Home › What We’ve Built › Case Studies

Case Studies

Seven engagements: a bank, a listed REIT, a residential developer, a steel and mining group, a steel producer and a professional services firm. Each one names the number it moved.

Featured

10 use cases

in production, built by the bank’s own engineers

OraclePostgreSQLGenerative AIMachine learningPythonBank core systems

BFSI · AI centre of excellence

A bank’s own engineers now ship AI use cases without us.

Canara Bank

The bank had a capable technology team, a mandate to adopt AI, and no internal practice for taking a model from an idea to something running against bank data. We built the centre of excellence inside the bank, trained the engineers, and worked beside them on live use cases until the standards were theirs. Fifteen executives took the generative AI leadership track alongside it.

Read the case study →

All case studies

Every engagement we can publish.

Sorted by the outcome each one moved.

Real Estate & REITs · Generative AI

4 weeks → 3 hours

valuation report cycle

A REIT closes its investor reporting cycle in an afternoon.

Mindspace REIT

Investor reporting was assembled by hand from valuation files, portfolio data and prior filings, against SEBI deadlines that do not move. The system reads the underlying records, drafts against the previous period, and flags what is missing before a person opens it.

SEBI filingsValuation recordsGenerative AI
→

Real Estate · Agentic AI

90 seconds

from enquiry to a profiled lead

Sales managers know the buyer before the first call.

One of India’s largest residential developers

Enquiries arrive from advertising, walk-ins, brokers and referrals, and the first fifteen minutes decide whether one converts. An agent profiles every enquiry across eight dimensions and writes the result into the CRM before a manager opens it, with nothing typed by hand.

SalesforceAgentic AILead Enrichment Agent
→

Steel & Mining · Custom build

Legacy → live

dispatch operations platform

Dispatch is visible while it runs, from mine to plant.

An integrated steel and mining group

The old system recorded what had happened and could not show what was happening. We built the dispatch and transport platform from the ground up: vehicle lifecycle, weighbridge integration, gate automation and the views the people running dispatch work from.

SAPWeighbridgeRFID and ANPRGPSFull-stack build
→

BFSI · Capability building

39 workshops

across three tiers, board to DevOps

One AI programme covering every tier of a bank.

A public sector bank

Designed against the bank’s EASE 8.0 reform indicators: ten executive workshops on AI strategy and governance, fourteen for branch, credit and operations teams on the tools they use daily, and fifteen hands-on technical bootcamps for data science, DevOps and cybersecurity.

AI strategyML for bankingGenAI and LLMsMLOpsCybersecurity
→

Steel & Heavy Industry · AI training

10 workshops

of ninety minutes each, online

Technical teams took AI from awareness to a daily work skill.

A national steel producer

Ten online sessions for engineers, automation specialists, data analysts and R&D scientists, built around the plant’s own SAP and sensor data and its thousands of SOPs and circulars. Each one taught a single capability and ended with something usable the next morning.

SAP dataSensor and SCADA dataDocument workflowsAI agentsAutomation
→

Professional Services · Developer enablement

96 hours

of in-person instruction

Engineering teams trained on agentic coding tools.

A global professional services firm

Ninety-six hours on Claude Code and AI-assisted development, taught in person over three days to the firm’s engineering teams. The exercises ran against the kind of codebases those teams maintain.

Claude CodeAI-assisted developmentDeveloper workflow
→
AI Consulting · Implementation · Training

Bring us the problem no one else has solved. Leave with a clear answer.

Tell us what you are trying to solve, fix, or decide. We reply within 48 hours.

Talk to us
An EigenSpark session in progress: a trainer presenting responsible-AI data rules to a room of analysts at workstations.