How Do the Three eIDAS Signature Levels Differ?

What are the three eIDAS levels?

Regulation (EU) No 910/2014 sets out three tiers, and each one raises the bar on identity proof and tamper evidence rather than changing what a signature is. SES is the floor. AdES adds cryptographic linkage to a named signer. QES adds supervised infrastructure behind that linkage, which is why only QES carries automatic equivalence.

The legal consequence of that top tier is written into the text. Under Article 25 of the Regulation, a qualified electronic signature has the equivalent legal effect of a handwritten signature, while Article 25(1) stops any electronic signature from being refused as evidence purely for being electronic.

So the choice isn't about legality. All three are admissible. What separates them is how much work you have to do in a dispute to prove who signed and that the file hasn't moved since. Picking a level is a decision about where you want that burden to sit, and who carries it.

SES provides basic electronic evidence

SES captures intent to sign and nothing more. No identity check is prescribed and no certificate is required. A typed name or a clicked checkbox qualifies.

The Regulation defines an electronic signature as data in electronic form attached to or logically associated with other data and used by the signatory to sign. That definition is deliberately wide, because the drafters wanted admissibility.

Which leaves the evidence problem with you. If a counterparty denies signing, your SES record is only as strong as whatever audit data your platform happened to collect around it. That's workable for internal approvals or low-value renewals where nobody disputes authorship. It gets thin fast when money or a regulator is involved.

AdES strengthens identity and integrity

AdES is defined by four conditions in Article 26, and a signature either meets all of them or it isn't advanced. The signature has to be uniquely linked to the signatory and capable of identifying them. It has to be created with data under their sole control and linked to the signed data so any later change is detectable, according to the European Union Agency for Cybersecurity (ENISA) overview of eIDAS standards.

The Regulation stays technology-neutral about how you satisfy those conditions. Public key cryptography is the usual route, but the certificate behind it doesn't have to be qualified, and identity proofing can range from a bank login to a video check.

That variation is the thing to watch. Two products can both call themselves AdES and produce evidence of different strength, because the gap sits in the identity proofing and key control. Ask your vendor what the certificate attests to and who verified it.

QES adds qualified trust services

QES is AdES plus two specific guarantees: a qualified certificate issued by a supervised provider, and a qualified electronic signature creation device (QSCD). Both elements are audited and listed, which is what turns the signature into something a court treats differently.

The European Commission's Digital Signature Service documentation describes the practical effect as a reversal of the burden of proof, with a presumption that the named person signed until proven otherwise.

Read that carefully, because it's the whole commercial argument for QES. You're buying a default position in litigation. The cost is friction, since the signer needs an identity credential they already hold or must obtain. QES also doesn't rescue a document from transaction-specific formalities. If a deed needs notarisation, QES won't substitute for the notary.

How do all three compare?

Here's the comparison in one view, with the caveat that AdES implementations vary while QES does not.

For public services, Commission Implementing Decision (EU) 2015/1506 requires Member States to recognise XAdES, CAdES and PAdES formats when they ask for advanced signatures. Format compliance buys you interoperability with public bodies. It says nothing about whether your identity evidence holds up, which is the part policies get wrong.

What does eIDAS actually establish?

eIDAS establishes what a signature level is and what legal effect follows from it without establishing that any given document is validly signed. Those are separate questions, and conflating them is the most common policy error inside operations teams.

The Regulation also draws a line most procurement documents blur. Since 1 July 2016, an electronic signature can only be used by a natural person to sign, per ENISA's guidance on qualified electronic seals. A company signs through a person. When an organisation itself needs to attest to a document's origin, the instrument is an electronic seal, governed by Article 36 rather than Article 26.

The practical consequence: your signature policy needs a second column for seals to cover invoices and bulk-issued documents where no individual is expressing consent. Teams that only map signature levels end up applying personal credentials to machine-generated output, which creates authority questions nobody wants to answer later.

How do trust services support signatures?

Trust services supply the external evidence that makes a signature verifiable by someone who wasn't there. Qualified trust service providers (QTSPs) issue the certificates, and timestamps fix the moment of signing. Validation services check all of it against published lists at the point of reliance.

The lists are the backbone. Each Member State publishes a trusted list of its supervised providers, and the Commission aggregates them into a List of Trusted Lists. As of April 2025, the Trusted List v6 format is enforced under ETSI TS 119 612 v2.4.1, which means validation tooling built on the older schema stops resolving correctly.

Keep two things apart in your documentation. eIDAS sets the legal requirements. ETSI specifications describe how to implement and validate them, and they change on their own schedule. My read, and this is inference rather than regulation: if your e-signing vendor can't tell you which ETSI version their validation stack tracks, your long-term evidence has an expiry date nobody has written down.

