Clean documents out. Structured evaluation in.
From the statement of work to the award record — on the authority side.
prüfr reads your draft statement of work and tender documents and shows where requirements are unclear, incomplete or contradictory — so you publish a cleaner tender and reduce the risk of queries, complaints and cancellations. After submission, prüfr supports the evaluation: qualification check, price check, scoring by your award criteria, and a structured award record with separated evaluation and approval. Constructive, authority-side. EU-hosted on Open Telekom Cloud in Frankfurt.
What does prüfr do?
prüfr supports contracting authorities from tender documents to award: check the specification, anticipate bidder questions, structure the evaluation of bids and prepare the award record. The award decision remains with the contracting authority. Before publication, prüfr checks the draft systematically for clarity, completeness and internal consistency — in the spirit of a clear and exhaustive specification per VOB/A §7 — and cross-checks main text, annexes and criteria against each other. After submission, prüfr is oriented to the steps of the legally prescribed evaluation process: qualification check, price check, scoring by award criteria and the award record. Every AI pre-assessment is confirmed or adjusted by a human; evaluation and approval are separated. prüfr is a standalone product with its own database and shares no data with auftragr — bidder data on the authority side stays on the authority side.
Who is prüfr built for?
For contracting authorities in federal, state and municipal government and in the utilities sector that want to sharpen their documents before publication and finish the evaluation cleanly documented.
Municipal contracting authorities
Limited capacity, many parallel procedures. From the review of the statement of work through to the award record — without a consulting budget for every procedure.
Requesting departments
The department writes the specification, the procurement office runs the procedure. prüfr helps both phrase requirements clearly and traceably.
Federal and state agencies
A clear, exhaustive specification is mandatory — and after that a weighted scoring matrix with traceable reasoning for every individual score.
Utilities-sector authorities
SektVO-specific procedures, complex technical specifications, long contract durations. Consistency across many positions, annexes and parallel procedures.
Central purchasing bodies
Standardised quality and scoring across all procedure types. One shared method per procedure, audit trail for internal review.
Public hospitals & municipal utilities
Medical technology, energy infrastructure, specialised services. Price checking and mixed-calculation detection at technical depth.
Procurement-law advisors
External advisors use prüfr as a shared workspace with their client — while preparing the documents and during the evaluation.
How prüfr works
From draft tender documents to a prioritised checklist — and from bid opening to a documented award record.
- 1
Upload the draft
Statement of work, eligibility and award criteria, other tender documents — as PDF, Word or GAEB. Still unpublished, in draft.
- 2
AI reads the documents
prüfr checks the requirements for clarity, completeness and consistency — positions against criteria, annexes against the main text, deadlines against the procedure.
- 3
You receive a checklist
Unclear wording, missing details, contradictory requirements and possible breaches of the clarity requirement — each with a source location and a concrete drafting suggestion.
- 4
You sharpen and publish
Every note is traceable to the source in the original document. You accept, adjust or discard — the decision stays with the contracting authority.
- 5
Qualification check
Bidder qualification against your eligibility criteria. Detection of exclusion grounds per §§ 123/124 GWB. Use of other entities' capacities and consortia are shown separately.
- 6
Price check
Comparison of all bid prices, detection of mixed calculation and abnormally low bids. Clarification needs are flagged — with draft wording for the query.
- 7
Scoring by award criteria
Evaluation by your award criteria with weighted scoring. Transparent reasoning for every individual score — with source evidence from the bid documents.
- 8
Award record
Structured award record with the usual mandatory fields. Separated roles: the approver and the evaluator are different people. Every override is logged with reasoning.
What prüfr does
Pre-publication quality assurance and post-submission evaluation — in one system.
Clarity check
Detects vague, open-to-interpretation or ambiguous requirements in the statement of work — in line with the clarity requirement of VOB/A §7.
Completeness check
Flags missing details: quantities without units, criteria without weighting, services without acceptance conditions.
Contradiction detection
Compares main text, annexes and criteria against each other and surfaces contradictions before bidders find them.
Anticipate bidder questions
Points to places that are ambiguous, incomplete or contradictory and therefore invite queries — so you can clarify them in advance.
Criteria consistency
Checks whether eligibility and award criteria are clear, measurable and consistent with the statement of work.
Drafting suggestions
For each weak spot, a concrete, neutral suggestion to sharpen it — as assistance, not a decision.
