Start with what you have: an RFP.Tell us what you want: an RFP response?
Bring an RFP, a BPD, an SoW, a requirements list, existing SOPs, or two release versions. The Neural Engine benchmarks every line against a vendor-neutral corpus of senior-architect-vetted implementation knowledge — telling you what's standard out-of-the-box, what's genuinely custom, what's missing, and where effort estimates drift before you commit.
Ways to start
Fit, gaps, and effort — line by line.
Drop in a statement of work or scope document. The Neural Engine benchmarks every line against the platform corpus and tells you what's standard out-of-the-box, what's genuinely custom and worth paying for, what's missing from the scope entirely, and where effort estimates drift from peer benchmarks — in either direction.
Standard vs. custom, before you negotiate.
Paste a requirements list or process spec. Every requirement is classified as standard (configurable out-of-box), custom (needs development), or absent from the corpus. Use the breakdown to shape what belongs in your SoW — or feed it directly into the Scope + SoW analysis for a fully grounded picture.
A single requirement is fine — you don't need a list.
From existing SOPs to RFQ requirements.
Drop in SOPs from your current system. Specforge maps each procedure to what the target platform handles natively versus what requires custom work. The output becomes requirements language for your RFQ — and a validation baseline to check any SoW you receive against it.
Have a BPD? See an independent assessment.
Drop in a business process design — your own, or one your integrator wrote for you. You get back the requirements it carries, laid out as a traceability matrix, and the parts of the process it has not covered. Each requirement is checked against the platform corpus: standard, configuration, or custom. Then the test cases and end-to-end scenarios that would prove the design, in one workbook. From there you can ask for the test execution to be automated, and for the test results to be validated from your SAP tables.
Sign in to run an assessment.
A freeform RFP, read as scope.
Drop in an RFP as it was issued — prose, tables, whatever arrived. Specforge pulls out the deliverables and requirements buried in it and classifies each against the corpus, so you can see what the RFP is actually asking for before you price a response to it.
What actually changed between releases.
Pick the release you run today and the one you're considering. Specforge lists what changed in between — new, changed, deprecated, removed, and what becomes mandatory — drawn from the release's own documentation, so the upgrade conversation starts from the delta rather than a summary of it.
See what your discovery notes become.
Requirement lists and process docs sit in separate files. Interface lists and accounting rules sit in others. Specforge connects them into one live map.
It runs from purchase order to financial statements. Every system, partner file and spreadsheet sits in the step where it happens. Badges mark every box that posts to the ledger or calculates a number for close.
Customer deductions
Every other tool answers
a different question.
There's no tool today that gives customers a corpus-grounded read on whether the scope they're being sold is right, complete, or padded. The existing tools all answer something else.
None of them tell a CIO whether a line labelled “custom” is actually configurable in a week on the current platform release, or where the scope is missing requirements the corpus considers standard, or where the effort estimates drift from peer benchmarks — in either direction.
The gap is structural. Customers don't have the platform depth to read a scope on their own — and no one is building that tool for them. Specforge is the corpus-grounded read that tells you what you're actually buying: what's standard, what's genuinely custom, and where the estimates don't hold up.
Signavio draws process steps. LeanIX lists systems. Specforge connects both to the partner files, spreadsheets, controls and ledger entries on one map.
A 14-year conviction. Built when the technology finally caught up.
The Specforge concept first appeared on LinkedIn in 2012 — a vision for a marketplace of reusable SAP delivery artifacts radiating from an intelligent core, packaging the institutional knowledge of every past engagement.
It was pitched at Deloitte, IBM, and PwC. None of them funded it. The technology to make it real didn't exist yet — large language models that could produce senior-architect-grade specs were still a decade away.
That changed in 2024. Specforge is the implementation of a thesis that was simply too early. The honeycomb is back, the engine is real, and every artifact is verified by a senior SAP architect before delivery.
Architect-verified, not AI alone
Every artifact is generated by a corpus-grounded engine and reviewed by a senior SAP architect before delivery. The reviewing architect's name, specialty, and years of experience are visible on every output. AI scales the production; human expertise guarantees the quality.