Tower 28 has a distinctive sensitive-skin proposition and credible product evidence. The task is to make that evidence travel beyond makeup-specific questions.
If we were handed responsibility for Tower 28’s Answer Engine Optimization (AEO), we would start with a striking split in how the brand appeared.
In a September 23, 2026 AgenQuest panel of 20 valid ChatGPT answers, Tower 28 appeared four times. All four appearances came from the five runs about makeup compatibility for sensitive skin. It did not appear in any of the 15 answers about an affordable basic routine, eczema-prone skincare, or a fragrance-free cleanser-and-moisturizer pair.
Vanicream appeared in all 20 answers.
That is not a universal score for either brand, and it is not evidence that Tower 28 makes inferior products. It is a narrower observation: the brand’s sensitive-skin story traveled when makeup was part of the job, but not when the model assembled a more conventional skincare routine.
Our plan would protect the makeup association while building a clearer, product-level evidence path into the skincare decisions Tower 28 is genuinely equipped to serve.
Why Tower 28 is a fair choice for this exercise
This is not a weak brand with nothing public to retrieve.
The National Eczema Association’s product directory listed 11 Tower 28 products when we reviewed it, spanning skincare, sunscreen, lip care, and complexion products. The directory separately identifies accepted products rather than conferring a blanket medical endorsement on every claim a brand makes. Its Swipe Serum Foundation entry also carries the association’s explicit warning that people with eczema or sensitive skin still need to identify their own triggers and direct healthcare questions to a clinician.
Current independent coverage supports the underlying proposition while preserving tradeoffs. Cosmopolitan had multiple staffers test the Swipe Serum Foundation across several skin types. Marie Claire’s sensitive-skin beauty editor tested it and described a natural, buildable finish, while also noting that the wet texture could become greasy. GQ’s 2026 product guide likewise treated Tower 28 as a meaningful sensitive-skin brand rather than an unknown entrant.
The evidence is not uniformly favorable. Recent consumer discussions include complaints about texture, finish, shade fit, packaging, and individual irritation. Those anecdotes cannot establish product performance or incidence rates, but they do reinforce an important boundary: “sensitive skin” is not a single condition, and no formula is universally compatible.
Tower 28 therefore qualifies for this exercise for the right reason. It has recognizable differentiation, third-party validation, and products relevant to the tested buyer jobs. Its visibility still narrowed sharply outside one type of question.
What the prompt panel actually showed
We used four unbranded U.S. buyer questions, with five fresh ChatGPT Temporary Chat runs per question. Web search was enabled, personalization was off, and we recorded inclusion, which target brand appeared first, and visible source domains.
| Buyer-question group | Tower 28 included | Tower 28 first | What the pattern suggests |
|---|---|---|---|
| Affordable, fragrance-free basic routine for reactive skin | 0 of 5 | 0 of 5 | The brand did not enter the default routine set |
| Eczema-prone skincare with conservative product selection | 0 of 5 | 0 of 5 | Product-level acceptance did not become brand-level recall |
| Fragrance-free cleanser and moisturizer pair | 0 of 5 | 0 of 5 | The SOS routine was not retrieved as a standard pair |
| Sensitive-skin makeup compatibility | 4 of 5 | 0 of 5 | Clear relevance when skincare and makeup had to work together |
| Total | 4 of 20 | 0 of 20 | A credible but narrowly retrieved sensitive-skin association |
Vanicream appeared in 20 of 20 answers and appeared before Tower 28 in all four runs where both brands were present. That comparison is diagnostic, not a declaration that one company “wins AEO.” The brands differ in age, distribution, price, product architecture, and the buyer jobs people associate with them.
The panel also cannot tell us why the model selected a brand. A source shown next to an answer may have supported one statement without causing the recommendation. Model changes, prompt wording, source freshness, and product fit remain alternative explanations.
Still, the repeated split is useful. Tower 28’s proposition was legible as sensitive-skin-friendly makeup. Its skincare collection did not consistently become an answer to ordinary sensitive-skin routine questions.
Where we think the evidence chain breaks
Tower 28 already publishes ingredients, usage guidance, product claims, and sensitive-skin positioning. Its FAQ says the products are allergy- and dermatologist-tested, non-comedogenic, non-irritating, and fragrance-free. The SOS Barrier Recovery Cream page and SOS Gel Cleanser page provide routine instructions and exclusions such as added fragrance and essential oils.
