✳ INITIAL CUSTOMER FOCUS / B2B SAAS

Support decisions your team can test before release.

Forgedata’s initial product focus is the engineering team adapting an open model to its SaaS support workflow. The first local reference covers structured ticket triage. Customer demand and production value remain hypotheses to validate.

THE TEAM WE AIM TO SERVE

A support operations lead working with a product or AI engineer.

The team has approved, resolved tickets and needs consistent routing, escalation, or structured outputs inside its existing support application.

This is an initial customer profile, not a claim of current customers, interviews, or pilot commitments.

Discuss your support workflow
AN EXAMPLE OF THE PROBLEM
“Someone we do not recognise downloaded our customer information.”

A valid JSON response can still make the wrong decision. In the local reference run, the adapter classified this synthetic security ticket as an access problem. Evaluation exposed that failure and blocked promotion.

Inspect the actual failed example
THE PROPOSED SOFTWARE VALUE

A workflow around the model, not just another training command.

The reference uses Transformers and PEFT. The product direction is the support-specific preparation, policy tests, and traceability around those tools.

Test the support policy

Measure exact output format and routing decisions. Make a failed acceptance check visible instead of relying on a lower training loss.

Trace a candidate to its data

Keep dataset fingerprints, a pinned base model, training settings, evaluation outputs, and saved-adapter checks together.

Keep the artifact portable

Use standard PEFT adapters and reproducible instructions. Target deployment compatibility must still be tested for each model and runtime.

IMPLEMENTED REFERENCE AND PRODUCT DIRECTION

What exists today.

These distinctions describe the current work, so a software roadmap is not mistaken for an operational SaaS backend.

CapabilityCurrent evidenceRemaining product work
Dataset preparationSynthetic JSONL, schema checks, exact deduplication, fixed splitsCustomer data permissions, import UX, version management
LoRA adaptationCompleted local run, saved adapter, reload verificationHosted job orchestration, access control, NVIDIA execution
Policy evaluationAll baseline/candidate predictions and a failed promotion gateCustomer-defined acceptance criteria and broader evaluation
DeploymentStandard local adapter artifactHosted inference endpoint, monitoring, licence and compatibility review
CHOOSING THE RIGHT APPROACH

Adaptation has to earn its place.

A strong prompt is the first baseline.

If a prompt already meets quality, cost, and maintenance requirements, a training run may add unnecessary complexity.

Retrieval supplies changing facts.

Current product documentation and account facts belong in an authorised retrieval or application layer. Fine-tuning does not replace current knowledge or access control.

Libraries provide the training building blocks.

The intended Forgedata contribution is coordination and repeatable policy evaluation. We have not demonstrated superiority over competing training platforms.

CUSTOMER VALIDATION IS NEXT

The questions a pilot needs to answer.

We plan to test this direction with SaaS teams before claiming operational impact or paid demand.

  • Which routing or escalation decisions actually need improvement?
  • Does adaptation improve on the team’s prompt and retrieval baselines?
  • What data permissions, security review, and deployment requirements apply?
  • Which acceptance criteria and buying process would justify a pilot?
YOUR NEXT CHAPTER IN AI

From your first dataset.
To your next release.

Bring a use case. Explore a model workflow built for your team.

Request access