The DAX 40, tested by agents.
How well Germany's leading brands let AI agents succeed.
Where does your brand rank?
4 of 40 member brands ended this wave without a talent score: no posting lock could be carried through to a completed measurement on the brand's own career surface. Per-brand findings are in verification and publish with the talent-wave editorial.
The task, frozen 2 Jul 2026, is the same for every brand: find one live software or IT posting on the brand's own career surface and prepare the application. AI Usability is derived from the observed per-agent-class access profile and the deepest step reached (posting found, spec match, application ready), by the same fixed rule for all brands. No application was submitted anywhere. Brands without a linked page carry their measurement in the machine-readable index record until the brand page ships.
Agent Success Score = (AI Visibility × 0.20) + (AI Usability × 0.70) + (Evidence × 0.10), on a 0–100 scale, shown 0–10. The two axes are measured separately; Evidence is a confidence layer, not a third axis.
AI Visibility: Can an AI agent find the brand, its products, and current pricing?
AI Usability: Can an AI agent complete a real task on the brand's own surface: configure, quote, or order?
How we measure agent success
Every brand in this index is tested against a public methodology. Tasks are chosen against a Task Selection Doctrine before any measurement runs. Two axes are scored separately on a 0–100 scale: AI Visibility, audit-pipeline-derived, whether an agent finds and recommends the brand; and AI Usability, derived from the observed per-agent-class access profile, whether an agent can use the surface to get the job done. Evidence is a confidence layer. Third-Party Interception runs alongside as a separate axis. No surveys, no self-assessments, no vendor input.
Real tasks, not checklists
Each brand is tested on a concrete journey (find, price, act) chosen to reflect what an AI-powered customer would actually try. Sector grids are frozen and public before the wave runs.
Two axes, one composite
AI Visibility (×0.20) and AI Usability (×0.70) are measured separately, then combined into one composite on a 0–100 scale. Evidence (×0.10) is a confidence layer, not a third axis.
Frozen, public tasks
Each wave's task grid is locked and published before measurement. Brands can extend with their own tasks under the same fairness review. They cannot subtract what's already public.
Continental, Daimler Truck, and SAP are the latest three brand-owned names measured on AI Visibility. Three industries (tires, trucks, enterprise software), one pattern: the agent finds the brand, then the close walks out to a retailer, a dealer, or a system integrator. 36 of 36 brand-owned names are scored on both axes; 0 more carry a measured AI Visibility while their fleet usability run is queued (composite withheld until it lands); one is a terminal public-surface stall (Symrise) that no agent could resolve to a spec-fidel bergamot path.
Read the full story
Tech & Telecom · DTE
Every agent class reaches the cart; the gap is the unbranded query, where the discounters answer first.
Pharma & Healthcare · QIA
A near-flawless surface that mass-market AI rarely finds.
Logistics · DHL
The core shipping flow is agent-ready; checkout and B2B breadth are the open edges.
Industrials · IFX
A reference-grade agent surface that AI finds about four times in ten.
Industrials · SIE
Browser-class agents configure the controller and reach order-ready; the simpler agents keep the spec but stall on a generic portal shell, and the order itself sits behind a Siemens ID login.
Pharma & Healthcare · MRK
A reference-grade B2B catalogue that text and code agents can't reach at all.
Insurance & Finance · DBK
Found and form-ready across every agent class; the account opening stops at the KYC step.
Aerospace & Defense · MTX
On MTU's own pages, everything an agent needs is there in plain HTML: the PW1100G-JM, its A320neo application, the Hannover maintenance operation and a reachable B2B contact. But asked for GTF engine service without the brand name, an agent finds MTU only about half the time. The site is ahead of its own visibility.
Insurance & Finance · CBK
Found and form-ready across every agent class; the account opening stops at the KYC step.
Pharma & Healthcare · FME
A working dialysis-center finder that closes on a paper form, not a treatment intake.
Industrials · ENR
Siemens Energy owns the H-class efficiency story; EPC contractors own the project decision.
Pharma & Healthcare · FRE
A working booking portal that only browser agents can drive end to end.
Chemicals & Materials · SY1
Symrise fixed the stall: concrete bergamot products now exist on symrise.com, but only inside the structured data; the visible text never says bergamot.
Real Estate · VNA
On the unbranded question the portals capture the search; on the surface, the listing reads clean and the landlord-owned inquiry form is reachable by every measured agent class.
Insurance & Finance · HNR1
Hannover Rück owns the low-cost story. Reinsurance brokers own the placement.
Insurance & Finance · MUV2
Munich Re owns the capacity story; reinsurance brokers own the placement.
Industrials · G1A
GEA Group owns the Niro spray-drying story; process-engineering firms own the project specification.
