From Invisible to Recommended: A 90-Day AI Visibility Campaign Playbook

Ninety days is not an arbitrary number. It is the minimum timeframe required for AI visibility signals to propagate through the system — from action to crawl to index to recommendation. Understanding why 90 days works helps you set expectations and resist the temptation to judge results too early.

According to Joel House, founder of MentionLayer and author of AI for Revenue, "Every client I have worked with wants to see results in 30 days. I tell them the same thing: you will see score movement in 30 days, but the share-of-model inflection point that actually drives revenue hits between weeks 6 and 8. The teams who commit to the full 90 days are the ones who build lasting positions."

Here is how the signal chain works. When you post a citation response on a Reddit thread, that response becomes part of the thread\'s content within minutes. Google re-crawls popular Reddit threads every 1-14 days. Once Google has indexed the updated thread content, AI models that use Google\'s index as a data source (which includes Gemini and significantly influences Perplexity) pick up the new signals on their next knowledge refresh. That refresh cycle varies by model but typically runs 2-6 weeks.

So the minimum latency from posting a citation to affecting AI recommendations is approximately 3-6 weeks for a single citation. But AI models do not change their recommendations based on a single signal. They need pattern recognition — multiple citations across multiple threads, consistent entity data, review corroboration. Building that pattern of signals is a 60-90 day process.

For a tier-specific version of this 90-day playbook — one that tells you exactly what to do differently if you’re in the Invisible tier (score 0–25), Emerging tier (26–50), Present tier (51–75), or Dominant tier (76–100) — see the AI Visibility Index study. It includes a 6-dimension self-assessment to place your brand in the right tier, then hands you a stage-specific plan. Tier 1 businesses are told NOT to bother with llms.txt yet (our 1,004-business data shows it provides zero benefit below DA 20). Tier 4 businesses are told to pursue Wikipedia entries and original research. The playbook below is the generalized version; the tiered version is more prescriptive.

The expected score trajectory for an active campaign looks like this:

• Baseline (Day 0): Composite score around 25-35 for most brands. Share of model: 0-10%.
• Month 1 (Day 30): Composite score around 35-50. Entity and citation pillars show the earliest gains. Share of model may begin moving but do not expect dramatic change yet.
• Month 2 (Day 60): Composite score around 48-62. Citation density reaches critical mass. AI presence pillar begins responding. Share of model: 15-25%.
• Month 3 (Day 90): Composite score around 60-78. All pillars contributing. The compound effect kicks in — citations reinforce entity signals, press coverage reinforces review credibility. Share of model: 25-40%.

The trajectory is not linear. Weeks 1-4 feel slow because you are building a foundation that has not yet propagated. Weeks 5-8 show acceleration as the first wave of signals reaches AI models. Weeks 9-12 deliver the compounding returns where multiple signal types reinforce each other. If you quit at week 4, you miss the entire payoff. For the strategic overview behind this playbook, see our complete guide to GEO.

This playbook breaks the 90 days into specific weekly actions with measurable milestones at each stage. Follow it exactly and you will have a defensible AI visibility position by day 90.

The first two weeks are about understanding your current position and fixing the foundation. Resist the urge to start seeding citations immediately — a citation response that links to a website with broken schema and inconsistent entity data is building on sand.

Day 1-2: Run the Full 6-Pillar Audit

Execute the complete audit across all six pillars. Record the baseline scores — you will compare against these every 30 days. Pay special attention to: - Citation pillar: How many threads exist for your keywords? Where are competitors mentioned and you are not? - AI presence pillar: What is your share of model across ChatGPT, Perplexity, Gemini, and Claude? - Entity pillar: How consistent is your brand information across platforms?

Document the baseline in a format you can reproduce monthly. Exact same prompts for AI testing. Exact same keywords for citation scanning. Consistency in measurement is as important as the measurement itself.

Day 3-5: Entity Cleanup

The entity audit will surface inconsistencies. Common issues: your LinkedIn description does not match your website, your Crunchbase funding information is outdated, your Google Business Profile uses an old business category, your social media bios have different value propositions. Fix every inconsistency flagged by the audit. This typically takes 2-3 focused hours.

