The Challenge: Inefficient Audience Targeting

Marketing teams are under pressure to deliver growth, but many still aim campaigns at broad, imprecise audiences. You group users by high-level demographics or generic interests, then hope your message hits the right people. The result: budgets leak on impressions and clicks from users who were unlikely to convert from the start.

Traditional approaches to audience targeting rely heavily on manual analysis, platform presets, and marketer intuition. You pull segments in Google Ads, Meta, or your DSP, try to interpret search terms and audience insights, then update targeting rules by hand. This process is slow, biased by personal experience, and almost impossible to keep consistent across channels and markets. As user behavior fragments and privacy rules change, guesswork simply can’t keep up.

The cost of not solving this is significant. Wasted spend on mismatched audiences inflates customer acquisition costs and pushes ROAS down. Sales teams receive low-quality leads. Valuable micro-segments that could convert at high rates remain hidden in your data. Meanwhile, competitors that use AI to refine their audiences can outbid you efficiently, dominate the best inventory, and set new performance benchmarks your manual setup can’t match.

The good news: this problem is highly solvable with the right use of AI. By combining your existing Google marketing data with Gemini, you can move from intuition-based targeting to data-driven audience design at scale. At Reruption, we’ve helped organisations build AI-first workflows that turn raw signals into actionable segments and negative audiences. Below, we’ll walk through a practical, non-theoretical approach you can use to systematically improve your audience targeting and ad performance.

Need a sparring partner for this challenge?

Let's have a no-obligation chat and brainstorm together.

Innovators at these companies trust us:

Our Assessment

A strategic assessment of the challenge and high-level tips how to tackle it.

From Reruption’s hands-on work building AI-first marketing workflows, we see a clear pattern: companies already sit on rich Google Ads and Analytics data, but struggle to turn it into precise audiences. Gemini is particularly strong here because it natively understands Google’s ecosystem and can interpret search terms, queries, and funnel performance into concrete audience targeting strategies that marketers can actually use.

Anchor Gemini in a Clear Audience Strategy, Not Just "Better Keywords"

Many teams try AI as a quick fix for keyword lists or ad copy, but ignore the underlying audience strategy. Before plugging Gemini into your marketing workflows, define what “good” targeting means for your business: which customer profiles are truly profitable, which intents matter, and which segments you actively want to exclude. This strategic clarity gives Gemini a frame to generate meaningful audience segment frameworks instead of generic ideas.

Use your existing LTV, margin, and sales feedback to describe your best and worst-fit customers in business terms. Then instruct Gemini to translate those profiles into search behavior, content consumption patterns, and likely channel touchpoints. That way, the AI operates as a strategic partner in sharpening your targeting, not just as a keyword expansion tool.

Treat Gemini as an Analyst for Cross-Channel Audience Signals

Marketing organisations are often structured by channel: search, social, display, affiliate. Audience insights get trapped in silos, and each team reinvents the wheel. A smarter approach is to use Gemini as a cross-channel analyst that can read exported reports from multiple platforms and identify patterns you’d miss manually.

Strategically, this means setting up a recurring workflow where audience performance data from Google Ads, Analytics, and potentially CRM exports are summarised and interpreted by Gemini. The goal is not to replace your channel experts, but to give them a shared, AI-generated perspective on which intents, creative angles, and demographic traits consistently drive profitable responses across channels.

Prepare Your Team for AI-Augmented Decision-Making

Even the best AI insights are wasted if the marketing team lacks the mindset or processes to act on them. Before scaling Gemini-based targeting, align your marketing, data, and performance teams on how AI-driven audience recommendations will be used. Clarify roles: who validates new segments, who owns negative audience decisions, and who is allowed to adjust bids or budgets based on Gemini’s suggestions.

This organisational readiness matters more than another dashboard. Encourage a test-and-learn culture where Gemini’s recommendations are treated as hypotheses to be validated via controlled experiments, not as unquestionable truths. This keeps risk under control while building trust in AI-assisted targeting.

