Technology
The same team builds it.
Strategy decks are cheap. A workflow map is worth paying for when the people who drew it also ship the system.
Plenty of firms will map your processes and hand you a recommendation. Plenty of others will build what you specify. The gap between those two is where most transformation programmes die — the recommendation turns out to be unbuildable, or the build turns out to solve a problem nobody had.
Capability
Four areas, in the order they matter.
AI & automation
Where the transformation work actually lands.
- Agent architectures with written scopes of authority
- Retrieval over private corpora, with citation
- Document extraction tuned to a client's own formats
- Workflow orchestration and exception queues
- Model evaluation harnesses, so accuracy is measured rather than claimed
Integration
The half of every project that decides the timeline.
- CRM, ERP and finance system APIs
- Middleware where no clean API exists
- Event-driven synchronisation between systems of record
- Legacy interfacing, including database-level and file-based exchange
- Identity and permission inheritance from existing directories
Product engineering
When the answer is that something needs building.
- Web applications and internal platforms
- Cross-platform mobile applications
- Customer and partner portals
- SaaS products and marketplaces
- Design systems and interface engineering
Data & reporting
Making a company's own information answerable.
- Metric layer definition — the unglamorous part that decides everything after it
- Warehouse consolidation from fragmented sources
- Governed natural-language querying
- Scheduled narrative reporting
- Permission-aware access
The stack
109 technologies, grouped by what each one is for.
Named because a technical buyer evaluating whether we can build the thing needs to see it. The positioning still holds — the tool is chosen after the problem, not before it.
01
Models and reasoning
Chosen per workload. Cost, latency, context length and where the data is allowed to go decide it — not preference.
- Anthropic ClaudeLong-context reasoning, tool use
- OpenAI GPTGeneral reasoning, structured output
- Google GeminiMultimodal, long context
- Meta LlamaOpen weights, self-hosted
- MistralEfficient open weights
- Amazon BedrockManaged, in-region
- Azure OpenAIEnterprise tenancy
- Google Vertex AIManaged on GCP
- vLLMSelf-hosted serving
- OllamaLocal and edge inference
02
Agents and orchestration
Where the scope of authority is actually enforced — what runs unattended, what escalates, and what is written to a log.
- Model Context ProtocolTool and data connectors
- Tool and function calling
- LangChain
- LlamaIndex
- Vercel AI SDK
- Semantic Kernel
- TemporalDurable, resumable workflows
- CeleryTask queues
- Apache AirflowScheduled pipelines
03
Retrieval and knowledge
Retrieval with citation, permission-aware, over a corpus the client decides on rather than everything on a shared drive.
- pgvectorVectors inside Postgres
- Pinecone
- Qdrant
- Weaviate
- Elasticsearch / OpenSearchHybrid keyword and vector
- Redis Vector
- Hybrid and re-ranked retrieval
04
Documents, speech and vision
Tuned against the client's own corpus. Generic extraction fails on exactly the formats that matter most.
- Azure Document Intelligence
- AWS Textract
- Google Document AI
- TesseractSelf-hosted OCR
- OpenAI WhisperTranscription
- DeepgramStreaming speech
- ElevenLabsVoice synthesis
05
Evaluation and monitoring
Accuracy is measured, not claimed. Every build ships with a harness so drift is visible before a client notices it.
- LangSmith
- RagasRetrieval evaluation
- Custom eval harnessesAgainst the client's own cases
- OpenTelemetryTracing and spans
- GrafanaDashboards and alerting
- SentryError tracking
06
Workflow automation and RPA
Most of this work is closing seams between systems that already exist. Where three Zaps genuinely solve it, we will say so rather than quote a build.
- n8nSelf-hosted, our usual default
- ZapierFast, hosted, per-task pricing
- MakeVisual branching scenarios
- Microsoft Power AutomateInside a Microsoft 365 estate
- Power Automate DesktopAttended desktop RPA
- UiPathEnterprise RPA at scale
- Automation AnywhereEnterprise RPA
- WorkatoEnterprise iPaaS
- Tray.aiHosted integration platform
- ActivepiecesOpen source, self-hosted
- WindmillScripts as durable workflows
- Node-REDEvent and device flows
- CamundaBPMN process orchestration
- PlaywrightBrowser automation where no API exists
07
Integration
The half of every project that decides the timeline. Systems access is audited before anything is quoted.
