By Tan Simin
How AI Changes the Consumer Growth Funnel
Five years working in consumer growth taught me a fairly simple lesson: growth is not primarily an execution problem. It is a strategy problem.
Five years working in consumer growth taught me a fairly simple lesson: growth is not primarily an execution problem. It is a strategy problem.
The hard part is understanding users deeply enough to identify the right problems, finding signals in data, deciding which opportunities matter, and translating those insights into a coherent growth strategy.
AI is often discussed as an execution layer - writing copy, generating experiments, analysing dashboards, building features faster. Those are useful capabilities, but I think they undersell what AI could mean for consumer growth.
The more interesting question is:
What happens when AI becomes part of how we understand users, design growth strategies, and eventually run the growth funnel itself?
I don't think AI simply replaces the traditional consumer growth funnel. Instead, I see an evolution:
Traditional growth → AI-assisted growth → AI-native growth.
The funnel remains recognisable. But the intelligence operating it becomes fundamentally different.
1. The traditional growth funnel isn't going away
Consumer growth has spent years building sophisticated mechanics around:
Acquisition → Activation → Retention → Sharing → Monetisation
!Source: Gemini
Source: Gemini
Recommendations, onboarding funnels, notifications, referrals, SEO, web-to-app conversion and reactivation are all well-understood growth systems. AI doesn't suddenly make these obsolete.
A traditional recommendation system already learns from behavioural signals: Did the user click? Did they finish watching? Did they like the content? Did they skip it? Those signals help the system determine what to show next.
The same is true for onboarding. Growth teams already segment users, identify their intent and progressively optimise the journey towards activation.
AI can make these systems smarter, but the underlying growth objective remains the same. The funnel is not necessarily changing. The intelligence behind the funnel is.
2. The first transformation: AI becomes a growth strategist
When people talk about AI and growth, the conversation often jumps immediately to execution:
- generate more creative
- automate campaigns
- write notifications
- personalise recommendations
- analyse experiment results
But growth teams spend enormous amounts of time on something that happens before execution: figuring out what to do in the first place.
A strong growth strategy requires connecting many fragmented signals:
- What are users doing?
- Where are they dropping off?
- Which cohorts behave differently?
- What problems are emerging?
- Which changes are behavioural versus structural?
- What opportunity is actually worth pursuing?
- What should we prioritise against everything else?
This is where I think AI can become a particularly powerful partner to a growth PM. Instead of simply asking AI to analyse a dashboard, imagine giving it access to behavioural data, experiment results, qualitative feedback and product context. AI can help synthesise those signals, surface patterns, challenge interpretations, identify potential problems and suggest opportunities, while the PM still needs to exercise judgement.
But the distance between raw data → insight → strategy → roadmap becomes much shorter.
This matters particularly in growth because growth problems are often highly interconnected. A decline in retention might not actually be a retention problem. It could originate from acquisition quality, onboarding, recommendation relevance, notification fatigue or a change in user intent. AI could help a PM reason across these connections much faster.
The opportunity isn't just AI executing the growth strategy. It is AI helping the PM construct a better growth strategy in the first place.
3. AI makes the funnel more granular
The second transformation is more obvious: personalisation. Consumer products have always segmented users.
A growth team might distinguish between:
- new and returning users
- high- and low-intent users
- different demographics
- different acquisition channels
- different behavioural cohorts
AI makes it increasingly possible to move from broad segments towards individualised situations.
The question shifts from:
“What should we do for users in this segment?”
towards:
“What does this particular user seem to need right now?”
A user doesn't have one permanent persona. Their intent changes constantly.
The same person can be browsing casually at 10pm one night, actively searching for something the next morning, and completely uninterested in the product a week later.
AI is particularly well suited to interpreting these changing states.
That means existing growth mechanics, recommendations, notifications, onboarding, reactivation, can become much more context-aware.
The mechanics remain but the decision-making inside them becomes more dynamic.
4. Retention may be where AI has the biggest impact
Of all the traditional growth stages, I think retention has particularly strong potential to benefit from AI. Retention ultimately comes down to a simple question:
Does this product continue to be useful and relevant enough for the user to come back?
AI can make a product substantially more personalised; AI can learn preferences, understand context and adapt its behaviour over time, and instead of every user experiencing roughly the same product, the product increasingly becomes shaped around the individual. That can increase relevance, which is one of the foundations of retention.
This is different from simply sending more notifications or adding more incentives to return.
The stronger version is: make the product itself more useful each time the user comes back.
That creates a potentially powerful retention loop:
User interacts → AI learns → product becomes more relevant → user gets more value → user returns → AI learns more.
The product becomes better at serving the user through repeated interaction. That is a much more interesting retention mechanism than simply optimising the frequency of reminders.
5. AI can compress the funnel
There is another change that I think is easy to overlook. AI can allow users to accomplish what previously required several steps in one. Imagine a traditional funnel where a user has to:
discover → search → compare → decide → execute
An AI system might allow the user to express their intent once and have the system handle much of the work in between. The user is still moving through the same underlying value journey, but the number of actions required from them decreases.