Design Guardrails to Mitigate Risk and Bias

AI-generated segments can sometimes drift into sensitive or non-compliant territory (e.g. proxies for protected characteristics, or segments that violate your brand’s guidelines). At a strategic level, you need clear guardrails for how Gemini is allowed to refine audiences. This includes defining forbidden targeting criteria, safe negative audiences, and compliance considerations with your legal and data protection teams.

Embed those constraints into your prompts and workflows from the start. For example, explicitly instruct Gemini to avoid using health, political, or other sensitive attributes when suggesting new audience ideas. This reduces risk and prevents uncomfortable surprises when ideas move from analysis to live campaigns.

Start with Focused Pilots in High-Impact Funnels

Trying to “AI everything” at once is a recipe for chaos. Instead, strategically select 1–2 high-impact funnels—such as branded and non-branded search for your top product line—and use Gemini-powered audience refinement there first. These funnels usually have enough volume to generate statistically valid results and enough business relevance to showcase impact.

Define clear success metrics (e.g. % reduction in cost per acquisition, improvement in conversion rate from new segments, or decrease in wasted spend from negative audiences) and a fixed test period. Once the pilot shows stable improvements, you can roll out the approach to additional campaigns, channels, and markets with more confidence.

Used deliberately, Gemini can transform inefficient audience targeting from a guessing game into a systematic, data-backed process that continuously improves ROAS. The key is to combine your business understanding with Gemini’s analytical power and to embed it into the way your team makes targeting decisions. At Reruption, we specialise in turning these concepts into working AI tools and workflows inside your organisation, not just slideware—if you want to explore a focused proof of concept or a production-ready setup, our team is ready to help you design and implement it.

Need help implementing these ideas?

Feel free to reach out to us with no obligation.

Real-World Case Studies

From Investment Banking to Energy: Learn how companies successfully use Gemini.

Goldman Sachs

Investment Banking

In the fast-paced investment banking sector, Goldman Sachs employees grapple with overwhelming volumes of repetitive tasks. Daily routines like processing hundreds of emails, writing and debugging complex financial code, and poring over lengthy documents for insights consume up to 40% of work time, diverting focus from high-value activities like client advisory and deal-making. Regulatory constraints exacerbate these issues, as sensitive financial data demands ironclad security, limiting off-the-shelf AI use. Traditional tools fail to scale with the need for rapid, accurate analysis amid market volatility, risking delays in response times and competitive edge.

Lösung

Goldman Sachs countered with a proprietary generative AI assistant, fine-tuned on internal datasets in a secure, private environment. This tool summarizes emails by extracting action items and priorities, generates production-ready code for models like risk assessments, and analyzes documents to highlight key trends and anomalies. Built from early 2023 proofs-of-concept, it leverages custom LLMs to ensure compliance and accuracy, enabling natural language interactions without external data risks. The firm prioritized employee augmentation over replacement, training staff for optimal use.

Ergebnisse

  • Rollout Scale: 10,000 employees in 2024
  • Timeline: PoCs 2023; initial rollout 2024; firmwide 2025
  • Productivity Boost: Routine tasks streamlined, est. 25-40% time savings on emails/coding/docs
  • Adoption: Rapid uptake across tech and front-office teams
  • Strategic Impact: Core to 10-year AI playbook for structural gains
Read case study →

Duolingo

EdTech

Duolingo, a leader in gamified language learning, faced key limitations in providing real-world conversational practice and in-depth feedback. While its bite-sized lessons built vocabulary and basics effectively, users craved immersive dialogues simulating everyday scenarios, which static exercises couldn't deliver . This gap hindered progression to fluency, as learners lacked opportunities for free-form speaking and nuanced grammar explanations without expensive human tutors. Additionally, content creation was a bottleneck. Human experts manually crafted lessons, slowing the rollout of new courses and languages amid rapid user growth. Scaling personalized experiences across 40+ languages demanded innovation to maintain engagement without proportional resource increases . These challenges risked user churn and limited monetization in a competitive EdTech market.

