Symrise.
Symrise fixed the stall: concrete bergamot products now exist on symrise.com, but only inside the structured data.
Found & recommended by AI agents
Can an agent use the surface to get the job done
This page measures two lanes. Commerce: can an agent find Symrise, and once it arrives, transact. Talent: can a candidate's agent find an open engineering role on Symrise's own career surface and prepare the application. After-sales, procurement, investor and press lanes run on different surfaces and are not yet measured.
Your brand isn't measured yet
What does an AI agent do with your website?
The same measurement as Symrise, free for your domain. Five agent classes, one real task, your score in 48 hours.
The stall is fixed. Machines can find Symrise's bergamot. Humans reading the page still cannot.
We test a real buyer scenario: a formulator needs a bergamot top note. The visibility pass asked three AI providers the question three ways: 18 measurements, all valid, zero errors.
The Q2 fleet test stalled: no agent could match the bergamot top note to a concrete product on symrise.com. The July 2026 re-probe overturns that. The Ingredient Finder now serves 213 products as server-rendered structured data. Seven carry bergamot in their descriptions, with SKU, CAS number and brand attached. The sample-request cart is one reachable link away.
One gap remains, and it is precise: the word bergamot exists only in the structured data. The visible page text and the odor-family filter say citrus. A coding agent finds the product; a text-reading agent walks past it.
One label, many breeds. From a plain reader to an autonomous operator, the kinds behind ChatGPT, Perplexity, and Claude Code:
Found, and able to transact?
Two questions, measured separately. A brand can be recommended and still un-buyable, or perfectly buyable and never found.
When someone asks an agent to source a bergamot top note for a fine-fragrance brief, does it route to Symrise?
Symrise comes up about a third of the time. The rest of the time, an agent recommends an alternative first.
Discoverability · 18-datapoint auditOnce an agent is on Symrise's site, can it request an ingredient sample?
Coding agent reach sample-ready; Plain reader and Search assistant get there only in some runs.
Browser-class agents reach sample-ready; text / search-class agents reach it only partially · breakdown aboveEvidence · 65 / 100 A measure of how provable and consistent the result is, grounded in cross-method ground-truth agreement (the methods that ran returned the same price), not a separate measured run. A confidence layer on the two scores above, not a third sales axis.
The talent lane.
A candidate's agent, sent to find an open software-engineering role at Symrise and prepare the application. Same brand, second surface: the career portal.
Closed to agents.
Every access class we can test without a browser hits the same wall: the Symrise career surface is a pure client-side single-page app, so the Cloud Engineer posting (req EA02813, Barcelona / Holzminden, Germany / Clichy) exists and is corroborated by three independent aggregator mirrors, but neither the job title nor a single requirement line is present in the raw HTML an agent without JavaScript execution would receive.
Furthest close-state reached: blocked · frozen task locked 2026-07-02
The score above is provisional where a class shows untested: untested classes are excluded from the calculation, so the pending check can only confirm or raise the value, never silently lower it.
AI Visibility · in measurement — the audit run for this lane has not landed yet. No Talent-lane composite is shown until both axes are measured.
What this does not yet cover
Not tested: form submission, login/account creation, and CV/cover-letter upload on the Talentlink apply flow — no login was attempted, no form was filled, and no application was ever submitted. The apply-entry URL itself returned HTTP 400 (missing site identifier) on a bare GET, so even login-requirement and artifact-requirement facts could not be read off without JS execution; this measurement only covers the HTTP/coding-class reachability of the posting and apply entry, not the full submission path.
Fairness note
This measurement covers automated access to Symrise's own career pages only. The tested Cloud Engineer posting (req EA02813, listed for Barcelona, Holzminden and Clichy) was live and independently corroborated on the lock date, and the low usability result reflects one specific property: the pages, served by the Talentlink recruiting platform Symrise uses, deliver no job content without JavaScript execution. Browser-based agents were not tested and are marked as such, no application was submitted and no forms were filled, and Symrise's presence on job boards such as LinkedIn or StepStone is recorded in a separate visibility measurement.
What this means for Symrise.
On specialty discovery, Symrise is not lacking product data. It now has one of the few server-rendered structured-data catalogues in the DAX 40, 213 products deep. What it lacks is the surface that translates that data into the buyer's words: the bergamot mapping lives in JSON-LD only, invisible to any agent that reads page text, and the odor filter speaks taxonomy (citrus) instead of the formulator's brief (bergamot top note). In a four-vendor market led by Givaudan, next to IFF and dsm-firmenich, the vendor whose ingredient answers match the question's vocabulary wins the shortlist.
What changes the outcome is closing the vocabulary gap between the catalogue and the question. The structured data exists; it needs the buyer's words on the visible surface. Named Hyperize offerings here: Answer Pages on the formulator's questions ("Which bergamot top note for a fine-fragrance brief?") that route to the concrete SKUs, and Reference Pages that lift the ingredient descriptions out of the JSON-LD into agent-readable text. If your catalogue speaks taxonomy and your buyers speak briefs, this is the test that decides whether agents find your products or a rival's.
The proof isn't more content. It's a repeatable lift in whether agents find, verify, cite, recommend, and complete the brand journey directly — re-measured each wave.
Commission an audit.
Where the BrandScore opens the question, an audit closes it. An interpretive engagement on your full surface, scored under the same methodology.
Get the auditFound with us.
Strategic partnership for brands building agent success as a long-term capability, not a one-off engagement.
ApplyAudit an adjacent property.
The BrandScore covers the primary domain. Get the same methodology applied to an adjacent property: a country site, a sub-brand, a category beyond the DAX-40 slate.
Get a SnapshotThis is Symrise. What about your brand?
