Thursday, May 22, 2025

Driving Growth with AI, How Adjust Is Reimagining AI User Acquisition

Driving Growth with AI, How Adjust Is Reimagining AI User Acquisition
Andrey Kazakov and Avi Blumenstein

 

 

 

Quick Read Summary

AI user acquisition is no longer about getting answers faster, it is about shrinking the distance between a business question and a decision.

The practical win starts with automated reporting that turns analysis into a query, with sources you can verify, then moves toward systems that surface opportunities and risks without waiting to be asked.

The bigger shift is operational. Growth leaders are moving from doing the work to designing the system that does it, with human judgement as the guardrail and acceleration as the payoff.

 

 

The real growth bottleneck is decision latency

Most growth teams do not lose because they lack dashboards. They lose because it takes too long to turn raw performance signals into a confident decision, especially when the questions come from product, finance, and executive stakeholders at the exact moment the business needs clarity.

Avi Blumenstein described the everyday reality: urgent questions, partial context, and the time sink of reconstructing an answer from notes and data. He has shifted that workflow into an AI first process that produces a structured draft response almost instantly, then relies on human verification before anything goes out.

“I’m getting like 80 percent of the way there within 60 seconds. I refine that a bit and then I’ve got a beautiful response.” Avi Blumenstein, Growth Marketing, Age of Learning

The implication for senior leaders is direct. If your team is still answering high stakes questions through manual reconstruction, your organization is paying a compounding tax in time, missed opportunity, and delayed mitigation.

Practical action: map the top twenty recurring stakeholder questions, then build an AI supported response workflow with explicit verification steps, including what must be checked, where the source of truth lives, and what is never safe to infer.

 

Automated reporting should behave like a conversation, not a dashboard

Speed only matters if the output is trustworthy. The most useful version of automated reporting is not a faster chart, it is a dialogue where the system can generate the report, show the underlying sources, and let the operator iterate toward the exact business question they are trying to answer.

Andrey Kazakov framed the sequence clearly, reporting comes first because decisions require data that arrives quickly and can be trusted.

Avi made the value tangible. Work that previously required building a report inside a BI style interface now becomes a single query, followed by light manual adjustment if needed.

“Where I would have gone into DataScape, maybe it would have taken me 15 minutes, 20 minutes to output a report, I can just put in a query right now and it’ll get me there.” Avi Blumenstein, Growth Marketing, Age of Learning

This is the strategic leap for AI user acquisition. When automated reporting becomes conversational, the growth operator spends less time building views and more time interpreting what matters, pressure testing assumptions, and moving to action.

Practical action: create a shared “question library” for growth, including the exact phrasing that reliably produces the analysis you need, plus the follow up questions that validate the output, such as cohort definitions, geo splits, and payback windows.

 

The next leap is proactive systems that surface what you did not think to ask

Reactive analysis is structurally limited. It depends on someone noticing a signal, deciding to investigate, and asking the right question. Kazakov’s vision moves beyond that constraint, toward an asynchronous agent that flags risks and opportunities as they emerge.

The examples are the ones that routinely cost growth teams money:

  • creative that is outperforming on one video network and should be ported to other channels faster
  • creative fatigue that requires intervention before efficiency collapses
  • anomalies caused by integration issues, country level measurement breaks, or platform specific issues
  • performance drops that correlate with app updates, then become debuggable when workflow context is available

Avi connected this to the operational pain of diagnosing whether a drop is marketing, product, or measurement, especially during international expansion and platform specific monetization issues.

The interpretation is that proactive intelligence is not a nice to have, it is a margin and growth protection mechanism. The faster you detect, the less you waste.

Practical action: define your “always watch” list, including thresholds for spend shifts, conversion anomalies, creative decay, and geo level breaks. Then decide which alerts can be automated now, and which require a human in the loop until trust is earned.