Lösung

Duolingo launched Duolingo Max in March 2023, a premium subscription powered by GPT-4, introducing Roleplay for dynamic conversations and Explain My Answer for contextual feedback . Roleplay simulates real-life interactions like ordering coffee or planning vacations with AI characters, adapting in real-time to user inputs. Explain My Answer provides detailed breakdowns of correct/incorrect responses, enhancing comprehension. Complementing this, Duolingo's Birdbrain LLM (fine-tuned on proprietary data) automates lesson generation, allowing experts to create content 10x faster . This hybrid human-AI approach ensured quality while scaling rapidly, integrated seamlessly into the app for all skill levels .

Ergebnisse

  • DAU Growth: +59% YoY to 34.1M (Q2 2024)
  • DAU Growth: +54% YoY to 31.4M (Q1 2024)
  • Revenue Growth: +41% YoY to $178.3M (Q2 2024)
  • Adjusted EBITDA Margin: 27.0% (Q2 2024)
  • Lesson Creation Speed: 10x faster with AI
  • User Self-Efficacy: Significant increase post-AI use (2025 study)
Read case study →

Ooredoo (Qatar)

Telecommunications

Ooredoo Qatar, Qatar's leading telecom operator, grappled with the inefficiencies of manual Radio Access Network (RAN) optimization and troubleshooting. As 5G rollout accelerated, traditional methods proved time-consuming and unscalable , struggling to handle surging data demands, ensure seamless connectivity, and maintain high-quality user experiences amid complex network dynamics . Performance issues like dropped calls, variable data speeds, and suboptimal resource allocation required constant human intervention, driving up operating expenses (OpEx) and delaying resolutions. With Qatar's National Digital Transformation agenda pushing for advanced 5G capabilities, Ooredoo needed a proactive, intelligent approach to RAN management without compromising network reliability .

Lösung

Ooredoo partnered with Ericsson to deploy cloud-native Ericsson Cognitive Software on Microsoft Azure, featuring a digital twin of the RAN combined with deep reinforcement learning (DRL) for AI-driven optimization . This solution creates a virtual network replica to simulate scenarios, analyze vast RAN data in real-time, and generate proactive tuning recommendations . The Ericsson Performance Optimizers suite was trialed in 2022, evolving into full deployment by 2023, enabling automated issue resolution and performance enhancements while integrating seamlessly with Ooredoo's 5G infrastructure . Recent expansions include energy-saving PoCs, further leveraging AI for sustainable operations .

Ergebnisse

  • 15% reduction in radio power consumption (Energy Saver PoC)
  • Proactive RAN optimization reducing troubleshooting time
  • Maintained high user experience during power savings
  • Reduced operating expenses via automated resolutions
  • Enhanced 5G subscriber experience with seamless connectivity
  • 10% spectral efficiency gains (Ericsson AI RAN benchmarks)
Read case study →

bunq

Banking

As bunq experienced rapid growth as the second-largest neobank in Europe, scaling customer support became a critical challenge. With millions of users demanding personalized banking information on accounts, spending patterns, and financial advice on demand, the company faced pressure to deliver instant responses without proportionally expanding its human support teams, which would increase costs and slow operations. Traditional search functions in the app were insufficient for complex, contextual queries, leading to inefficiencies and user frustration. Additionally, ensuring data privacy and accuracy in a highly regulated fintech environment posed risks. bunq needed a solution that could handle nuanced conversations while complying with EU banking regulations, avoiding hallucinations common in early GenAI models, and integrating seamlessly without disrupting app performance. The goal was to offload routine inquiries, allowing human agents to focus on high-value issues.

Lösung

bunq addressed these challenges by developing Finn, a proprietary GenAI platform integrated directly into its mobile app, replacing the traditional search function with a conversational AI chatbot. After hiring over a dozen data specialists in the prior year, the team built Finn to query user-specific financial data securely, answer questions on balances, transactions, budgets, and even provide general advice while remembering conversation context across sessions. Launched as Europe's first AI-powered bank assistant in December 2023 following a beta, Finn evolved rapidly. By May 2024, it became fully conversational, enabling natural back-and-forth interactions. This retrieval-augmented generation (RAG) approach grounded responses in real-time user data, minimizing errors and enhancing personalization.