Update your website structured data and schema markup. At minimum, implement Organization schema with sameAs links to all your verified profiles. Add FAQ schema to your most important landing pages with questions that match the AI prompts you are targeting. Add Product or SoftwareApplication schema if applicable to your business type.

Day 5-7: Technical Foundation

Check your robots.txt — our AI crawler decision framework explains exactly which bots to allow and which to block. Check your site\'s canonical tags, sitemap, and indexing status. Ensure your key pages load fast and render properly for bots. These technical hygiene items are prerequisites for AI models to crawl and understand your site.

Day 8-10: Keyword Research and Target Identification

Expand your keyword list from the initial 10-20 to 40-60 terms. Group keywords by intent: informational ("how to..."), commercial ("best...", "top..."), transactional ("buy...", "pricing..."). Prioritize commercial and transactional keywords for citation seeding because they drive buying-intent AI queries.

Run the discovery scan against your expanded keyword list. This produces your initial thread queue — the Reddit, Quora, and Facebook Group discussions where you will seed citations. Sort by opportunity score and flag the top 30-50 threads for the upcoming sprint.

Day 11-14: Set Up Monitoring and Prepare Content

Configure your monitoring prompts — the 15-20 buying-intent questions you will test against AI models weekly. These should mirror the queries your potential customers actually ask. Set up weekly automated testing across ChatGPT, Perplexity, Gemini, and Claude.

Develop your brand brief, tone guidelines, and response rules if you have not already. These documents control the quality of every citation response you generate. Invest time here — a well-crafted brand brief is the difference between responses that earn upvotes and responses that get downvoted.

Week 2 Milestone: Audit complete, entity issues resolved, schema updated, 30-50 target threads identified, monitoring configured. You should have a clear picture of the gap between your current position and where you need to be.

This is the highest-intensity phase of the campaign. The goal is to place your brand in 30-50 high-authority threads with authentic, valuable responses. The citation seeding playbook covers the tactical details — this section focuses on pacing, prioritization, and quality control.

Week 3: First Wave (8-10 responses)

Start with your highest-opportunity threads — the ones with the best combination of Google rank, comment count, relevance score, and recency. For each thread, generate three response variants (casual, expert, story) and select the one that best matches the thread\'s existing tone and culture.

Post 2-3 responses per day, spread across different subreddits and platforms. Do not dump 10 responses in the same subreddit on the same day — this looks unnatural and can trigger mod attention. Variety in platform, community, and timing creates a more authentic pattern.

After posting, monitor engagement for 48 hours. Track upvotes, replies, and any moderator actions. If a response gets removed or downvoted, analyze why and adjust your approach for the next batch. Common early mistakes: being too salesy, not matching the subreddit\'s culture, not engaging with the thread\'s specific context.

"The single highest-information gain signal you can send to an AI model is a citation response that earns genuine community engagement," says Joel House. "A response with 40 upvotes in a thread ranking position 3 in Google is worth more to your AI visibility than 10 responses with zero engagement in threads that do not rank."

Week 4: Second Wave (10-12 responses)

By week 4, you have learned what works for this brand in these communities. Double down on the response style and platforms that earned the best engagement in week 3. Expand into secondary threads that you deprioritized initially.

Start tracking which threads are ranking well in Google. A citation response in a thread ranking position 3 for a high-volume keyword is worth more than one in a thread ranking position 47. Re-prioritize your queue based on live SERP data, not just the initial opportunity score.

Weeks 5-6: Sustained Pace (8-10 responses per week)

Maintain the posting pace while shifting focus to thread diversity. If your first 20 responses were all on Reddit, expand into Quora. If they were all in one subreddit, branch into adjacent communities. The goal is a broad citation footprint across multiple platforms and communities, which sends a stronger signal to AI models than deep presence in a single location.

By the end of week 6, you should have 30-50 active citation placements across your keyword universe. Many of these will have accumulated upvotes and replies, making them more authoritative.

Week 6 Milestone: Run a mid-campaign citation pillar check. Your citation score should have improved by 15-25 points from baseline. If it has not, diagnose: Are the threads you targeted actually ranking in Google? Are your responses getting engagement? Did any get removed? Course-correct before moving to the next phase.

Share of model may begin to move during weeks 5-6 for Perplexity (which refreshes most frequently), but do not expect ChatGPT or Gemini movement yet. This is normal and does not indicate a problem.

