02Catalogue
Six categories. Each one a real, running system.
Pick one and we scope it; combine several and we sequence them. Every solution below has a defined input, a defined output, and a named team that owns it inside the customer's organisation.
SOLUTION · 01
AI Assistants
Copilots for internal teams — grounded in your data, your workflows and your tools. Not a chatbot bolted to a wiki: an assistant that understands the task, retrieves what's relevant, drafts the work, and writes back to the systems where the work actually lives.
- Input
- Natural-language task or question · scoped to a workflow
- Output
- Drafts · decisions · actions written into your systems
Where it fitsSalesOperationsEngineeringFinance
SOLUTION · 02
Document Intelligence
Read large volumes of contracts, invoices, claims, certifications or technical drawings. Extract structured data, summaries and decisions. Flag exceptions. Route to the right person. The system learns the cases it should not handle alone.
- Input
- PDFs · scans · email attachments · scanned forms
- Output
- Structured records · summaries · flagged exceptions · routing
Where it fitsLegalClaimsProcurementFinance Ops
SOLUTION · 03
Knowledge Systems
Retrieval-grounded answers from your own corpus — wiki, tickets, SharePoint, code, policies, manuals. Every answer is cited and source-linked. The system declines when sources don't support the claim, instead of inventing one.
- Input
- Natural-language question · scoped access to your corpus
- Output
- Cited answer · linked sources · escalation when insufficient
Where it fitsSupportEngineeringHRCompliance
SOLUTION · 04
Workflow Automation
End-to-end processes that read, decide and act inside your stack. Triggers, branches, retries, fallbacks. We treat them like real distributed systems — observable, idempotent and recoverable when something goes wrong.
- Input
- Trigger event · business context · system access
- Output
- Work performed inside your CRM, ERP, ticket platform, document store
Where it fitsOperationsCustomer SuccessFinanceRevOps
SOLUTION · 05
AI Product Features
AI capabilities built into the product you already ship — assistants, generators, smart filters, anomaly surfaces, recommendations. Evaluated continuously, observable, accountable to the same SLAs as the rest of your product.
- Input
- Existing product surfaces · customer interactions · operational data
- Output
- Embedded AI features · production-grade · evaluated per release
Where it fitsProductEngineeringGrowth
SOLUTION · 06
Decision Support Tools
Aggregate operational signals, model the option space, brief the decision-maker. The system surfaces evidence, weighs trade-offs and notes its own uncertainty. It is not the decider, and is explicit about that. Every run leaves an audit trail of every input and assumption.
- Input
- Trigger event · operational data · constraints · policy
- Output
- Ranked option set · recommendation · dissent · uncertainty notes
- Mode
- Advisory only · always reviewed · full audit trail per run
Where it fitsPricingCreditSupplyM&A DiligenceExecutive Briefings