Enterprise Integration in 2026: What Fixed-Bid Delivery Actually Looks Like

by Green Dolphin Software, Integration practice

Most enterprise integration consulting is sold on time-and-materials. The consultant benefits when scope expands. The buyer pays for discovery phases before the first line of code ships. We've watched this pattern fail too many Fortune 500 buyers, and it's why Green Dolphin Software was built differently.

Every Green Dolphin integration engagement is a fixed-bid Statement of Work, returned within 3 business days of the intake form. No T&M. No discovery-phase billing. No quietly-creeping invoice. The architect doing the SOW is the same architect writing the production code.

What a typical integration engagement looks like

The three-tier MuleSoft API pattern is the spine of most of our integration deliveries. Below is a synthetic but representative diagram — a $50K Standard tier engagement showing a Salesforce ↔ NetSuite bidirectional sync running on MuleSoft Anypoint with full observability and error handling.

Sample target-state architecture: Salesforce ↔ NetSuite bidirectional sync via MuleSoftThree-tier MuleSoft API architecture. Experience APIs at top, Process APIs in middle with canonical model, System APIs at bottom connecting to Salesforce and NetSuite. Anypoint MQ dead-letter queues, Datadog observability, and PagerDuty alerts shown.EXPERIENCEAPIsPROCESSAPIsSYSTEMAPIsSales PortalWeb / MobileSalesforce UISales CloudOperations PortalInternal adminProcess API — Canonical Model + OrchestrationAccount Sync↔ bidirectionalContact Sync↔ bidirectionalOrder ReplicationSF → NSInvoice ReplicationNS → SFSystem API — SalesforceREST API · Bulk API · Composite APIAccount · Contact · Opportunity · Invoice__cSystem API — NetSuiteSuiteTalk REST · SuiteScriptCustomer · Contact · Sales Order · InvoiceAnypoint MQDead-letter queuesReplay endpoint3 retries · 30s/5m/30mDatadogMetrics · logs · tracesPagerDutySev1/Sev2 alertsSynchronous flowAsync / event flowError / DLQ pathSynthetic example · not real client data

What you're seeing:

  • Experience APIs at the top — UI consumers (Salesforce, Operations Portal, Sales Portal) that don't know which back-end systems hold which records.
  • Process API in the middle — the canonical model and orchestration layer where business logic lives. Four flows: bidirectional Account sync, bidirectional Contact sync, SF→NetSuite Order replication, NetSuite→SF Invoice replication.
  • System APIs at the bottom — vendor-specific encapsulations of Salesforce REST/Bulk/Composite APIs and NetSuite SuiteTalk REST. Each System API exposes a clean, vendor-agnostic schema upward.
  • Anypoint MQ dead-letter queues on the right for error recovery with automatic retry (3 attempts, exponential backoff: 30s / 5m / 30m).
  • Datadog for metrics, logs, traces. PagerDuty for Sev 1/Sev 2 alerts.

This is what a procurement team sees when they ask "what are we actually buying?" — a documented, audit-ready architecture, not a slide deck.

The five service categories — and where Integration fits

Green Dolphin offers five service categories. Three are implementation services that use the same tier ladder (Starter $25K → Custom $100K+). Two are design-only services at $25K+:

  1. Integration — build new integrations (this article)
  2. Migration — move existing integrations from one platform to another
  3. AI / Agentic Solutions — Salesforce Agentforce, Agent Fusion, LLM integration, RAG, custom agents
  4. Architecture & Design — design package without build, for teams with build capacity
  5. Data Architecture — picks the right data platform + designs the data layer

Today's focus is Integration. The other four are covered in separate posts.

The four Integration tiers

$25,000 Starter (~3 weeks)

One bidirectional integration or up to 3 unidirectional flows sharing a pattern. Up to 4 data objects. Volume up to 25K records/day sustained. Most common fit: a new Salesforce ↔ NetSuite customer/order sync, or a single MuleSoft API exposing one internal system, or a Workato recipe set automating one RevOps workflow.

$50,000 Standard (~6 weeks)

Where most of our engagements actually land. Multi-system or bidirectional with a canonical model. Up to 3 systems, up to 8 objects, volume up to 100K records/day per object. Identity federation across multiple SaaS apps. Healthcare EDI 837/835 pipelines. The middle tier where complexity starts justifying canonical-model investment.

$75,000 Enterprise (~8 weeks)

Platform initiative scale. AI integration layer, multi-org migration scoping, Center for Enablement (C4E) governance establishment. Up to 5 systems, up to 15 objects, volume up to 250K records/day per object. Reusable accelerators and API templates. Architecture review board onboarding for the client team. This tier is where integration starts becoming a practice, not a project.

$100,000+ Custom (10-12+ weeks)

Regulated environments (HIPAA / FedRAMP / PCI-DSS), full platform migrations, multi-system rationalization where you're consolidating 5+ legacy integrations. Compliance documentation pack. Phased rollout with milestone gates. Embedded coaching for the client architecture team. Audit-trail logging and immutable evidence store integration.