You just saw how an AI agent treats a DAX 40 brand.
The same measurement runs free against your domain: five agent classes, one real buying task, your Agent Success Score in 48 hours, in the exact format of this page.
No intermediary stands between agents and Symrise. The gap is being found, not the channel.
B2B specialty fragrance has no displacing marketplace intermediary; the sample flow closes on Symrise's own portal. The competitive risk is direct vendor substitution (Givaudan, dsm-firmenich, IFF) rather than aggregator capture.
Hyperize-selected tasks.
One task from the public sector grid. Task list is frozen before each wave runs.
Symrise Bergamotte Top-Note fragrance-ingredient sample (cosmetic B2B)
- Close state
- sample-ready
- Bottleneck
- Concrete bergamot SKUs exist and the sample cart is reachable, but the bergamot mapping lives only in structured data; a text-reading agent cannot see it.
Fairness note
Wave 6 Chemicals + Materials visibility measurement plus an in-session usability re-probe (2026-07-21). Frozen task: Symrise Bergamotte Top-Note (sample_ready close, B2B parfumerie sample-request flow). AI Visibility from the audit pass (Q2-2026, openai/perplexity/anthropic, DE language, 18/18 valid datapoints). AI Usability derived from the in-session fleet phases (text + code) run 2026-07-21 against the frozen task; observed close state sample_ready. The Q2 stall verdict (bergamot never resolved to a concrete product) was falsified by this re-probe: the Ingredient Finder now server-renders concrete bergamot SKUs. Pending agent classes (browser/ACT) are excluded from the derivation (usability-derivation/v1.2). The sample-cart add was NOT attempted (NO-SUBMIT). Fairness Review pending the sector fairness grid.
How the score was produced.
Discoverability is audit-pipeline-derived. 3-provider sample (openai, perplexity, anthropic), 3 query variants per task, 2 runs per variant · 18 valid datapoints scored against a five-state handoff cascade. [S1]
AI Usability is derived from the access-profile above (usability-derivation/v1.2): how far the best agent reached (close state) modulated by how many agent classes succeeded. The per-class profile is the truth; the score is a reproducible summary of it, not a separate rating. Fleet phases (HTTP / Coding / Browser / ACT) produce the profile. [S2]
Agent Success Score = (AI Visibility × 0.20) + (AI Usability × 0.70) + (Evidence × 0.10)
On a 0–100 scale, displayed 0–10. AI Usability bundles the agent's reach + completion; AI Visibility is audit-derived discoverability. Weighting is public; the per-prompt derivation is not.
Measurement scope
Confidence C · one measured task on a 3-provider track (openai, perplexity, anthropic). Confidence promotes to B with a second task plus a fourth provider on the next wave.
Measurement timeline.
Each wave appends; nothing overwrites. Frozen Wave Rule.
- Entry · 01
25 May 2026
Wave · ProtocolWAVE-Q2-2026-W6-CHEMICALS
ars-methodology/v1.1
Wave 6 kickoff. AI Visibility audit dispatched. Agent-fleet run brief dispatched. State in_measurement until both data sources land.
- Entry · 02
21 Jul 2026
Wave · ProtocolUSAB-INSESSION-2026-07-21
ars-methodology/v1.1
In-session usability re-probe (text, code) falsifies the Q2 stall. In Q2 no agent could match the bergamot top note to a concrete product on symrise.com (phases 2-4 terminally deferred, surface_stalled). The 2026-07-21 re-probe found a changed surface: the Ingredient Finder server-renders 213 Product JSON-LD blocks, seven SKUs carry bergamot in their descriptions (with SKU/CAS/brand), and the sample-request cart is GET-reachable. text=partial (bergamot only in JSON-LD, visible odor filter says citrus), code=success. Stall lifted; brand flipped in_measurement -> scored. Observed close: sample_ready. Usability derived (usability-derivation/v1.2). The sample-cart add is JS/AJAX and was NOT attempted (NO-SUBMIT).
Evidence and provenance.
Public methodology references and internal evidence pointers behind every claim above.
- [S1]Accessed · 25 May 2026
AI Visibility audit · Symrise Wave Q2 2026 (dispatched 2026-05-25)
Internal · Hyperize evidence
- · AI Visibility score
- · the close state reached (sample_ready)
- [S2]Accessed · 21 Jul 2026
In-session fleet re-probe · chemicals · Symrise text + code (2026-07-21)
Internal · Hyperize evidence
- · the per-class access profile
- · the falsification of the Q2 stall verdict
- · the close state observed (sample_ready)
- Accessed · 25 May 2026
Public · hyperize.ai
- · fairness declaration
- · Third-Party Interception framing
Read the doctrine. Challenge the score. Extend the slate.
Task Selection.
The fairness doctrine behind the slate above. Five failure modes, six criteria, public before each wave.
Read ChallengeDisagree with this score.
Send evidence under public Fairness Review. Failed reviews are documented with the named failure mode.
Challenge ExtendSubmit your own task.
Open Surface Run · additive measurement. The Hyperize-selected slate stays frozen; your task gets the same methodology.
SubmitEditorial coverage
The DAX 40 Agent Success Index is a point-in-time snapshot of the agent-success of public digital touchpoints. Results are not statements about product quality, company performance, service quality, or the legal obligations of the brands named. Brand names and logos remain the property of their respective owners and are used solely for identification and reporting purposes in the context of editorial coverage (§ 23 MarkenG, Art. 5 GG).
Brands wishing to respond, engage, or correct a factual error may contact hello@hyperize.ai. Responses received are published in full alongside the findings. Full methodology and editorial-coverage notice: coverage statement.