- REST and GraphQL APIs
- Webhooks and event streams
- Apache Kafka
- RabbitMQ / Amazon SQS
- MuleSoftEnterprise service bus
- BoomiHosted iPaaS
- OAuth 2.0, SAML, SSOPermission inheritance
- SFTP and file-based exchangeWhere no API exists
- Database-level integrationLegacy systems of last resort
08
Business systems we connect to
We integrate before we replace. Adding a migration to the critical path is how transformations fail.
- Salesforce
- HubSpot
- Microsoft Dynamics 365
- Zoho
- Pipedrive
- SAP
- NetSuite
- Odoo
- Xero / Sage / QuickBooks
- Zendesk / Intercom / Freshdesk
- Shopify / WooCommerce / Magento
- Microsoft 365 and Google Workspace
09
Application engineering
When the answer is that something needs building, it is built into the tools people already open.
- TypeScript
- React
- Next.js
- Node.js / NestJS
- Python
- FastAPI
- Django
- Tailwind CSS
10
Mobile applications
Where the workflow leaves the desk — drivers, field engineers, clinicians, shop floors. A single codebase across iOS and Android changes the economics of putting a tool in front of people.
- FlutterOne codebase, iOS and Android
- Dart
- React Native
- ExpoManaged build and delivery
- Swift / SwiftUINative iOS
- Kotlin / Jetpack ComposeNative Android
- FirebaseAuth, messaging, crash reporting
- Offline-first syncField work without signal
- Push notificationsAPNs and FCM
- App Store Connect / Play ConsoleRelease and staged rollout
11
Data and reporting
The metric layer first. Conflicting numbers are almost always a definitions problem rather than a technical one.
- PostgreSQL
- MySQL
- MongoDB
- Redis
- Snowflake
- BigQuery
- dbtTransformations and definitions
- Metabase / Looker Studio
12
Cloud, delivery and security
Deployed where the data is allowed to live, with the arrangement documented rather than assumed.
- AWS
- Microsoft Azure
- Google Cloud
- Docker
- Kubernetes
- Terraform
- GitHub ActionsCI and delivery
- Vercel
- Client-tenancy deploymentWhere residency requires it
These are technologies we build with, not partnerships, certifications or endorsements. The list is deliberately longer than any single engagement needs — what gets used is decided by the workflow and by the systems you already run, after the assessment rather than before it.
How it fits together
Every capability earns its place by what it contributes.
Nothing here is the headline. Each one exists because a transformation needs it, and it gets used only when the workflow calls for it.
- AI
- Intelligence
- Automation
- Efficiency
- Software
- Infrastructure
- Mobile
- Customer and employee experience
- Web
- Digital infrastructure
- Data
- Decisions
- UI/UX
- Usability
- Cloud
- Scalability
- SEO / GEO
- Distribution
How we build
Four commitments that shape every engagement.
We integrate before we replace
Adding a migration to the critical path is one of the most reliable ways to make a transformation fail. Replacement is a finding of the assessment, never an opening assumption.
Configuration is handed over
Thresholds, rules and prompts belong to you. Your operations lead should be able to change one without raising a ticket with us.
Instrumented from day one
Throughput, stall points and time-in-stage are measured before anything changes, so the improvement is a number rather than an impression.
Built to be handed to someone else
Documentation, access and architecture are prepared on the assumption that another team may maintain this. If that makes us easier to leave, so be it.
The tool is chosen after the problem
We do not lead with a framework, a platform or a vendor partnership. What gets used is whatever the workflow and your existing estate actually call for.
No technology for its own sake
If the honest answer is a better form, a cleaner process or a configuration change in software you already own, that is what the assessment will say.
You are evaluating whether we can actually build this.
Evaluating whether we can actually build it?
The gallery has 7 working systems you can open and inspect. Same components, same code we build with.