Scoring matrix
Your award criteria with weighting. Weighted scoring per bid, individual scores with reasoning.
Exclusion-grounds check
Check per §§ 123/124 GWB. Capacity borrowing and bidding consortia are shown separately in the evaluation.
Mixed-calculation detection
Automatic detection of suspicious pricing patterns in bills of quantities. Flagged for further clarification.
Clarification flagging
Abnormally low bids, contradictory documents, missing evidence — with draft wording for the query to the bidder.
Separated roles
Approver ≠ evaluator. Overrides are logged with reasoning.
Structured award record
Record with the usual mandatory fields, source evidence per individual score. Export as PDF and structured document.
EU-hosted, flexible models
EU-hosted on Open Telekom Cloud in Frankfurt. Language models from various providers incl. European models — exclusively with processing inside the EU.
Compared to current contracting-authority work
Where prüfr removes time and risk from the procedure — and where responsibility stays with the contracting authority.
| — | Today | With prüfr |
|---|---|---|
| Document review | Manual, dependent on capacity and experience | Structured checklist as a basis for the review |
| Clarity of requirements | Ambiguities often surface only through bidder questions | Ambiguous wording flagged in advance |
| Contradictions across annexes | Hard to find manually across many documents | Automatic cross-check of main text, annexes and criteria |
| Time per award record | Manual: binds scarce resources in the authority | Structured pre-assessment + documented review |
| Documentation gaps | A single missing piece of evidence can kill the award | Mandatory fields per stage, no stage without documentation |
| Consistency across procedures | Depends on evaluator and current workload | Every procedure evaluated with the same framework |
| Traceability of scoring | Reasoning is often only reconstructable after a query | Record with source evidence per individual score |
| Decision authority | Stays with the contracting authority | Stays with the contracting authority. AI structures. Humans decide. |
Frequently Asked Questions
What stage is prüfr at?
prüfr is an MVP and open to first pilot and design partners. There is no public price list. We work with a few contracting authorities that help shape the scope — register pilot interest at blackswanai.de/en/pilot-interest.
Does prüfr replace procurement officers or legal advisors?
No. prüfr structures and documents the review and the evaluation process — the decision stays with your staff. AI structures. Humans decide. Every AI pre-assessment is confirmed or adjusted by a human.
Does prüfr check the legal soundness of the documents?
prüfr surfaces weak spots in clarity, completeness and consistency and is oriented to the process and documentation steps of VgV. It does not replace the professional and legal review by your contracting authority or procurement lawyer — that remains necessary and your responsibility. We make no certification or conformity promises.
How are evaluation and approval separated?
The approver of the award record is a different person than the evaluator. If an approver overrides an AI pre-assessment, the change is logged with reasoning — traceable.
Which documents does prüfr read?
Statements of work, eligibility and award criteria and further tender documents in the usual formats (PDF, Word, GAEB) — as drafts before publication and as incoming bids after submission.
Which procedure types does prüfr support?
Open procedure, restricted procedure, negotiated procedure with prior competition, innovation partnership — the VgV procedures are supported. UVgO procedures below the EU threshold as well. SektVO utilities procedures with separate configuration.
Can prüfr handle alternative bids?
Yes. Alternative bids are evaluated separately and shown in the scoring matrix. The decision whether an alternative bid is admissible and how it is weighted stays with the authority.
How does prüfr integrate with our existing procurement platforms?
prüfr imports documents and bids from the usual German procurement platforms (DTVP, Deutsche eVergabe, Vergabe24, state platforms). Export of the finished award record as PDF and structured document. Direct API integration into specific platforms is available on request.
How does onboarding work?
The base configuration covers your award criteria, weighting and approval workflow. For complex tenant structures (multiple authorities in one organisation) there is configuration support. We discuss specifics in the pilot conversation.
How are our drafts and bid data handled?
Processing runs on EU infrastructure (Open Telekom Cloud). AI processing also takes place exclusively inside the EU — we do not use language-model providers outside the EU; the subprocessors used are named in our Terms and Privacy Policy. prüfr has its own database and shares no data with auftragr. We clarify the details in the pilot conversation.
The other product
prüfr covers the authority side — creating and awarding. The bidder side has a standalone product, with its own database and no coupling to prüfr:
Help shape prüfr
prüfr is an MVP and open to first pilot and design partners. Register your pilot interest — a 30-minute conversation will tell you whether it fits your authority.
AI structures. Humans decide.