Skip to main content
Download free report
Softblues
Softblues
Our Work

AI Case Studies

45 projects, from AI agents and voice assistants to document automation and internal copilots. Each one lists the problem, what we built and the numbers that came out of it.

45
Projects delivered
7
Flagship builds
15
Sectors covered
50+
AI projects

Anthropic Partner Network member · Google Cloud Partner · Top-5 UK AI Firm

Flagship AI projects

The 7 projects we point to most often.

The SofiaHR candidate assessment pipelineSpeech enters as a waveform, passes through six stages, and fans out to ten fine-tuned models that score the candidate profile.SOFIAHR · ASSESSMENT PIPELINE10 MODELSSPEECH → SIX STAGES → SCORED PROFILE2–3 hrs → mins

HR & Recruitment · AI Business Automation

SofiaHRIn beta · 10 clients

AI candidate assessment, 2–3 hours to minutes

How Softblues automated SofiaHR's candidate-assessment process: a 2 to 3 hour expert linguistic analysis, scored across 10 psychological dimensions, now runs in minutes at 90%+ of human-analyst accuracy.

2–3 hrs → minsPer-Candidate Analysis

90%+ accuracy vs expert analysts

15 technologies · Fine-tuned Gemini Flash (10 models) + Claude SonnetView case study
The clinical research pipelineOne question in plain English climbs through nine agents in sequence and comes out as an analysis, with no code written by the researcher.CLINICAL RESEARCH · NINE AGENTSQUESTIONin English123456789ANALYSIS0 LINESof code writtenONE QUESTION → NINE AGENTS → A PAPER12–18 mo → weeks

Healthcare · AI Business Automation

In beta

AI-Powered Clinical Research Platform

How Softblues automated a 12 to 18 month manual research process into a nine-agent AI pipeline that turns a plain-English question into a publication-ready analysis in weeks.

Months→WeeksResearch Cycle

9 ai agents orchestrated

17 technologies · Claude Sonnet + LangChainView case study
The Softblues Claude operating systemFive functions, strategy, marketing, finance, sales and delivery, arranged around one shared system hub.SOFTBLUES · SIX CONNECTED SPACESSTRATEGYFINANCESALESDELIVERYMARKETINGSYSTEM HUBFIVE FUNCTIONS · ONE SHARED HUB6 spaces

Softblues · AI Business Automation

Running in-house

The Softblues Claude Operating System

Softblues runs its own company on Claude: six connected agents, one per function, each with its own operating manual, tools and routines, joined through a shared hub. We use it before we sell it, so we can show it live rather than in a slide deck.

6Connected Spaces

3+ functions run by agents

14 technologies · Claude (Anthropic) + CoworkView case study
The Roof Maker voice assistantCalls arrive into one voice backed by eight agents, and leave by one of two routes: a routine ticket, or an escalation to a person.ROOF MAKER · ONE VOICE, EIGHT AGENTSCALLS INONE VOICEEIGHT AGENTSTICKETA PERSONROUTINE HANDLED · URGENT ESCALATED24/7

Customer Support · AI Business Automation

Roof MakerLive

From Missed Calls to 24/7 Support

Roof Maker's support line went quiet after hours, and routine questions ate up specialist time during the day. Softblues automated the front line of their customer support: an AI voice agent that answers every call, troubleshoots, logs the ticket in HubSpot, and escalates anything urgent to a person. 24/7, in production.

24/7Support Coverage

8 ai agents

12 technologies · Python + LiveKitView case study
Order-to-schedule automation for a secure logistics operatorMessy intake documents feed a rules engine of more than seventy rules, which splits work into orders that run themselves and exceptions a person decides.SECURE LOGISTICS · MANAGE BY EXCEPTIONMESSY INTAKE70+ RULESCLEAR · RUNS ITSELFEXCEPTION · A PERSON DECIDESIN THE CLIENT'S OWN TENANT70+ rules

Secure Logistics · AI Business Automation

Discovery done · pilot designed

Order-to-Schedule Automation for a Secure Logistics Operator

A cash-in-transit operator ran its scheduling on email and spreadsheets: hundreds of order emails a day, hand-built run sheets, and one person's knowledge holding it together. We ran a four-week discovery, mapped the whole process and its 70+ rules, and designed an order-to-schedule automation that runs inside the client's own Microsoft tenant.

15–30 hrs/wkScheduling Time to Reclaim

70+ business rules mapped

14 technologies · Claude Sonnet + Claude OpusView case study
The compliance file review pipelineA client file passes through four review stages of around forty checks, every finding cited back to its source, ending with a human sign-off that is kept.FILE REVIEW · EVERY FINDING CITEDTHE FILESTAGE 01STAGE 02STAGE 03STAGE 04~40 CHECKSHUMAN SIGN-OFF, KEPTFOUR STAGES · THEN A PERSON~40 checks

Financial Services · AI Business Automation

Proposed

Automating the Monthly Compliance File Review

A regulated financial-advice firm reviews every client advice file by hand each month: dozens of mechanical checks per file plus a judgement call on suitability, all under a regulator's eye. We proposed a four-stage multi-agent pipeline that does the mechanical work and drafts the review, with a compliance reviewer keeping the final sign-off. It is designed to run entirely inside the firm's own Microsoft tenant, in the EU.

~40Checks per File

4 pipeline stages

11 technologies · Claude Sonnet + Claude OpusView case study
The Claude Code audit sequenceA fractured monolith moves through audit, week-one lockdown, rebuild and handover, ending as modular components owned in-house.CODE AUDIT · 58K LINES BEFORE GO-LIVEONE MONOLITHAUDIT01LOCKDOWN02REBUILD03HANDOVER04WEEK ONEMODULAR2 GO-LIVE BLOCKERS CLOSED58K lines

Food Production · Claude Code Setup

Audit delivered · rebuild proposed

Auditing an AI-Built App and Making It Safe to Scale

A food producer's own team built a real operational app with AI: procurement, stock and traceability, and it works. Underneath, it was a 58,000-line monolith with critical security holes and no tests, weeks from go-live. We ran a Claude Code audit and discovery, closed the security exposure, and set out a secure, modular rebuild plus an AI development setup the team keeps building on.

58KLines Audited

4 priority findings

12 technologies · Claude Code + ClaudeView case study

The full back catalogue

38 further projects, grouped by industry.

Want the same for your process?

Bring us one workflow that costs you time every week. We will tell you whether AI is the right answer and what it would take.