How might self-learning systems change Product Growth from a human-controlled system to an AI-driven system?
Now that artificial intelligence can produce code, text and images, a common question that is top of mind for growing companies and engineering leaders is “Can AI agents generate growth for business on its own?”
Mohit Agrawal, who has been in the industry for a long time and is now the Head of Growth Engineering at a leading fintech company, has been trying to find an answer to this question by rethinking the application of generative AI and intelligent agents to the marketing strategies based on the Product-Led Growth (PLG).
Agrawal states, “In the past, PLG was fundamentally based on the creation of value cycles — designing a great product, letting users experience it and which ultimately drives adoption. AI agents have systematically disrupted this notion. They enable self-learning within the product so it’s capable of learning, personalizing and growing with the customers on its own.”
From Static Funnels to Adaptive Systems
For more than ten years, growth teams have been carefully perfecting their funnels, campaigns, and A/B tests that aim to guide users in a very predictable manner through predetermined set of steps. Agrawal thinks that model is getting out of date quickly.
“AI is changing the very nature of growth into a self-learning system; one that keeps on learning, adapting, and improving,” he remarks. “Instead of manually deciding what to test, the smart agents can monitor user activities, come up with hypotheses, and conduct micro-experiments all the time. Growth is no longer reactive, but it has become adaptive.”
The AI systems mentioned can sort out behavioral signals in thousands and make changes to onboarding flows, recommendations, and pricing experiments without any human intervention. The end product is the personalization that occurs in real-time and which is way beyond the capacity of any human team.
Personalization Without the Playbook
Agrawal perceives generative AI as a data engine and a creative partner at the same time.
“The personalization of the future won’t be based on rules and segments anymore,” he expands. “It will be based on AI reading the user’s intent in natural language and replying in the same natural manner.”
Imagine an AI agent that is able to communicate, observe users’ browsing habits, and pick up on the overall context and then skillfully create the perfect in-app message, onboarding flow, or upsell suggestion at any moment. Not only that, but it also has the capacity to A/B test hundreds of micro-variations, figure out what resonates, and roll out the winning experience immediately.
“AI could not only create but also run thousands of mini-experiments in a single day instead of waiting for weeks to build a growth experiment,” Agrawal comments. “Hypothesis generation and validation become almost instantaneous.”
AI-Powered Upsells and Offers
Agrawal thinks that AI will not only change the content but also when and how upsells are revealed.
“Upsells should not be perceived as selling – rather, they should be perceived as a kind of predicting your next step,” he states.
The intelligent agent could recognize when a customer’s action, financial status, or position in the life cycle signals willingness for an upgrade, and then it would automatically select the appropriate offer in terms of tone, timing, and channel.
Instead of displaying the same prompt to each user, the system could come up with personalized value propositions like: “You have reached your savings milestone – this is how to speed it up even more.” Each message would be customized to each person’s objectives and AI would compose it instantly.
Generative AI and Paid Marketing
AI’s presence will not just be limited to the product. Agrawal foresees paid marketing turning into a completely self-optimizing process — using generative models that are constantly creating and testing different versions of creatives.
“AI will connect what occurs in the product and what occurs in the paid acquisition channel,” he said. “It will create new ads, personalize them for each group of target audience and modify campaigns depending on the real-time performance”
This implies that ad copy, landing pages, and visuals could be produced and improved through AI systems allocating budget in different channels like Google, Meta, and TikTok based on the conversion data returned from the product itself.
In Agrawal’s opinion, this technology revolutionizes marketing by turning it from a handmade operation into a responsive system of smart agents — each one extracting knowledge from data, adjusting bids and improving creatives with very little human intervention.
AI Agents Running Website Experiments
According to Agrawal, the same intelligence that is already affecting marketing and in-product personalization will very soon be able to perform similar tasks for web experiments and brand optimization as well. “AI agents will be testing non-stop the way your brand communicates,” he explains. “Layouts, tones, imagery, and messaging — all will be done in your company’s authentic voice. The AI agents will not stick to the conventional A/B Testing with just two variants, but they will come up with hundreds of visual and copy versions that would match the brand’s tone and values. They will know which stories appeal to which audience — whether they are professionals, investors, or first-time users — and will make it all better starting from the homepage headlines to the product visuals.”
“Your site will turn into a self improving surface,” Agrawal points out. “It will do so by constantly refining and testing new ideas, positioning, messaging, layout, and in creative ways every hour.”
Eventually, these machines will assume the role of brand consultants, giving the power of AI to communicate brand positioning, clarity and ultimately conversion while preserving brand identity and integrity
The Future: Self-Learning Growth Systems
According to Agrawal, the future of growth engineering will not be dependent on dashboards or static campaigns. Rather, it would be a self-learning ecosystem, an AI-driven system that is always observing, experimenting, and adapting at every customer’s unique touchpoint.
“Growth will no longer be a quarterly plan but rather a continuous evolution,” he claims. “The growth leader’s responsibility will transition from control to coaching — from conducting experiments to training intelligent systems to drive growth strategy.”
Future interactions will see the automatic generation of content, offers, marketing creatives, and brand messaging; the automatic testing of hypotheses; and AI agents working together across different channels to achieve significant outcomes — not only clicks, but also greater customer value.
While businesses are trying to see how AI can be used along with Product-Led Growth, people like Mohit Agrawal are already thinking about laying the necessary foundation for what might be a major change in the way products — and brands — grow their market share.
This new age will not only see growth being engineered, but also it being learned, co-created, and continuously reimagined by AI agents.
