How we engage

The business process operating layer for enterprise AI agents

Multipliers comes with 192 pre-built vertical AI agents, remixable to your specific enterprise needs. Our Forward Deployment Engineers wire it into your existing AI substrate, business systems, and governance — so your team operates the fleet from day one.

Why FDE-led

Why coordination is the hardest part.

Most enterprise AI projects stall at the same point: the software is installed, but no real work is flowing through it. The gap isn’t technology — it’s coordination.

Getting an agent into production needs three people to align. Your business sponsor defines what success looks like. Your process owner knows how the work actually flows day-to-day, with all its edge cases. Your IT lead owns systems access, security, and governance handoff. No one person has all three sets of expertise, and asking one team to drive the whole thing is why pilots get stuck.

Our Forward Deployment Engineer is the dedicated bridge across those three roles. They know the platform inside-out so wiring is fast, they’ve done this before so they ask the right questions early, and they translate between business and IT so no one waits on someone else to finish first.

Engagement model

Four stages, from first conversation to operational fleet.

Each stage has a defined outcome, a defined deliverable, and a clear handoff to the next. You can stop after any stage and the value produced is durable.

Stage 1

Discovery & Scoping

2-3 weeks

Stage 2

Spec & Pilot Setup

3-4 weeks

Stage 3

Pilot Run

4-8 weeks

Stage 4

Operationalize

Ongoing

Stage 1 · 2-3 weeks

Discovery & Scoping

Outcome

Pilot scope locked, success criteria signed off, FDE has access to the systems we’ll wire.

What we do

  • Process inventory
  • System access mapping
  • ROI hypothesis
  • Pilot proposal

What you bring

  • Business sponsor 1-2 hrs/week
  • 1 process owner per pilot process
  • IT lead engaged in week 2 for access reviews

DeliverableScoped pilot proposal + Private Offer pricing (if procuring via marketplace).

Stage 2 · 3-4 weeks

Spec & Pilot Setup

Outcome

Pilot agents activated in your tenant, connector wiring complete, sandbox testing green.

What we do

  • Connector configuration (Salesforce / ServiceNow / SAP / Shopify / SFCC / etc.)
  • Spec-card review with your team
  • Sandbox runs with synthetic data
  • Governance dashboard tour

What you bring

  • 1-day FDE workshop with your process owner
  • IT review of OAuth scopes + data access

DeliverableActivated agents in shadow mode, ready for live pilot.

Stage 3 · 4-8 weeks

Pilot Run

Outcome

Agents running against your real systems, ROI measured, iteration in place.

What we do

  • Live execution monitoring
  • Weekly business review with sponsor
  • Iteration on prompts / policies / guardrails based on real-world results
  • Shadow-mode comparator validating ≥90% match rate before live writes

What you bring

  • Weekly 1-hour review
  • Escalation path to subject matter experts when agents flag edge cases

DeliverableShadow → testing → active publish gate cleared; first wave of measured ROI in value-chain dashboard.

Stage 4 · Ongoing

Operationalize

Outcome

Full agent fleet operational with governance + training transferred to your team.

What we do

  • Handoff governance ownership to your IT lead
  • Train your process owners on storefront, governance, and metrics dashboards
  • Monthly health reviews
  • Expansion plays (more agents, more processes, more sponsors)

What you bring

  • Continued sponsor engagement
  • Willingness to expand scope as confidence builds

DeliverableProduction-operational AI agent fleet your team owns + measurable enterprise KPI movement.

Who you work with

Three roles on our side. Three on yours.

The engagement runs as a tight six-person operating team. Any missing role from either side slows everything down — so we name them upfront.

On our side

  • Forward Deployment Engineer (FDE)

    Senior engineer dedicated to your account; owns wiring + activation + iteration.

  • Solutions Architect

    Process expertise; helps shape the pilot scope and identifies expansion plays.

  • Customer Success

    Ongoing partnership beyond pilot; quarterly business reviews.

On your side (we expect)

  • Business Sponsor

    VP-level outcome owner; defines success criteria; available 1-2 hours/week.

  • Process Owner

    Knows the day-to-day workflow we’re automating; available for 1-day workshop + weekly reviews.

  • IT Lead

    Owns system access, security review, governance handoff; engaged weeks 2-3 and at operationalize stage.

What we bring

A pre-built library, a wiring platform, and structural governance.

You’re not starting from a blank canvas. You’re not paying for a code shop. You’re not bolting on governance after the fact.

A pre-built library

192 catalog agent templates · 39 named process bundles (Phase 5 + 6) covering Order Orchestration, Billing Exception, Channel Enablement, Field Service, Renewal Intelligence, Promo Launch, Back-in-Stock, B2B Onboarding, and more. You’re not starting from a blank canvas.

