Independent Salesforce & AI Engineering

We architect Salesforce.
We build the AI that runs on top of it.

A small studio of senior engineers designing CRM architecture, secure integration, validated data migrations, and enterprise AI applications — with a deep focus on life sciences. The people you meet in the first call are the people who write the code.

Book an Architecture Review → See what we build

No junior hand-offs · No reseller incentives · Senior architects only

Reference architecture
Legacy CRM On-premise ERP & LIMS SAP · Oracle Documents PDF · Scans · Files Migration & integration layer Profiling · Mapping · Reconciliation · Audit trail Salesforce core Sales Service Experience Data 360 Apex · LWC · OmniStudio · Platform events · SSO & mTLS AI application layer Document AI Service agents Process agents Grounded on your data · Human in the loop · Fully logged
25 Combined Salesforce certifications across the team
15 yrs Average enterprise experience per engineer
4 clouds Sales · Service · Experience · Data 360
100% Senior engineers on every engagement
Capabilities

Two engines, one architecture.

Most firms do platform work or AI work. The interesting problems live where they meet — an agent is only as good as the data model underneath it, and a CRM is only as valuable as the work it takes off people's desks.

Salesforce Platform Engineering

Architecture, security and heavy custom development — well past what configuration and low-code can reach.

  • Core CRM architecture Sales and Service Cloud designed around your actual process, not a demo org. Object models, sharing, automation strategy, and a governance model your admins can live with.
  • OmniStudio guided experiences OmniScripts, FlexCards, Integration Procedures and DataRaptors for complex guided processes — built as maintainable components rather than the sprawl these projects usually become.
  • Identity, SSO & secure integration Single sign-on with third-party applications over SAML 2.0 and OpenID Connect, Connected Apps and OAuth 2.0 JWT bearer flows, SCIM user provisioning, Named and External Credentials, and mutual TLS with full certificate lifecycle management where the counterparty requires it.
  • Experience Cloud portals Partner, customer and patient-facing sites on Lightning Web Components, with the external identity, licensing and sharing architecture done properly the first time.
  • Data 360 & unified data models Harmonising fragmented sources into one addressable profile — the prerequisite for any AI that has to be right, not just fluent.
  • Validated migration & large data volumes High-stakes migrations with profiling, field-level mapping, automated reconciliation and an audit trail that survives inspection — plus Big Objects, selective indexing and archiving strategies for orgs that have outgrown their storage model.
  • Custom development & DevOps Apex, LWC, Platform Events, Change Data Capture and Bulk API 2.0, delivered through SFDX source control and CI/CD — with Salesforce Shield encryption, event monitoring and field audit trail where compliance calls for it.
Sales Cloud Service Cloud Experience Cloud Data 360 OmniStudio SAML / OIDC SSO mTLS & certificates MuleSoft Shield Apex / LWC

Enterprise AI Applications

Named, scoped AI systems that do a specific job inside your operation — grounded on your data, measured, and logged end to end.

  • Document intelligence Extraction, classification and semantic search over dense operational and regulatory documents. Answers that cite the source page, so a reviewer can verify rather than trust.
  • Migration validation agent Automated reconciliation across source and target during a migration — anomaly detection, exception queues and a generated evidence pack instead of a month of manual sampling.
  • Service & process agents Tier-one resolution and back-office automation on Agentforce 360 or your own model stack, using hybrid reasoning so deterministic business rules stay deterministic and only the genuinely ambiguous parts reach the model.
  • Agent interoperability over MCP Model Context Protocol servers exposing your internal tools and data to agents through one governed gateway — so an agent reaches your systems through a reviewed, permissioned interface rather than a pile of bespoke integrations.
  • Evaluation & observability The part most projects skip. Test sets built from your real cases, measured accuracy before anything reaches production, and drift monitoring afterwards — because an agent that is right most of the time is a liability until you know exactly which times.
  • Enterprise mobile applications Field, rep and site-facing apps that sync cleanly with Salesforce and keep working when the network doesn't.
Agentforce 360 MCP Hybrid reasoning Retrieval & grounding Evals & observability Private models iOS & Android
Focus Industry

Built for life sciences.

Pharma is where our two engines matter most. The data is enormous, the systems are old, the migrations are consequential, and every automated decision has to be explainable months later to someone who wasn't in the room.

How we work in pharma →

  • Migrations designed as documented, reproducible, evidence-producing events
  • Traceability and audit logging designed in from the first sprint, not retrofitted
  • Measured agent accuracy with human checkpoints wherever the stakes require them
  • Commercial, medical affairs and field operations on one coherent data model
  • Validation-aware delivery that fits how your quality organisation already works
Why a small studio

Three things a large SI structurally cannot offer you.

Direct access to seniority

You work with the architects themselves. There is no account manager between you and the build, and no quiet substitution of the senior team for juniors once the contract is signed.

Independent by design

We hold no reseller agreement and earn nothing on licences. When we tell you a cloud is the wrong fit, or that the problem is your data model rather than your platform, nothing about our revenue depends on the answer.

Weeks, not fiscal quarters

A small team with no internal approval chain moves at a different speed. Scoping happens in days, decisions happen on the call, and your first working increment lands early enough to change your mind about it.

RULE 01

Named team, in writing

The engineers assigned to your work are named in the statement of work. If someone has to change, we tell you before it happens.

RULE 02

No pyramid staffing

We don't win work with architects and deliver it with juniors. If a task is genuinely junior, we say so and price it accordingly.

RULE 03

Handover is a deliverable

Documentation and knowledge transfer are scoped from the start. The goal is that your team can own the system without us.

How to start

Start with a two-week Architecture Review.

A fixed-scope, fixed-price engagement. You get a senior architect's written assessment and a plan you own outright — whether or not you build anything with us afterwards.

Deliverables
  • 01Findings document. What your current architecture does well, where it will break, and why.
  • 02Risk register. Ranked technical and delivery risks, each with a concrete mitigation.
  • 03Target architecture. A diagrammed end state for platform, integration, data and AI layers.
  • 04Phased roadmap. Sequenced increments with effort ranges and dependencies.
  • 05Working session. A live walkthrough with your team, recorded, questions included.
Format
Duration2 weeks
Price$7,500 fixed
Your time~4 hours
CreditedAgainst a build
OwnershipYours, fully

If you go on to a build with us, the review fee comes off the first invoice.

PHASE 01

Review

Two weeks. We assess what exists, name the risks, and hand you a roadmap you can act on with anyone.

PHASE 02

Build

Small senior team, short increments, working software in front of your users early and often.

PHASE 03

Operate

Documentation and handover built for your team to own it — or a light retained arrangement if you'd rather we stayed.

Joan Labs

Products we're building.

Client work teaches us which problems recur. When one recurs often enough, we build a product for it. These are in active development — we'll talk candidly about where each one stands.

In development

Document intelligence for pharma operations

Extraction, classification and cited semantic search across dense regulatory and operational document sets, deployable inside your own tenancy.

In development

Migration validation agent

Continuous reconciliation between source and target during large migrations, with anomaly detection and a generated evidence pack.

In development

Process automation agents

Agentic automation for high-volume back-office workflows in regulated environments, with human checkpoints and complete audit logging.

Next step

Tell us what's breaking.

A 30-minute call with a senior architect. No slides, no discovery deck — just a direct read on whether we're the right team for the problem.

Book an Architecture Review → Life sciences work