WHY PACKS
Three reasons to start
with a ready product.
A pack is not generic. It is a proven base you customize through configuration, not from-scratch engineering.
Speed-to-value
Go-live in 4-8 weeks, not 4-8 months. The base has been through dozens of deployments, so what is left is configuring it for your context.
Inherited maturity
MLOps, observability, governance, and quality gates all come ready. You do not build the infrastructure. You inherit what already runs in production.
Predictability
Scope, timeline, and cost are set before the contract. You know what arrives in week 4 and week 8, with no spec drift.
THE SIX PACKS
One for each
class of problem.
Each pack covers a set of recurring use cases. You choose by the problem, not the name.
BlueAssistant
Generative-AI customer service and support, grounded in your data and protocols.
Use cases
L1/L2 support · Internal onboarding · Corporate FAQ
BlueDocs
Document extraction, classification, and validation with multimodal NLU.
Use cases
KYC · Onboarding · Document back office
BlueConnect
Smart integrations across systems, APIs, and heterogeneous data sources.
Use cases
Migrations · Data fabric · Orchestration
BlueOps
Operations automation and mature MLOps in regulated environments.
Use cases
MLOps · Pipelines · Observability
BlueRisk
Real-time anomaly, fraud, and risk detection.
Use cases
Anti-fraud · Credit risk · AML
BlueDecision
Predictive and prescriptive models for business decisions.
Use cases
Forecasting · Pricing · Allocation
COMPARISON
Pack or Custom?
Four questions.
A pack moves faster when the problem is known. Custom wins when the problem is unique. Here is the practical difference:
HOW IT WORKS
Four phases.
From diagnostic to operation.
A fixed pattern across all 6 packs. Same cadence, same gates, same documentation.
Fit-check (1 wk)
A fit diagnostic: does the pack cover your case? Which modules? What customizations? Output: a sized proposal, or "this case is Custom".
Configuration (2-4 wk)
Pack setup in your environment, connection to your sources, rule and parameter configuration, and tuning of the industry templates.
Validation (1-2 wk)
End-to-end validation with real data, model tuning, quality gates, key-user training, and a rollout plan.
Go-live + Op
Production with an on-call squad for the first 4 weeks. After that, run it internally or as AI as a Service, your call.
WHEN NOT TO USE
A pack is not for everything.
Be honest.
Some cases are a bad fit for a pack. If any of the below sound like you, go straight to Custom.
- 01 A problem unique in the market, with no equivalent recurring case.
- 02 Highly proprietary integrations with non-standard legacy systems.
- 03 Regulatory risk modeling with a sector-specific framework that the templates do not cover.
- 04 An operation that requires a tech stack other than AWS for strategic reasons.
WHY BLUEMETRICS
Four things that set us apart
from off-the-shelf packs.
A pack is not a black box. Not off-the-shelf SaaS. Not a "closed platform".
The code is yours
A pack ships as software with source code, not off-the-shelf SaaS that locks in your operation. You can run it internally, with AaaS, or switch vendors.
White Box AI from day one
Every model decision is traceable. Logs, governance evidence, explainability. It matters in regulated sectors and helps in any sector.
Templates by industry
Not one generic pack configured three times. Specific templates for banking, healthcare, retail, and industry, built from real deployments.
Quarterly evolution roadmap
Clients get base improvements every quarter at no extra cost. The pack you get today is better than yesterday's.
FREQUENTLY ASKED
FAQ
It includes the base license, initial configuration, integration with up to 3 standard sources or systems, the applicable industry templates, end-to-end validation, key-user training, and squad support for the first 4 weeks after go-live. The exact scope is sized during the fit-check (1 week, no cost).
Yes, and that's the point. Customization happens through configuration (rules, parameters, templates), not by rewriting code. For customizations that go further, such as new integrations, custom models, or non-standard flows, we have an honest conversation: either we extend the pack as an additional project, or we tell you the case is pure Custom.
This is common. A pack covers 70-80%, and the rest becomes extension scope: a small custom project on top of the pack base, with the same squad. You pay for the delta, not the whole thing. If the delta is more than half the problem, that's a sign the case is pure Custom, and we'll say so.
You get the code for your configuration and any specific customizations. The pack base belongs to BlueMetrics, licensed for perpetual use. You're never trapped: run it internally, contract AaaS, or move to another vendor and take your configuration with you. The base is replaceable.
AWS is the technical default. We're an Advanced Partner and the pack base is tuned for SageMaker, Bedrock, and Textract. For clients with a stack consolidated on another cloud for strategic reasons, we handle portability during the fit-check or recommend Custom, depending on the effort involved.
The investment range depends on the pack, the industry, and your company size. The 1-week fit-check is free and ends with a sized proposal: scope, timeline, and investment spelled out before you sign. No surprises, no hidden consultant hours.