Do signatures work across borders?

Within the EU, QES travels. A qualified signature based on a certificate issued in one Member State must be recognised as qualified in every other, which removes the format argument from cross-border deals. AdES and SES get admissibility.

National contract law is where it stops being simple. Where German law requires the written form, Section 126a of the Bürgerliches Gesetzbuch permits electronic substitution only with a qualified signature, as a 2017 European Commission paper on e-signature platforms sets out. Some documents under German law still require a notary and can't be signed electronically at all.

So EU-wide recognition answers a narrower question than most policies assume. It confirms the signature is qualified. It doesn't confirm the agreement satisfies the form the governing law demands, and it says nothing about acceptance outside the EEA. Check form rules before you check formats.

Which level should each agreement use?

Use a test. Document lists go stale the moment you enter a new market or a new sector rule lands. A test survives because it asks the same questions about whatever lands in the queue.

Run each agreement type through three checks before you assign a level:

1. What happens if the signer denies signing, and what would you need to show?

2. Does any law or sector regulator prescribe a form for this document?

3. How long must the evidence stay verifiable after signing?

The amended Regulation nudges in a specific direction here. Member States are encouraged to consider advanced signatures for day-to-day transactions where they provide sufficient security and confidence.

Read that as a signal about proportionality. Defaulting everything to QES is a real cost in abandoned signing sessions, and defaulting everything to SES is a real cost in disputes you can't win. Route anything touching statutory written form or regulated transactions to legal review rather than guessing.

What identity evidence is necessary?

Match the verification to what an impersonator would gain. A low-value service renewal and a €4m share transfer carry different incentives, so they warrant different proof that the person at the keyboard is who the file says.

ETSI TR 119 460 treats identity proofing as a question of assurance level and notes that requiring a digital signature from an applicant is a highly relevant means of confirming identity, with policy requirements set on the certificate's quality level.

Signer authority deserves its own check, separate from identity. Verifying that Anna signed is not the same as verifying Anna could bind the company, and eIDAS proves the first while saying nothing about the second. Where mandate matters, pair the signature level with a board resolution or a registry extract in the same file.

What integrity evidence is necessary?

Decide how long the signature has to remain verifiable, because that drives the format more than the level does. Certificates expire and revocation data disappears, so a signature that validated in 2026 can fail validation in 2033 without anyone touching the document.

ETSI EN 319 142-1 defines four levels of PAdES baseline signatures with incremental requirements for maintaining validity over the long term, each building on the one below it.

If you have retention duties measured in decades for pensions or property, the long-term variants are the difference between an archive and a folder of unverifiable PDFs. Ask whether your platform applies timestamps and embeds validation data at signing, or whether it expects you to do it later.

Which laws govern the agreement?

Start with governing law, then check where the signer physically is, because both can impose form rules and they don't always agree. Sector regulation adds a third layer that contract law alone won't surface.

Employment law shows how sharp these rules get. The Mecklenburg-Vorpommern Regional Labour Court ruled in May 2023 that an agreement giving an employee the right to terminate early is subject to the written form requirement of Section 623 of the German Civil Code, as Deloitte Legal Germany reported.

Note what that case turned on. The issue was whether the document fell inside a statutory form category in the first place. Which means the jurisdiction question has to be answered before the signature question, every time. Build the governing-law check into your contract intake, and your signature level follows almost automatically.

How should this framework stay current?

Date-check your sources annually, because the eIDAS framework is mid-revision and the implementing acts are still landing. Separate what the Regulation requires from what your policy recommends, and label the second clearly so nobody mistakes internal preference for legal obligation.

Two deadlines matter right now. Regulation (EU) 2024/1183 requires every Member State to provide at least one European Digital Identity Wallet by the end of 2026, and wallet users will be able to sign with legal validity free of charge.

That second detail is the one worth planning around. When qualified signing stops being a paid credential and becomes a wallet function, the friction argument against QES weakens considerably, and policies built on "QES is too hard for signers" will need revisiting. Review this page against European Commission and ETSI publications each year. Narrower pages on QES and PAdES should link back here.

Can Agrello support your signing policy?

Yes, if what you need is signing workflows with identity-linked events and records you can show to someone who wasn't in the room. Agrello is a document management and e-signing platform built in Estonia, listed among the country's business solutions in the e-Estonia DigiExpo catalogue.

Before you compare vendors, write down what your policy actually requires. Which agreement types need verified identity. Which need long-term validation data embedded at signing. Which fall under a written-form rule that only QES satisfies.

Then bring that document to a conversation with Agrello and ask which signing methods are available for your signers' jurisdictions and how audit history is recorded. A format label is not a validity guarantee, and any vendor worth working with will tell you the same thing.