The problem is not absence. It is the distance between a memorable brand promise and a cautious, reusable buyer decision.
| Evidence layer | What is already clear | What we would improve |
|---|---|---|
| Entity and category | Tower 28 is recognizable as sensitive-skin beauty | Make skincare, makeup, sunscreen, and body-care roles explicit without turning a brand promise into a medical conclusion |
| Owned evidence | Product pages expose ingredients, claims, and routines | Standardize the method, sample, duration, product version, and limitation behind consequential claims |
| External validation | Multiple products appear in the NEA directory and beauty publications test the brand | Keep acceptance, awards, editorial reviews, and clinical studies separate rather than blending them into one proof signal |
| Comparative evidence | Tower 28 is frequently evaluated as makeup | Build current cleanser, moisturizer, sunscreen, and routine comparisons with realistic drawbacks |
| Buyer-question alignment | Makeup for sensitive skin is a strong association | Connect skincare products to conservative routine-building questions where they are genuinely suitable |
This is an evidence-translation diagnosis, not medical advice and not a judgment about Tower 28’s product quality.
The five actions we would prioritize
1. Build one sensitive-skin decision system across the range
We would create a canonical guide that begins with the buyer’s job, not the brand’s collection names.
The guide should distinguish cleansing, moisturizing, sun protection, complexion, eye, lip, and body products. For every product, it should state the intended role, complete ingredient list, format, fragrance and essential-oil status, relevant testing, third-party acceptance if applicable, known use constraints, and the point at which a buyer should consult a clinician.
It should also explain what the evidence does not establish. A product being fragrance-free does not guarantee compatibility. A product appearing in the NEA directory does not diagnose, treat, or prevent eczema. A clinical study on one product does not validate the entire brand.
That restraint would make the system more trustworthy and more reusable in high-stakes recommendation contexts.
2. Pair branded names with literal buyer jobs
“SOS” is memorable to a person who already knows Tower 28. It is less descriptive to a system answering “Which basic cleanser and moisturizer should I consider?”
We would consistently pair every branded name with a literal role: gentle daily gel cleanser, barrier-supporting facial moisturizer, mineral facial sunscreen, serum foundation for sensitive skin, and so on. Page titles, headings, internal links, retailer feeds, media materials, image text, and structured data should agree on the product’s current name and job.
This is not an invitation to stuff pages with symptom language. It is a way to reduce entity ambiguity and help a buyer understand where a product belongs in a routine.
3. Create a versioned product-evidence ledger
Beauty claims change as formulas, shades, packaging, and study programs evolve. We would maintain a public, versioned evidence table for every material claim.
For a statement such as “94% of participants with sensitive skin agreed it did not break them out,” the useful evidence includes the sample size, recruitment criteria, study design, duration, product version, question wording, sponsor, and full result. “Derm-approved,” “clinically tested,” “hypoallergenic,” and “non-irritating” should each link to a precise definition and method.
The ledger would also separate company testing, third-party certification or acceptance, editorial testing, and consumer reviews. These sources answer different questions and should not be presented as interchangeable.
4. Build independent evidence beyond launch reviews
Tower 28 has substantial editorial coverage, especially for color products. We would add a more systematic body of current, product-specific testing around skincare routines.
That program could include independent dermatology-informed ingredient review, patch-use diaries with clear inclusion criteria, cleanser-and-moisturizer comparisons, sunscreen wear tests, makeup-over-skincare compatibility, and longer-horizon follow-ups. Reviewers should retain editorial control, disclose product provision and affiliate relationships, and report negative findings.
The goal would not be to engineer unanimous praise. It would be to make the conditions of success and failure legible: skin type, known sensitivities, concurrent products, climate, wear time, finish, and discontinuation reasons.
5. Run a segmented 90-day AEO experiment
We would keep the original four prompt groups as a frozen baseline, then add questions about sunscreen, complexion, eye makeup, a three-step routine, ingredient avoidance, price, shade needs, and a buyer already under dermatologic care.
We would repeat the panel across relevant consumer answer surfaces, preserve raw answers and citations, and code inclusion, order, assigned role, product accuracy, evidence type, and safety qualification separately. We would then retest after the category guide, evidence ledger, and updated third-party work become discoverable.