Ergebnisse

  • 100,000+ questions answered within months post-beta (end-2023)
  • 40% of user queries fully resolved autonomously by mid-2024
  • 35% of queries assisted, totaling 75% immediate support coverage
  • Hired 12+ data specialists pre-launch for data infrastructure
  • Second-largest neobank in Europe by user base (1M+ users)
Read case study →

Samsung Electronics

Manufacturing

Samsung Electronics faces immense challenges in consumer electronics manufacturing due to massive-scale production volumes, often exceeding millions of units daily across smartphones, TVs, and semiconductors. Traditional human-led inspections struggle with fatigue-induced errors, missing subtle defects like micro-scratches on OLED panels or assembly misalignments, leading to costly recalls and rework. In facilities like Gumi, South Korea, lines process 30,000 to 50,000 units per shift, where even a 1% defect rate translates to thousands of faulty devices shipped, eroding brand trust and incurring millions in losses annually. Additionally, supply chain volatility and rising labor costs demanded hyper-efficient automation. Pre-AI, reliance on manual QA resulted in inconsistent detection rates (around 85-90% accuracy), with challenges in scaling real-time inspection for diverse components amid Industry 4.0 pressures.

Lösung

Samsung's solution integrates AI-driven machine vision, autonomous robotics, and NVIDIA-powered AI factories for end-to-end quality assurance (QA). Deploying over 50,000 NVIDIA GPUs with Omniverse digital twins, factories simulate and optimize production, enabling robotic arms for precise assembly and vision systems for defect detection at microscopic levels. Implementation began with pilot programs in Gumi's Smart Factory (Gold UL validated), expanding to global sites. Deep learning models trained on vast datasets achieve 99%+ accuracy, automating inspection, sorting, and rework while cobots (collaborative robots) handle repetitive tasks, reducing human error. This vertically integrated ecosystem fuses Samsung's semiconductors, devices, and AI software.

Ergebnisse

  • 30,000-50,000 units inspected per production line daily
  • Near-zero (<0.01%) defect rates in shipped devices
  • 99%+ AI machine vision accuracy for defect detection
  • 50%+ reduction in manual inspection labor
  • $ millions saved annually via early defect catching
  • 50,000+ NVIDIA GPUs deployed in AI factories
Read case study →

Best Practices

Successful implementations follow proven patterns. Have a look at our tactical advice to get started.

Use Gemini to Mine Search Term Reports into Actionable Audience Intents

Start by exporting your Google Ads search term reports for the last 30–90 days, including impressions, clicks, conversions, and cost. Combine this with performance columns so Gemini can see which terms drive value versus waste. Your goal is not just more keywords, but a richer map of user intent that informs which audiences to target or exclude.

Feed these exports into Gemini (chunked if necessary) and ask it to cluster queries into intent groups, then map them to audience characteristics and potential segment names. Include business context: products, pricing, and your definition of a qualified lead or high-value customer. This allows Gemini to distinguish between research intent, price shoppers, and high-intent buyers.

Example prompt for Gemini:
You are a performance marketing analyst.
You receive a Google Ads search term report with columns:
- search_term
- impressions
- clicks
- conversions
- cost
- conversion_value

Tasks:
1. Cluster search_term values into intent groups.
2. For each intent group, describe:
   - Typical user problem or need
   - Likely stage in the funnel (awareness, consideration, purchase)
   - Whether this group is likely high, medium, or low value, based on metrics.
3. Suggest audience segment ideas and negative audience ideas based on these groups.
4. Output in a concise table-like structure I can translate into new campaigns.

Expected outcome: a set of 8–20 intent-based audience groups and a short list of negative intents you can use to refine your targeting and exclusions.

Generate Positive and Negative Audience Frameworks from CRM and Funnel Data

Export anonymised CRM or conversion data from Google Analytics/GA4 (e.g. medium, campaign, landing page, basic demographics if available, and outcome such as LTV band or qualification status). The goal is to show Gemini which combinations of attributes tend to produce high-value versus low-value outcomes.