With the citation foundation in place, weeks 7-10 shift focus to the signals that build authority and credibility — the signals AI models use to decide whether your brand *deserves* to be recommended, not just whether it is *mentioned*.

Weeks 7-8: PR Campaign Launch

Develop 3-5 story angles that connect your brand to current industry trends, newsworthy data, or genuine expertise. Our guide on digital PR in the AI era covers which types of coverage AI models actually cite. The stories should be legitimately newsworthy — not thinly veiled press releases. AI models weight earned media from recognized publications much more heavily than syndicated press releases.

Target publications in your industry vertical plus general business/tech outlets. For B2B: industry trade publications, relevant podcasts, business journals. For B2C: consumer publications, lifestyle media, local business press. For SaaS: TechCrunch, Product Hunt, industry blogs, podcast appearances.

Distribute 2-3 press releases through wire services for broad coverage. Simultaneously, pitch 5-10 journalists directly with personalized angles relevant to their recent coverage. The wire service coverage builds volume; the direct pitches build high-authority placements.

Track every piece of coverage earned: publication name, URL, whether it includes a backlink, estimated domain authority. Also track any unlinked mentions — these are opportunities for follow-up outreach to convert into linked citations. This data feeds directly into the press pillar of your next audit.

Weeks 8-9: Review Generation Campaign

Launch a structured review request campaign targeting your two highest-priority review platforms. For most businesses, this means Google Reviews plus one industry-specific platform (G2 for SaaS, TripAdvisor for hospitality, Avvo for legal, etc.).

Identify your best review candidates: customers who have expressed satisfaction verbally, customers who completed onboarding successfully in the last 30 days, customers who recently expanded their engagement. Send personalized review requests with direct links to the review form.

Target 15-25 new reviews during this phase. A realistic conversion rate from review request to published review is 15-25%, so you need to send 80-150 requests. Automate the request triggers but personalize the message.

Weeks 9-10: Content and Continued Seeding

Maintain citation seeding at a reduced pace — 2-3 responses per week to keep the momentum without burning out your team. Focus these on newly discovered high-opportunity threads from your ongoing automated scans.

Create 2-3 pieces of content designed to be cited by AI models. This could include a comprehensive guide, an industry report with original data, or a comparison resource. Publish on your own site with strong schema markup, then promote through social channels and industry communities. The goal is to create pages that AI models reference directly as authoritative sources.

Week 10 Milestone: You should have: 3-5 pieces of press coverage earned, 15-25 new reviews published, 35-55 total citation placements active, and your AI presence pillar should be showing the first signs of movement. Share of model across all models should be measurably higher than baseline — typically 15-25% up from the starting 0-10%.

The final two weeks close the loop. This is where you prove the campaign worked, identify what to optimize, and set the stage for the next 90 days.

Day 75-77: Run the Complete Re-Audit

Use exactly the same methodology as the baseline audit — same keywords, same AI prompts, same platforms scanned. Consistency in measurement is critical. Run all six pillars and calculate the new composite score.

Compare pillar-by-pillar: - Citations: How many new threads now mention your brand? How has the competitor gap changed? - AI Presence: What is your share of model across each AI platform? Which prompts are you appearing in that you were not before? - Entities: Are all inconsistencies resolved? Has your knowledge panel appeared or improved? - Reviews: What is the new total review count and average rating? How does velocity compare to competitors? - On-Page: Is your schema markup complete? Has your citable content and sitemap coverage improved? - Press: How many new third-party mentions? What is the authority distribution?

Document every metric in the same format as the baseline so the comparison is visual and immediate.

Day 78-82: Calculate ROI

Apply the ROI framework to the 90-day period. Layer 1: audit score delta (the headline number). Layer 2: share of model gains per AI platform. Layer 3: activity summary (citations posted, press earned, reviews generated). Layer 4: revenue attribution from AI-referred traffic and citation link clicks.

Build the comprehensive report that tells the 90-day story: "You started at 28/100. We took these actions. You are now at 62/100. Here is the revenue impact. Here is the plan for months 4-6."

"In our experience running AI visibility campaigns at MentionLayer, we\'ve found that the re-audit is the single most important deliverable of the entire 90 days," says Joel House. "It is simultaneously the proof that the campaign worked, the retention tool for the next quarter, and the sales asset for your next 10 prospects."