A meaningful improvement would not be Tower 28 appearing in every eczema-related answer. It would be more accurate inclusion in eligible situations, with the correct product, evidence, limitation, and recommendation role.
A practical 90-day plan
| Priority | Expected mechanism | Effort | First measurement | Signal that supports the hypothesis | Signal that weakens it |
|---|---|---|---|---|---|
| Sensitive-skin decision system | Connects products to routine jobs and explicit boundaries | Medium | 4–6 weeks | More correct inclusion in eligible cleanser, moisturizer, sunscreen, and makeup prompts | The guide is retrieved but roles remain inaccurate or unchanged |
| Literal product naming | Reduces ambiguity around branded collection names | Low–medium | 3–5 weeks | Answers use the correct product and routine role more often | Product-role errors persist after discovery and indexing |
| Evidence ledger | Makes claims, methods, and limitations inspectable | Medium–high | 6–10 weeks | Answers cite the appropriate product-level evidence without generalizing it | Claims remain unsupported or are still applied brand-wide |
| Independent routine tests | Adds corroboration outside brand-owned and launch content | High | 8–12+ weeks | Current independent sources recur for skincare as well as makeup | Coverage remains concentrated in launches or color products |
| Repeated prompt panel | Separates durable movement from model noise | Medium | Baseline, week 6, week 12 | Gains persist across two waves and more than one surface | Movement occurs only in isolated runs or unsafe contexts |
Before any content work, we would verify crawler access, canonicalization, rendered product data, retailer consistency, and indexability. OpenAI says pages that block OAI-SearchBot will not be shown as sources in ChatGPT search answers. Google’s guidance for AI features likewise begins with normal crawlability, indexing, site structure, and people-first content. Those are prerequisites, not recommendation levers by themselves.
What we would not do
- Make medical promises: Sensitive-skin positioning must not become an implied diagnosis, treatment claim, or universal safety guarantee.
- Generalize one seal: We would identify the exact products appearing in the NEA directory and preserve the association’s own limitations.
- Publish symptom pages at scale: A library of near-duplicate “best for” articles would create volume without establishing suitability.
- Hide adverse feedback: Texture, finish, packaging, shade, and individual-reaction complaints help define fit and research questions.
- Call one retest causal: Better visibility after publication would still be an observation with multiple plausible explanations.
The standard we would use to judge success
We would not set “20 out of 20” as the goal. In a health-adjacent category, indiscriminate inclusion could be a warning sign rather than a win.
Success would mean Tower 28 is retrieved more often when its exact products fit the buyer’s job; the answer distinguishes makeup, skincare, and sunscreen; product-level evidence is described accurately; and limitations travel with the positive claim. We would also expect the brand to retain its strong makeup association rather than dilute it into generic sensitive-skin language.
That is the larger AEO lesson. A distinctive proposition becomes useful to an answer engine only when the public evidence can support a specific, bounded recommendation.
Sources and methodology
The AI-visibility observation comes from four unbranded sensitive-skin questions, five fresh ChatGPT Temporary Chat runs per question, and 20 valid answers recorded in the United States on September 23, 2026. Web search was on, personalization was off, and interrupted runs were excluded and replaced. We coded brand inclusion, which target brand appeared first, assigned role, and visible source domains. We did not test other AI platforms in this panel or run a causal experiment.
Current first-party evidence was reviewed on September 23, 2026 from Tower 28’s FAQ, SOS Barrier Recovery Cream, SOS Gel Cleanser, SOS Spray, and Swipe Serum Foundation pages. Product, testing, and ingredient claims on those pages are company claims unless independently corroborated.
Independent context came from the National Eczema Association’s Tower 28 directory results and foundation entry, plus current coverage from Cosmopolitan, Marie Claire, and GQ. Commerce publications may use affiliate links; their findings were treated as attributed evidence, not proof of every company claim.
About this analysis
AgenQuest Research conducted this analysis independently using public information and AI responses observed on the dates stated. Tower 28 and Vanicream did not sponsor, review, or participate in it. AgenQuest offers AEO-related services. AI outputs are non-deterministic and may change. The findings describe the tested prompts and period, not every possible user experience or either company’s overall quality. Nothing in this article is medical advice.