Ask Gemini to describe patterns in the data and to propose a framework of positive audiences (who you want more of) and negative audiences (who you want fewer impressions from). Include clear instructions about privacy and compliance so Gemini avoids suggesting sensitive attributes.

Example prompt for Gemini:
You are helping refine advertising audiences.
Here is sample data of past leads and customers with columns like:
- source / medium
- campaign
- device category
- country
- basic demographic buckets (if available)
- final_status (e.g. Won, Lost, Unqualified)
- LTV_band (e.g. <100, 100-500, >500)

1. Identify patterns that distinguish high LTV or Won customers from Unqualified/Low LTV.
2. Propose 5–10 positive audience segment definitions using only non-sensitive attributes.
3. Propose 5–10 negative audience or exclusion ideas to reduce wasted spend.
4. For each segment, explain why it should perform better or worse.

Expected outcome: a prioritised list of concrete audience rules and exclusion ideas you can translate into Google Ads audiences or custom segments in GA4.

Ask Gemini to Draft Audience-Specific Creatives and Message Variants

Once you have intent-based segments, use Gemini to generate creative variants aligned to each audience. Export your top-performing headlines, descriptions, and landing page copy from Google Ads and your website. Share these with Gemini as examples of brand voice and what has worked historically.

Then ask Gemini to produce 3–5 ad copy variants per audience segment that address the specific problem, language, and objections of that group. Include your brand and legal guidelines in the prompt so outputs are usable with minimal edits.

Example prompt for Gemini:
You are a performance copywriter for [Brand].
Here are examples of high-performing search ads and our brand guidelines: <paste>.
Here is an audience description:
"SMB owners searching for 'fast implementation marketing software' who value speed over price."

Tasks:
1. Write 5 Google search ad headlines (max 30 chars) and 4 descriptions (max 90 chars)
   tailored to this audience.
2. Emphasise implementation speed and ease, while staying within our brand voice.
3. Avoid mentioning discounts or pricing.

Expected outcome: a library of audience-tailored creatives you can plug into responsive search ads or A/B tests, reducing manual copywriting time and improving relevance.

Build a Recurring Gemini-Based Targeting Review Ritual

To keep your targeting sharp, establish a weekly or bi-weekly ritual where Gemini reviews updated performance data. Export fresh campaign and audience performance from Google Ads/GA4, then use a standard prompt to have Gemini summarise what’s working, what’s wasting spend, and which segments should be scaled or cut.

Document this as a repeatable process: where data is pulled from, how it’s prepped, which prompts are used, and how decisions are logged. Over time, you can semi-automate this with scripts or simple internal tools that feed data into Gemini and present recommendations in a structured format for your team to approve.

Example prompt for Gemini:
You are my weekly performance marketing analyst.
I will paste updated performance exports for:
- Campaign level
- Audience/segment level (if available)
- Search term level

Tasks:
1. Highlight 5–10 audiences or queries that are clearly underperforming.
2. Suggest specific negative audience or keyword exclusions to reduce wasted spend.
3. Highlight 5–10 high-performing patterns and recommend how to scale them
   (e.g. new segments, bid adjustments, new campaigns).
4. Summarise in a short, actionable report with priorities A/B/C.

Expected outcome: a consistent cadence of AI-assisted targeting improvements, with clear decisions each cycle rather than ad-hoc, reactive tweaks.

Use Gemini to Translate Business Strategy into Campaign Structures

When marketing receives a new strategic directive—entering a new market segment, launching a product, or targeting a specific industry—Gemini can help bridge the gap between abstract strategy and concrete campaign structures. Provide Gemini with your strategic brief, ICP descriptions, and historical performance insights, then ask it to propose campaign, ad group, and audience structures that reflect the strategy.

This is especially useful for teams that need to roll out consistent structures across markets or brands. Use Gemini’s output as a blueprint, then refine with your channel experts before implementation.