Day 83-86: Analyze What Worked and What Did Not

Not all activities produce equal returns. Analyze which citation placements earned the most engagement. Which subreddits and platforms delivered the best results? Which response variants (casual, expert, story) performed best for this brand? Which PR angles generated coverage and which fell flat?

This analysis shapes your month 4-6 strategy. If Reddit citations drove 80% of your pillar improvement while Quora drove 5%, shift resources accordingly. If expert-variant responses outperformed casual variants 3:1, adjust your generation approach.

Day 87-90: Plan the Next Phase

Based on the re-audit results and performance analysis, build the month 4-6 plan. Typical priorities: - Continue citation seeding at maintenance pace (5-8/week) on highest-performing platforms - Expand into new keyword clusters that were not covered in the initial sprint - Launch a second PR wave with fresh story angles - Intensify the review generation campaign on the platform showing best results - Begin tracking competitor movements more closely — are they starting their own GEO campaigns?

Week 12 Milestone: Comprehensive re-audit complete, ROI report delivered, month 4-6 plan approved, and the client or stakeholder has clear visibility into what was accomplished and what comes next.

Setting realistic expectations is essential. Here are the ranges we see across campaigns, broken down by pillar.

Citation Pillar: +15 to +25 points

Starting from a typical baseline of 15-25, expect to reach 35-50 after 90 days. The improvement comes directly from the citation seeding sprint. Brands that start at a very low baseline (under 10) may see larger gains because the first citations have an outsized impact on a near-zero score. Brands in highly competitive categories with entrenched competitors may see gains at the lower end of the range.

AI Presence Pillar: +10 to +20 points

This is the lagging indicator. Starting from a typical baseline of 5-15, expect to reach 20-35 after 90 days. AI presence improvements depend on citation velocity reaching critical mass AND AI models refreshing their knowledge base. Some models (Perplexity) update rapidly; others (ChatGPT) update on longer cycles. Do not judge this pillar too early — week 6-8 is when meaningful movement starts.

Entity Pillar: +20 to +30 points

Starting from a typical baseline of 25-45, expect to reach 55-75 after 90 days. Entity improvements are the fastest to appear because they depend on fixing discrete issues (inconsistent descriptions, missing profiles, incomplete schema) rather than accumulating external signals. Most of this gain happens in weeks 1-2.

Review Pillar: +5 to +15 points

Starting from a typical baseline of 20-50, expect to reach 30-60 after 90 days. Review improvement is the hardest to move quickly because you are dependent on customers taking action. The review generation campaign in weeks 8-9 seeds the pipeline, but many reviews will arrive in months 4-5. Set expectations accordingly — this pillar shows the largest gains in the second 90-day cycle.

Press Pillar: +10 to +20 points

Starting from a typical baseline of 10-30, expect to reach 25-45 after 90 days. Press gains depend on the quality of your story angles and the competitiveness of your media landscape. Industries with active trade publications (SaaS, finance, health) tend to see faster PR results than niche categories with fewer outlets.

Composite Score: +25 to +45 points

The weighted composite typically moves from the 25-35 range to the 55-75 range over 90 days. This translates to moving from "invisible to AI" to "emerging contender" in your category. Going from contender to leader typically requires a second 90-day cycle.

These ranges assume active, consistent execution. Campaigns that have execution gaps — missed weeks, paused citation seeding, delayed PR — see results at the lower end or below these ranges. Consistency matters more than intensity.

Having run and analyzed dozens of 90-day campaigns, these are the failure modes that show up repeatedly. Avoiding them is as important as following the playbook.

Pitfall 1: Trying to do everything at once. The playbook phases exist for a reason. Teams that try to run citation seeding, PR, and review campaigns simultaneously from week 1 end up doing all three poorly. The phased approach ensures each activity gets the attention and resources it needs. Entity cleanup and citation seeding are the foundation — do not dilute week 1-6 energy on activities scheduled for weeks 7-10.

Pitfall 2: Low-quality citation responses that get downvoted or removed. The single most common failure in citation seeding is writing responses that sound like marketing copy instead of authentic community participation. If your responses are getting downvoted or removed by moderators, the problem is almost always tone — too promotional, too generic, too disconnected from the thread\'s actual conversation. Read the citation seeding playbook and follow the rules: provide genuine value before mentioning the brand, match the platform\'s culture, acknowledge other options, include specific details from the thread.

