Els Labs
Case studies

Products we're proud to have shipped.

Real projects, real outcomes. Each case study walks through the challenge, how we solved it, and the measurable results our clients saw after launch.

Cashflow
Mar 2026
£1.42M
↑ 24% vs last quarter
Inflow
£2.1M
Runway
19 mo
SaaS PlatformFintech12 weeks

A real-time finance platform for 4,000+ SMEs

Meridian had outgrown their Excel-and-email approach to financial reporting. Their legacy tool was slow, error-prone and couldn't support the multi-entity structure their growing client base demanded. Month-end close routinely took three to five working days.

We rebuilt the entire platform as a multi-tenant SaaS product with live cashflow dashboards, automated bank reconciliation and an AI-powered assistant that answers natural-language questions about a company's financial position. The new system cut month-end close from days to minutes and became the primary daily tool for over 4,000 SME finance teams.

Challenge

Replace a brittle, single-tenant reporting tool with a scalable platform that serves thousands of tenants with real-time data, strict data isolation and zero planned downtime during migration.

Solution

A ground-up rebuild using Next.js and PostgreSQL with row-level security for tenant isolation, Stripe for subscription billing, and an OpenAI-powered assistant embedded into the dashboard. We shipped incrementally over 12 weeks with a blue-green migration strategy that moved tenants without interruption.

Results

Report generation dropped from 3.1 seconds to 0.4 seconds. Activation rate increased 42% within the first quarter. The launch saw zero downtime and zero data incidents across all migrated tenants.

3.1s → 0.4s
Report load time
+42%
Activation rate
£0
Downtime at launch
Next.jsPostgreSQLStripeOpenAI
View detailed case study
TODAY · BOOKINGS
17 jobs
Next
Boiler service
10:30 · SW1 · J. Patel
Leak repair12:00 · E8
Inspection14:15 · N1
Mobile AppiOS & Android10 weeks

Field-service scheduling for 600 engineers

Ashcroft manages a nationwide network of 600 field engineers handling boiler installs, gas safety inspections and emergency repairs. Their scheduling was a patchwork of spreadsheets, phone calls and a decade-old web portal that engineers refused to use on site.

We built a cross-platform mobile app that intelligently routes jobs based on location, skill set and availability, captures customer sign-off with photo evidence, and syncs everything offline-first so engineers in basements and rural areas never lose work. The back office gained a real-time dispatching dashboard that replaced four separate tools.

Challenge

Replace fragmented scheduling tools with a single mobile app that works reliably offline, handles complex multi-skill routing, and integrates with Ashcroft's existing ERP and compliance systems.

Solution

A Flutter app with offline-first architecture using local SQLite and background sync via Firebase. The routing engine factors in travel time, engineer certifications and parts availability. A Node.js backend handles job allocation, compliance logging and real-time push notifications to dispatchers.

Results

Missed appointments dropped by 40% in the first two months. The app achieved a 4.8-star rating on the App Store within its first quarter. All 600 engineers were onboarded and actively using the app within six weeks of launch.

600
Engineers onboarded
-40%
Missed appointments
4.8★
App Store rating
FlutterFirebaseNode.js
View detailed case study
AI ASSISTANT
Which contracts renew in the next 30 days?
7 contracts worth £214krenew before 9 Apr. 2 are flagged at-risk — want me to draft outreach?
Draft emailsExport list
AI IntegrationB2B8 weeks

An AI copilot for a sales operations team

Lattice's sales team was drowning in data. Contract details lived in the CRM, pricing history in spreadsheets, and product specs across a dozen Confluence spaces. Reps spent hours each week hunting for answers before they could respond to prospects.

We embedded a retrieval-augmented AI assistant directly into their CRM sidebar. The copilot indexes contracts, pricing sheets, product documentation and historical deal data, then answers natural-language questions with cited sources. Reps now get accurate, grounded answers in seconds instead of context-switching across five tools.

The system includes guardrails that prevent hallucination, an admin panel for managing indexed sources, and a feedback loop that continuously improves answer quality based on rep ratings.

Challenge

Build an AI assistant that answers complex sales questions accurately using scattered internal data, without hallucinating or exposing confidential deal information across team boundaries.

Solution

A retrieval-augmented generation pipeline built with LangChain and Pinecone. Documents are chunked, embedded and stored with metadata-based access controls. A React sidebar component renders answers with inline citations and confidence scores. The Python backend handles re-ranking, prompt engineering and audit logging.

Results

Sales reps save an average of six hours per week on information retrieval. Answer accuracy reached 89% within the first month, verified against manual spot-checks. The project paid for itself in under three weeks through reduced time-to-quote alone.

6 hrs/wk
Saved per rep
89%
Answer accuracy
3 weeks
Payback period
PythonLangChainPineconeReact
View detailed case study
Let's build

Have a product worth building properly?

Tell us where you're headed. We'll come back within two working days with a clear, honest view of how we'd build it.