What's included on every Integration engagement (Starter through Custom)

We documented this on our /pricing page but it's worth repeating because it shapes the buyer conversation:

  • Source code in Green Dolphin's GitHub during the engagement. At delivery acceptance, repositories transfer to the client's GitHub Org with full commit history. Client owns the IP outright.
  • Comprehensive README.md per repository covering architecture, prerequisites, setup, build, run, test, deploy, environment variables, runbook for ops, and troubleshooting.
  • Design package: high-level topology diagram, integration landscape, detailed sequence diagrams per business event, detailed per-API design (RAML 1.0 / OpenAPI 3.x with request/response schemas, error models). Editable in Lucid/Miro plus PDF export.
  • Unit test coverage ≥80% on application code. Coverage gate enforced. Report delivered with handover.
  • Postman collection covering every workable API endpoint with example requests, expected responses, and assertions. Used for both QA and post-deploy smoke tests.
  • Deploy to all environments — Sandbox + UAT + Prod (Starter is Sandbox + Prod only). Each deploy verified end-to-end via the Postman collection.
  • Post-deployment Postman tests in every environment as smoke/regression. Failures block promotion.
  • Error handling + structured logging: idempotent retry, dead-letter queues, structured JSON logs with correlation IDs. Logs flow to client's log aggregator (Datadog, Splunk, Elastic) or the native integration platform (Anypoint Monitoring, Boomi AtomSphere logs, Workato logs). Client picks.

What is NOT included (and where it goes instead)

The boring discipline of saying-no-to-things is what makes the fixed bid sustainable:

  • CI/CD pipeline build-out is covered by an optional Managed Services agreement, not the fixed bid. Implementation deploys into your existing CI/CD; standing up new CI/CD is a separate engineering scope.
  • Performance tuning beyond initial sizing — Managed Services.
  • License procurement for any third-party tooling — client-owned.
  • Data migration / cleansing / MDM beyond what the in-scope sync needs — separate engagement.
  • Custom UI development on connected systems (Lightning components, Suitelets) — out of scope by default.
  • DR / failover architecture — separate engagement.
  • Compliance certification of the delivered system (HIPAA / FedRAMP / PCI-DSS) — included only when called out at SOW signature for Custom tier.

Which platforms we deliver on

iPaaS: MuleSoft Anypoint, Workato, Dell Boomi, Informatica IICS, SnapLogic, Jitterbit, Celigo, Tray.io, TIBCO, Talend, Oracle Integration Cloud (OIC), SAP CPI, Apigee, Kong.

Salesforce: all Clouds (Sales, Service, Marketing Engagement, MCAE/Pardot, Commerce, CDP/Data Cloud, Health Cloud, Financial Services Cloud, Custom Force.com, plus Agentforce and Einstein).

ERP: NetSuite, SAP S/4HANA, SAP ECC, Oracle EBS, Workday.

Cloud-native: AWS (Lambda, EventBridge, Step Functions, AppFlow), Azure (Functions, Logic Apps, Service Bus, Data Factory), GCP (Cloud Functions, Workflows, Pub/Sub, Dataflow).

Custom code where it fits: Java (Spring Boot, Quarkus), .NET (ASP.NET Core), Node.js (NestJS, Fastify), Python (FastAPI, Flask).

If you're on a platform not listed, ask. The intake form takes 10 minutes regardless.

How communication runs during the engagement

Three channels:

  • Slack Connect (dedicated shared channel) — default day-to-day, daily async updates, design discussion
  • Email — formal status updates, change orders, written decisions; copies the Slack channel for archival
  • Phone / video call — kickoff, weekly check-in, design reviews, and major-decision discussions

Major decisions are always confirmed in writing. The Slack Connect channel from our greendolphinsoftware.slack.com workspace is the operational spine; email is the audit trail.

The 3-business-day SOW promise

You submit the intake form. Within 3 business days you receive:

  1. A fixed-bid Statement of Work with scope (must-haves and won't-haves), phase plan, acceptance criteria, assumptions, out-of-scope list, change-order policy, and fixed price.
  2. A target-state architecture diagram specific to your environment (not the synthetic one above — yours).
  3. A weekly milestone plan with explicit demo and review gates.

If you sign, 50% kicks off the engagement; the other 50% is due on delivery acceptance. Net 15. ACH or wire. The intake form, SOW writing, target-state diagram, and initial 30-min clarification call are all free.

Ready to scope?

If you have an integration that needs to ship and you've been burned by T&M consulting before, submit the 6-step intake form. Fixed-bid SOW in 3 business days. $25K floor, $25K increments. The architect you talk to is the architect who writes the code.

For a representative synthetic SOW + three target-state architecture diagrams (MuleSoft three-tier, Workato citizen-developer, Boomi hybrid runtime) plus four downloadable sample design deliverables drawn from real engagements, see /samples. Same depth you'd receive on an actual engagement.

More articles

AI Cost Optimization for Enterprise Workloads: Prompt Caching, Evaluation Frameworks, and the 80% Reduction Levers

Enterprise AI bills compound silently. The same workload that costs $4,200/month in November will hit $25,000/month by July without intervention. A vendor-neutral playbook for the five levers that produce 80% cost reduction without compromising quality: prompt caching, model tiering, response truncation, batch routing, and evaluation-driven optimization. Plus the audit framework that catches drift before invoices do.

Read more

MuleSoft Center for Enablement (C4E) Playbook: Crossing the 30-API Wall

Most enterprise MuleSoft estates hit a governance wall around their 30th API — not from technical limits, but from reuse, naming, and review drift. The Center for Enablement (C4E) framework is the answer. A vendor-neutral playbook covering the five C4E pillars (API standards, reusable assets, reuse model, security guardrails, Architecture Review Board cadence), what real C4E governance looks like vs vendor-mandated theater, and the $75K Enterprise tier we ship it with.

Read more

Ready to scope an integration?

Six-step intake. Fixed-bid SOW returned in 3 business days. $25K floor, $25K increments.

Office