels, one for Stories, and a fourth for the Explore page. Each of these predicts what you will enjoy based on your unique past behavior, which explains why two friends sitting on the same couch will see two completely different worlds when they open the app at the same moment. While AI controls the mechanics of this distribution, it is actually being programmed to protect the human element rather than replace it. Adam Mosseri has made it clear that the system is being tuned to prioritize real, human-made content over AI-generated filler, creating a digital environment that values the weight of a person’s actual voice and vision.
With engagement rates hovering around 0.7% in 2026, what specific metrics are the most important for creators who want to break through the noise?
In a landscape where the engagement rate is as low as 0.7%, you cannot afford to ignore the three primary signals that the AI uses to measure reach: watch time, likes per reach, and sends per reach. Of these, the “send” is perhaps the most powerful because when a user shares a post with a friend, it carries a level of trust and personal recommendation that an algorithm simply cannot fake. There is also the critical factor of engagement velocity, which tracks how a piece of content performs in its first 30 minutes of life. A Reel that catches fire immediately will almost always outrank a post that has been growing slowly for three days, as the system is designed to reward content that sparks an immediate, genuine reaction. By focusing on these “trust points,” creators can signal to the AI that their content has real value, helping it bypass the filters that are constantly scanning for fake followers or bot-driven activity.
How is AI automation changing the daily workflow for brands and creators who used to spend hours on repetitive production tasks?
We have moved away from the era where growing an account meant losing dozens of hours every week to the “busywork” of writing captions, resizing images, and planning complex schedules. Modern AI automation tools can now take a single video and repurpose it into five different formats in just a matter of minutes, allowing creators to focus on high-level strategy rather than technical grunt work. These tools are incredibly proficient at drafting initial caption versions or predicting which specific audience segment will vibrate with a particular visual style. However, there is a very clear line in the sand; Instagram’s system is now smart enough to suspend accounts that rely too heavily on reposting or low-effort automation. The goal is to use AI to handle the “how” of production—like scheduling at the best times or generating on-brand visuals—so the human brain can stay focused on the “why,” which involves story, voice, and judgment.
Direct messaging has become a primary hub for commerce and interaction, but how is AI managing the complexity of these personal conversations?
The shift toward DMs as a primary business channel has been quiet but profound, as people are increasingly choosing to message a brand rather than sending an old-fashioned email. AI agents have stepped in to meet the expectation for instant replies, and research from Harvard Business School shows that these tools help human agents respond to inquiries about 20 percent faster. You typically see two tiers here: basic rule-based bots that handle simple triggers like “order status,” and more advanced AI agents that can hold a natural conversation and resolve unusual queries using a brand’s specific data. These advanced systems are turning the platform into a legitimate marketplace by handling refunds and disputes across any time zone, day or night. Crucially, the most effective systems always include a “human handoff” feature, ensuring that when a conversation becomes emotional or high-stakes, a real person steps in to provide the empathy and rights protection that code cannot provide.
What is your forecast for the future of AI on Instagram?
I expect to see a much tighter feedback loop where AI agents move beyond just answering questions to actually resolving complex customer service issues from start to finish. At the same time, the “war on fake” will intensify, with Instagram deploying even more advanced detection systems and penalties for any account that leans too heavily on AI-generated shortcuts or coordinated engagement pods. The most successful brands won’t be the ones with the most advanced AI, but rather the ones that manage it the best, blending the efficiency of automation with a voice that is strictly aligned with their brand values. We are entering a phase where AI handles the scale of communication, but humans are the only ones who can provide the meaning that builds long-term trust. Ultimately, the algorithm is being designed to reward this exact combination: using technology to clear the path so that real human connection can happen more frequently and more effectively.
