Skip to content
Annatech_

Work

Case studies, with the numbers left in

Client names are confidential, as they should be. Everything else - architectures, constraints, volumes, outcomes - is here in enough detail that an engineer can judge the work.

Case studies

8 engagements, end to end

Applied AI / Intelligent Automation

Teaching a local LLM to write production UiPath code

A knowledge-injection spec plus a static-analysis toolchain lets a locally hosted open-weight model generate UiPath XAML that passes validation before Studio ever opens - with no client code leaving controlled infrastructure.

Read the case study

Intelligent Automation / SAP

Calibration certificates into SAP DMS at 50,000 documents a year

A two-phase unattended process resolves equipment from serial numbers, verifies it in SAP and files certificates into the document management system - roughly 50,000 documents a year, fully audit-trailed.

Read the case study

Applied AI / Document Intelligence

AI validation of US tax-exemption certificates across seven state form families

Azure Document Intelligence OCR feeding a deterministic rule engine - 8 generic checks plus 65+ state-specific rules - triages certificates into pass, fail or human review with per-rule evidence.

Read the case study

Intelligent Automation / Security

Unattended automation under mandatory MFA: 8-12 prompts down to 2

When mandatory MFA broke an unattended SharePoint + SAP process, consolidating authentication to two strategic points with session reuse cut MFA cycles by roughly 80% - without touching security policy.

Read the case study

Enterprise Workflows / Pega

Standardized offline Pega branch reviews on a 1,500-rule platform

A review methodology plus LLM knowledge file reconstructs rule logic from branch exports, rule XML and tracer logs - 31 standardized reviews delivered without competing for live-instance access.

Read the case study

Private AI / Infrastructure

A private AI platform with zero external AI APIs

Dual-backend LLM inference, 30+ agent tools behind tiered auth, voice, image generation and self-healing infrastructure - a complete AI platform where no prompt, document or token ever leaves the network.

Read the case study

Applied AI / Research Automation

Unattended LLM market research across 11,500 cities, every night

A self-hosted research pipeline that prices housing in 11,543 cities worldwide with a language model on our own hardware - every figure checked by deterministic rules before it is published, 8,000+ cities priced without a person in the loop, and a public globe that makes zero external requests.

Read the case study

Applied AI / Customer-facing LLM

A public AI assistant on a company website, answered by a local model

The assistant on this site runs on an open-weight model on our own hardware. A small gateway adds the company knowledge, per-visitor limits, a transcript store and automatic failover to a second backend - the pattern for putting an LLM in front of customers without handing their questions to a cloud vendor.

Read the case study

Also delivered

The longer tail

Talk to the person who will actually build it

One architect, end to end: scoping, architecture, delivery, operations. Write a paragraph about your problem and you will get an engineering answer, not a sales call.

[email protected] Capability statement (PDF)