A wiring platform, not a code shop

Per-client variation lives as data (connector configs, agent specs, tenant policies). Same codebase serves every customer. Onboarding a new system = inserting a row, not writing code.

Governance built in

Every agent born compliant with our 5-contract runtime: MCP server · A2A AgentCard · MVE outcomes · Context Engine integration · Value-chain bindings. Your governance dashboard surfaces the same data we use to ship agents to production.

Substrate-agnostic

Where compute runs.

When you procure Multipliers — whether direct, via AWS Marketplace, Azure Marketplace, or Google Cloud Marketplace — you control where your compute runs. We’re substrate-agnostic by design. An agent isn’t a single deployable thing; it’s distributed compute across four layers, and each layer can be routed to your cloud independently of how you procure us.

Layer 1

LLM inference

Routes to your chosen substrate via per-tenant config. Bedrock for AWS-procured customers, Vertex AI for GCP, Azure AI Foundry for Azure, or direct lab APIs. Your model usage shows up on your cloud bill; counts toward your EDP / MACC / committed-spend.

Layer 2

Tool execution

When an agent calls a custom tool (e.g. “lookup account health”), the tool runs wherever you host it. Customer-built tools deployed as MCP servers on AWS Lambda, Azure Functions, or Cloud Run hit your bill, not ours.

Layer 3

Sidecar compute

Specialized workloads (audio transcription, time-series forecasting, image processing) route to your cloud’s services (AWS Transcribe, Azure ML, Vertex AI endpoints) via per-tenant configuration.

Layer 4

Orchestrator loop

The ~2-second-per-run coordinator runs on our SaaS by default. For customers requiring full VPC deployment (data residency, regulated industries), we ship a container + Helm chart that runs in your EKS / AKS / GKE.

Layer

Marketplace procurement

Bill via cloud

Substrate routing

Per-tenant config

Full VPC self-host

Regulated

LLM inference

Bedrock / Vertex / Foundry per cloud
Per-tenant model picker
Same — substrate of your choice

Tool execution

Your MCP servers on your cloud
Per-tool routing rules
Same — runs in your network

Sidecar compute

AWS Transcribe / Azure ML / Vertex endpoints
Per-tenant sidecar adapter
Same — your cloud services

Orchestrator loop

Our SaaS control plane
Our SaaS control plane
Helm-deployed in your EKS / AKS / GKE

The result: an AWS-procured customer can have ~70-80% of agent compute consumed against their AWS bill (Bedrock + Lambda + Transcribe), without leaving our SaaS control plane. A regulated customer can have 100% of compute inside their VPC by deploying the orchestrator container. Same codebase, same agent library, different routing decisions.

What we don’t do

Honest about scope.

  • We’re not a model lab

    We don’t train or host LLMs. We use the model substrate you’ve already invested in.

  • We’re not a per-client custom code shop

    Every customer runs the same codebase. Variation lives in configuration data, not branches.

  • We don’t replace your cloud

    We extend it. Your AWS / Azure / GCP investment becomes more valuable, not duplicated.

  • We don’t ship AI strategy consulting

    We ship operational agents. Strategy is your CTO’s call; execution is ours.

Comfort signals

How we’re built for enterprise from day one.

Multi-tenant from day one

Cross-tenant isolation enforced structurally (every query, every API call, every audit row). Not bolted on.

Cross-stack OAuth runtime

Unified OAuth across all integrated platforms with PKCE, refresh coordination, and audit-trail uniformity. We handle credential lifecycle so you don’t.

End-to-end value attribution

Every agent run rolls up through process metrics to enterprise KPIs. The CFO sees revenue impact, not tokens consumed.

Built for enterprise security review

Single-tenant data isolation enforced structurally (every query layer asserted in tests); cross-stack audit events on every state change with PII-clean payloads; idempotent writes everywhere; documented data flows and configurable retention. Ready for your SOC 2 / ISO 27001 / HIPAA security questionnaire.

At a glance

Roles & timeline summary.

Stage

Timeline

Multipliers role

Your role

Outcome

Discovery

2-3 weeks
FDE + SA
Sponsor + Process Owner
Scoped pilot proposal

Spec & Setup

3-4 weeks
FDE + SA
Process Owner + IT Lead
Agents activated in shadow

Pilot Run

4-8 weeks
FDE + CS
Sponsor + Process Owner
Publish gate cleared; first ROI measured

Operationalize

Ongoing
CS
IT Lead + Process Owners
Production fleet your team owns

Ready to scope a pilot?

Email us at contact@kiwana.ai with a brief on the process you’d like to automate and your existing AI substrate (Bedrock, Vertex AI, Azure AI Foundry, or direct). Our solutions architect will respond within 2 business days with a discovery call calendar link.