Consumer & Retail · HEN3
Henkel is findable. The Persil Color Megaperls page is reachable, with the JETZT KAUFEN button, but the 40-load pack size (1.4 kg) and the 75 ml dosage do not show up on the product page.
Chemicals & Materials · BAS
AI agents find BASF, and the Glysantin datasheet is accurate. BASF's own site offers no way to buy, so the purchase runs through Amazon and the building-materials aisle.
Consumer & Retail · BEI
AI agents find Beiersdorf. The Eucerin Sun Sensitive Protect Face SPF 50+ page confirms the product claim, but the purchase runs through drugstores and pharmacies.
Automotive · VOW3
Volkswagen's Golf 8 Basis configurator works for browser agents at the deep path; coding and discovery agents land on the new Golf 2025 instead, a catalogue-drift story rather than a broken surface.
Industrials · SHL
Siemens Healthineers owns the MAGNETOM 3T story. The contract belongs to hospital procurement.
Automotive · MBG
Mercedes-Benz is machine-readable, but not yet transactable; agents read the configurator cleanly, then stall at a booking built only for humans, and the simplest agents never load the page.
Automotive · DTG
Daimler Truck owns the Actros story; truck dealers own the fleet quote.
Aerospace & Defense · RHM
Agents confirm the KF41, its tracked design and the Hungary order, and the inquiry route works. But the exact configuration, turret and combat weight, cannot be quoted from the public pages. And asked for a tracked infantry fighting vehicle without the brand name, agents reach Rheinmetall only about half the time.
Utilities · EOAN
On the unbranded question the comparison portals and the city utility capture the query; on the surface, only the coding and the autonomous agent reach the contract form.
Aerospace & Defense · AIR
AI agents read the H145 pages and reach the live sales form, but they never capture the full specification in one pass. Asked for a twin-engine EMS helicopter without the brand name, they reach Airbus just under half the time.
Automotive · CON
Continental owns the premium-tire story; tire retailers and comparison sites own the purchase.
Chemicals & Materials · HEI
AI agents can reach Heidelberg Materials' cement pages. But the purchase path for a 25kg sack of CEM I 42,5R (DIN EN 197-1) delivered in NRW slips toward Bauhaus and Hornbach listings instead of ending at the brand's own regional dealer portal.
Consumer & Retail · ZAL
Discovery is there, completion is not: HTTP hits an Akamai shell, the browser cannot reach an exact-variant bag publicly, and /checkout is session-gated.
Consumer & Retail · ADS
Akamai 403 at homepage entry across every measured agent class; no product, cart, or checkout surface becomes reachable.
Insurance & Finance · ALV
Comparison portals capture the category before allianz.de; the simplest agents never get through the door.
Pharma & Healthcare · BAYN
Everyone knows Aspirin. No AI agent can buy Aspirin Forte 1000mg from Bayer.
Tech & Telecom · SAP
SAP owns the S/4HANA story; system integrators own the deal.
Automotive · BMW
The €47,000 318i is configurable in a browser one time in three, and not at all by the simpler agent classes; Shadow DOM web components block every standard read path.
Chemicals & Materials · BNR
The buy-critical specs are not on the page, and agents rarely find Brenntag for them either.
Every brand has an Agent Surface, whether they built it or not. This index is its public deployment, applied to ourselves first. 36 brands open the methodology. The rest follow as the wave clock advances.
Listings platforms and exchanges are measured on the same probes as the brand-owned surfaces alongside them (Vonovia and Scout24 share the identical unbranded rental query), but their close-state mechanics are structurally different. They render here, separately, so the contrast is legible and the brand ranking stays comparing like with like.
Agent Success Score = (AI Visibility × 0.20) + (AI Usability × 0.70) + (Evidence × 0.10), on a 0–100 scale, shown 0–10. AI Visibility measures how discoverable a brand is to AI agents (audit-pipeline-derived). AI Usability measures how well agents complete a real task on the brand's own surface (derived from the observed access profile). Evidence is a confidence layer, not a third axis. Task Selection Doctrine published. Wave Q2 2026 Pilot.
Universe: DAX 40 composition as of 2026-07-01, reviewed after each Deutsche Boerse index review.
Data license: the DAX 40 Agent Success Index dataset is published under CC BY 4.0. Reuse and cite with attribution to Hyperize (hyperize.ai). Scores are shown 0–10, measured 0–100.
Can an agent finish the job on your brand?
Find out how AI agents experience your brand — and where revenue gets lost.
In the pilot wave?
Request the early Reference Page draft for your brand and see the Q3 audit task slate before it locks.
Request preview →Not in the wave yet?
Get a free Agent Success snapshot for your brand. Same methodology, off-cycle. Results in 48 hours.
Free snapshot →Editorial 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.