 

Creative and channel allocation is where AI driven pattern finding pays off

Creative performance is one of the most expensive, least systematized parts of user acquisition. Teams often have the data but not the time to connect performance patterns across networks, formats, and audiences.

Kazakov highlighted creative analysis as a prime candidate for insight generation, including the specific value of understanding how to move a creative concept from one place to another based on performance patterns.

“Having a picture and understanding on how you can move certain creative from one place to another is also really valuable.” Andrey Kazakov, CEO, Adjust

The claim here is not that AI replaces creative judgement. It changes the throughput of learning. If your testing velocity is constrained by analysis effort, you are leaving growth on the table.

Practical action: standardize a creative taxonomy that includes concept, hook, format, and audience cues. Without consistent labels, AI driven creative insight becomes shallow. With them, you can scale cross channel recommendations and shorten the path from winner identification to redeployment.

 

The biggest upside is cross functional, finance and product become first class consumers of growth insight

Growth data has always been relevant to finance and product, but too often it is trapped in growth tooling and growth language. Kazakov pointed out that finance teams want to forecast profitability and payback, and that the same performance dataset can support those simulations when paired with richer internal context.

He also noted that bringing in internal metrics like predicted lifetime value, or workflow context from tools like Asana, can make the system meaningfully more valuable because it explains the “why” behind the “what.”

The implication is that growth leaders should treat measurement and insight as an enterprise asset, not a departmental function. This is where forecasting becomes more than a slide deck exercise. It becomes an operational capability, tied to pacing, ROAS discipline, payback targets, and product release cycles.

Practical action: align on a shared “decision model” with finance and product, including the definitions of payback, incrementality assumptions where relevant, and the internal metrics you will use to contextualize performance, such as pLTV models and retention curves.

 

Governance matters, move from moderated insight to automation in deliberate stages

Automation is not a switch, it is a trust ladder. The credible model is staged: reporting first, then forecasting, then automation, with clear human oversight until reliability is proven.

A key concept in this progression is moderation, an insight is produced, a human confirms the recommendation, and only later does the organization choose to remove that checkpoint.

“There will be an insight that is moderatable. Someone confirms the advice. Eventually you can decide you do not want it moderated anymore.” Andrey Kazakov, CEO, Adjust

This is the operational playbook growth leaders need. AI user acquisition will increasingly include systems that suggest next steps, and eventually execute them, but only the organizations with disciplined guardrails will capture upside without creating risk.

Practical action: define three tiers of action, informational, recommended, and executable. Keep “executable” narrow at first, with explicit rollback plans and audit trails, and expand only when error rates, confidence scoring, and business impact justify it.

 

Conclusion

Driving growth with AI is ultimately about replacing friction with feedback loops. Automated reporting reduces the time cost of knowing what happened, proactive systems reduce the time cost of noticing what matters, and cross functional forecasting reduces the time cost of aligning on what to do next.

The teams that win will not be the ones with the most tools. They will be the ones that operationalize AI user acquisition as a system, with strong data foundations, clear verification norms, and a measured path from insight to automation.

 

 

Speakers

Andrey Kazakov, CEO, Adjust.  Andrey leads Adjust with a focus on building artificial intelligence driven systems that support modern growth teams. His work centers on developing tools that simplify reporting, enhance decision making, and create intelligent workflows that help marketers operate at scale. Andrey brings deep experience across advertising, measurement, and data platforms, and is guiding Adjust toward a future where artificial intelligence acts as a strategic partner for user acquisition teams worldwide.

Avi Blumenstein, Growth Marketing, Age of Learning. Avi oversees growth marketing at Age of Learning, where he focuses on performance strategy, creative optimisation, and data informed decision making. He is an early and active user of artificial intelligence tools in his daily workflow, applying them to improve reporting speed, internal communication, and campaign analysis. Avi offers a practitioner’s view of how artificial intelligence can support real world growth challenges and help teams deliver stronger results with greater efficiency.

 

Loading