AI can effectively compress the funnel. From a growth perspective, that can be enormously valuable.
If users previously needed three actions to reach a meaningful value moment and AI reduces that to one, the conversion rate between intent and value can improve dramatically.
This suggests an important shift in how growth teams should think about AI:
Don't only ask whether AI creates new funnel stages. Ask whether AI can remove existing friction between stages.
Sometimes the biggest growth opportunity is not adding another intervention, it is eliminating three steps altogether.
6. Sharing is different
Not every part of the growth funnel changes equally. For example, I think AI's impact on sharing is less direct.
AI can certainly help determine:
- what content a user is likely to share
- who they might share it with
- when they are most likely to share
- which intervention might prompt sharing
But the actual act of sharing is still fundamentally behavioural. This is an important distinction.
AI can become extremely good at identifying the conditions for a growth action without necessarily performing that action for the user. That means traditional growth mechanics still matter. The opportunity is to make them much more intelligent.
7. The really interesting shift: from reactive to anticipatory growth
This is where I think AI-native consumer growth becomes genuinely different. Consider how recommendation systems work today. They largely learn from behavioural signals:
User watches → system learns → system recommends something similar.
The user is continuously providing signals, and the system responds, but what happens when the user's behaviour starts changing?
Imagine a user who normally enjoys the recommendations they receive. Suddenly, completion rates fall. They skip more content. They stop engaging.
A traditional system might continue optimising recommendations based on the available signals.
An AI-native system could reason differently:
“This user appears to be losing interest in this type of content. What else might satisfy the underlying need?”
Perhaps the user wants a different content format, to search for something specific or maybe they are bored with their current consumption pattern entirely. The product could then introduce an alternative behaviour at an appropriate moment without disrupting the user's experience or damaging the existing business metrics.
The goal isn't simply: “Get this user to consume one more piece of content.”
It becomes: “Understand why this user's experience is deteriorating, infer what might restore value, and intelligently intervene.”
That is a fundamentally more sophisticated growth loop.
8. From optimisation to intervention
This creates an interesting evolution in the role of a growth system. Traditional growth systems are largely reactive. They observe behaviour and optimise the next step. AI makes it possible to move towards anticipatory growth.
The system can continuously ask:
- What does this user appear to want?
- Is their intent changing?
- Are they becoming less engaged?
- What underlying need might explain that?
- What could help them next?
- When should we intervene?
- Should we intervene at all?
The last question is particularly important. A smarter system does not necessarily mean more interventions. If AI constantly interrupts users with “helpful” suggestions, it can easily destroy the experience it is trying to optimise.
The real challenge is therefore not prediction alone - it’s judgement.
Knowing when to act, when not to act, and which action creates genuine user value. That is where I think the growth PM's role becomes even more important.
9. The next generation of growth PMs
This leads me to a broader conclusion. AI does not make growth PMs less important. It changes where their leverage comes from.
When AI can analyse data, identify patterns, generate hypotheses and execute parts of the growth strategy, the PM's value increasingly shifts towards: problem framing, strategic judgement, prioritisation and understanding users.
The PM has to decide:
- Which signals matter?
- What does the user actually need?
- Which problem is worth solving?
- What should AI optimise for?
- What should AI never optimise for?
- When should the product intervene?
- What does “better” actually mean?
This is especially important because a growth system can optimise itself into a bad product.
A system that maximises consumption isn't necessarily creating more value.
A system that maximises notifications isn't necessarily improving retention.
A system that maximises conversion isn't necessarily producing better users.
The growth PM still defines the objective function. AI can become extraordinarily good at navigating towards that objective, but someone still needs to decide where the product should be going.
10. From funnel optimisation to intelligent growth systems
I don't think the future of consumer growth is a world where AI replaces the funnel. I think it is a world where the funnel becomes increasingly intelligent, compressed and adaptive.
The evolution might look something like this:
| Traditional consumer growth | AI-native consumer growth |
|---|---|
| Segment users | Understand individual context |
| Analyse behaviour | Interpret behaviour and intent |
| Optimise predefined funnels | Dynamically adapt the journey |
| React to user signals | Anticipate changing needs |
| Add interventions | Decide whether intervention is needed |
| Personalise experiences | Continuously learn the individual |
| Optimise each funnel stage | Compress unnecessary stages |
| Build habits around existing behaviour | Potentially create entirely new behaviours |
And that last point is, to me, the most exciting. We've spent years using technology to make existing behaviours more efficient.
But the much harder and more interesting question is:
What new consumer behaviours become possible when a product can understand, predict and act on human intent in ways that weren't previously practical?
I don't have a definitive answer yet, and that's precisely why it is worth thinking about. The biggest AI growth opportunities may not come from making today's funnels perform 10% better.
They may come from discovering new ways for products to become useful, form habits and create value that today's consumer growth playbook doesn't yet have a name for.
That is where AI-native growth begins.