Example prompt for Gemini:
You are a senior performance marketing strategist.
Here is our new go-to-market strategy and ICP description: <paste>.
Here is a summary of our past campaign structures and what has worked: <paste>.

Design a Google Ads account plan that includes:
1. Recommended campaign structure (search, Performance Max, etc.).
2. Suggested audience segments and exclusions per campaign.
3. Example ad messaging angles for each audience.
4. Notes on risks and what to monitor in the first 4 weeks.

Expected outcome: faster, more consistent translation of strategy into execution, with better upfront targeting assumptions and fewer rounds of trial-and-error.

Track Impact with Clear KPIs and Reasonable Expectations

To prove value and avoid overpromising, define clear before/after metrics for your Gemini-driven targeting improvements. Focus on KPIs tied directly to audience quality, such as cost per qualified lead, conversion rate by audience, wasted spend percentage (spend without conversions), and ROAS per segment.

In many cases, teams can realistically target 10–25% reduction in wasted spend on low-intent or misaligned audiences within 8–12 weeks, alongside incremental gains in conversion rate for refined segments. The exact numbers will depend on your baseline, but the key is to measure at the segment level, not just at the overall account level, so you can see where Gemini has made a real difference.

Need implementation expertise now?

Let's talk about your ideas!

Frequently Asked Questions

Gemini can analyse your existing Google Ads, Analytics, and CRM exports to uncover patterns that are hard to see manually. It clusters search terms and performance data into intent-based groups, proposes positive and negative audience segments, and generates audience-specific creative ideas.

Instead of guessing which users to target, you can use Gemini’s analysis to focus spend on high-intent, high-value segments while systematically excluding low-quality traffic. Over time, this shifts your campaigns from broad, inefficient targeting towards precise segments that convert at a higher rate.

You don’t need a full data science team to benefit from Gemini, but you do need basic performance marketing and data-handling skills. Typically, a performance marketer or marketing operations person can export Google Ads/GA4 reports, anonymise data if needed, and work with Gemini via prompts.

For more advanced setups—like automating recurring analyses or integrating Gemini into internal tools—you’ll benefit from light engineering support (e.g. for scripting, APIs, or dashboards). Reruption often bridges this gap by pairing marketing teams with our engineers so the workflows become sustainable instead of one-off experiments.

For most organisations, you can see early impact on audience quality and wasted spend within 4–8 weeks. In the first 1–2 weeks, you’ll focus on setting up data exports, running initial analyses in Gemini, and implementing new segments and negative audiences in your campaigns.

The following weeks are about collecting enough data to validate the changes and iterating based on performance. Sustainable improvements—like a 10–25% reduction in wasted spend or a noticeable lift in conversion rate for key segments—typically become visible after one to two optimisation cycles.

The direct usage cost of Gemini for analysis and content generation is typically minor compared to your media budget. The main investment is internal time (marketing, ops, potentially engineering) and any support from external partners like Reruption.

From an ROI perspective, the benchmark is your current level of inefficiency: if even 10–15% of your spend goes to low-intent or misaligned audiences, reducing that waste with Gemini pays back quickly. Additional upside comes from better-performing segments and creatives, which can lift ROAS and lower cost per acquisition without increasing budget.

Reruption works as a Co-Preneur inside your organisation: we don’t just recommend tools, we build and ship working AI workflows with you. For Gemini-based audience optimization, we can start with our AI PoC offering (9,900€) to prove technical feasibility and demonstrate impact on a concrete funnel—using your real Google Ads, Analytics, and CRM data.

From there, our team helps you design prompts, data pipelines, and simple internal tools so marketers can use Gemini reliably without becoming prompt engineers. We embed with your performance team, challenge assumptions, and move fast—from first prototype to a production-ready audience optimization workflow that fits your compliance and reporting standards.

Contact Us!

0/10 min.

Contact Directly

Your Contact

Philipp M. W. Hoffmann

Founder & Partner

Address

Reruption GmbH

Falkertstraße 2

70176 Stuttgart

Social Media