Pitfall 3: Not tracking results. Teams that do not run the mid-campaign check at week 6 and the full re-audit at week 12 are flying blind. They cannot course-correct if something is not working, and they cannot prove ROI when renewal conversations arrive. Measurement is not optional — it is the mechanism that makes GEO a repeatable, improvable process rather than a one-time experiment.

Pitfall 4: Giving up at week 4 when AI presence has not moved. Week 4 is the patience test. Citation seeding is in full swing, but AI presence has not budged. This is expected — the signal propagation timeline means AI models have not yet processed the new citations. Teams that interpret static AI presence as "this is not working" and abandon the campaign miss the inflection point that typically arrives at weeks 6-8. Commit to the full 90 days before drawing conclusions.

Pitfall 5: Not re-auditing to prove progress. Some teams do great execution work but never run the comparison audit. They know subjectively that things have improved, but they cannot quantify the improvement for their client or executive team. The re-audit is not just a measurement tool — it is the sales tool for the next 90-day engagement. Without a clear before/after comparison with specific numbers, renewals are harder to close.

Pitfall 6: Ignoring community feedback. When a citation response earns a thoughtful reply — someone asking a follow-up question, challenging a claim, or sharing their own experience — respond. Community threads are living conversations. A brand that posts and disappears looks like a bot. A brand that engages in follow-up conversation builds credibility that extends beyond the original citation.

The most important thing about GEO is that it compounds. Unlike paid advertising where results stop when spending stops, AI visibility builds cumulative equity that accelerates over time.

Here is how the compound effect works. More citations → higher authority signals → more AI mentions → more organic citations. Once your brand appears in AI recommendations, real users start mentioning you in their own Reddit posts and Quora answers. Those organic mentions create additional citations that you did not seed, which further strengthen your AI presence. This flywheel takes 60-90 days to start spinning but accelerates meaningfully in months 4-6.

Month 4-6 Strategy: Expand and Deepen

Maintain citation seeding at a steady pace of 5-8 responses per week. This is lower intensity than the initial sprint but sufficient to keep building citation density. Shift focus from your core keywords to adjacent keywords and emerging topics. Expand into new platforms — if you dominated Reddit in the first 90 days, deepen your Quora presence.

Launch a second PR wave with fresh story angles. Use the data from your first campaign to craft more targeted pitches. If a particular publication ran your story, pitch them a follow-up. If a journalist covered your competitor, pitch them the contrasting angle.

Intensify review generation. The review request pipeline you built in weeks 8-9 should be generating steady reviews by now. Scale the triggers, refine the messaging, and expand to additional platforms. The review pillar often shows its largest gains in months 4-6 as the pipeline matures.

Month 7-12 Strategy: Dominate and Defend

By month 7, your composite score should be in the 65-80 range. The focus shifts from building to defending and optimizing. Monitor competitor activity — are they starting their own GEO campaigns? If so, increase your pace in the areas where they are gaining ground.

Expand your keyword coverage aggressively. The first 90 days likely covered your top 40-60 keywords. Months 7-12 should expand to 100-200 keywords across long-tail variations, seasonal terms, and emerging topics. Each new keyword cluster generates a new wave of citation opportunities.

Begin experimenting with YouTube GEO — our platform-by-platform GEO guide covers video optimization for AI citation. YouTube content is increasingly referenced by AI models, especially Perplexity and Gemini, and video citations are a defensible moat that most competitors are not building yet.

The Long-Term Trajectory

Brands that execute consistently for 12 months typically reach composite scores of 75-90 and share of model rates above 50% for their core queries. At this level, you are a category leader in the AI recommendation channel. New competitors entering the space face a significant accumulation gap — your 12 months of citation density, review volume, and press coverage cannot be replicated overnight.

The brands building AI visibility now are creating a durable competitive advantage. Every month of execution widens the gap with brands that have not started. The best time to start was 6 months ago. The second best time is today. Run the audit, see your score, and begin. Start your free AI visibility audit now, or explore our citation engine to see how MentionLayer automates the seeding workflow.