The business had outgrown a normal lead tracker

Nexium Connect began as a focused internal system for managing outreach, leads, approved clients and onboarding. As the company's workflow matured, the problem became larger than storing contact details. The directors needed to know who owned each opportunity, what had happened, what was due next, which leads had converted and which business actions still required attention.

That information was spread across calls, WhatsApp messages, spreadsheets, notes and separate operational tools. The cost was not only lost time. Duplicate records could enter the pipeline, follow-ups could be missed, client history could become fragmented and director decisions could sit outside the system where the work was happening.

The new Nexium Connect was rebuilt as an operating platform rather than a static CRM: one shared workspace for sales execution, client movement, onboarding, approvals, tasks and AI-assisted action.

From recording work to helping the team execute it

The platform now follows the full relationship from early outreach through conversion and delivery. Leads can be imported, assigned, edited inline, filtered by owner and moved through clear pipeline stages. Once a lead converts, the connected client record carries the relationship into onboarding, payment tracking, documents and delivery preparation.

  • A dashboard surfaces pipeline movement, owner performance, active clients, revenue and operational priorities.
  • The outreach workspace supports spreadsheet-style editing, Google Sheets and Excel imports, owner assignment and bulk actions.
  • Duplicate protection matches incoming records and merges repeated rows instead of quietly multiplying the pipeline.
  • Converted leads remain linked to their client records so the sales history is not lost at handover.
  • Onboarding brings agreements, invoices, welcome packages, checklists and shared files into the same workflow.
  • A director workspace handles approvals, decisions, follow-ups and a shared project board with clear ownership and status.
Connected business lifecycle one record from first contact to delivery
01Research & importProspects enter through manual capture, Excel or Google Sheets.
02OutreachOwner, status, notes, demo links and next actions stay visible.
03ConversionThe lead becomes a client without breaking its history.
04OnboardingDocuments, payments, files and delivery actions are coordinated.

Meet Layla

Layla was created because the team did not need another chatbot that only gives advice. They needed an assistant that understands how Nexium Connect works, can read the live operational context and can help complete real tasks inside the platform.

Her job is to turn normal human instructions into safe, structured business actions. A director can speak or type naturally and ask Layla to find a lead, create a prospect, update contact details, change ownership, prepare a follow-up, create a director task, summarise the pipeline or organise a bulk workflow. Layla handles the translation between what the user means and what the system needs.

This makes Nexium Connect faster to use, especially on mobile or immediately after a call. Instead of opening several pages and forms, the team can tell Layla what happened and what needs to happen next. She reduces repetitive administration, shortens the path from decision to action and helps keep the database current while the work is still fresh.

  • Sales support: finds prospects, creates and updates leads, changes stages, assigns owners and prepares follow-ups.
  • Operations support: creates director tasks, records next actions and summarises outstanding work.
  • Pipeline intelligence: answers questions using live lead and task context rather than generic information.
  • Voice-first capture: transcribes spoken instructions so updates can be made directly after meetings or calls.
  • Controlled execution: previews planned actions and requires confirmation before destructive or large-scale changes.
  • Data protection: checks for duplicates, respects permissions and keeps AI credentials on the backend.
Layla in action: "Find the Cape Town leads waiting for a demo, assign them to Enrico and create a high-priority follow-up task." Layla understands the request, checks the live records, prepares the changes and asks for confirmation before executing anything sensitive.

Why we built her

CRM systems often become outdated because every call, note, assignment and next step still depends on someone remembering to update several fields manually. Once the admin starts falling behind, the dashboard stops reflecting reality and the team goes back to WhatsApp messages, spreadsheets and memory.

Layla was built to close that gap. She gives the team a faster way to capture information and act on it while preserving the controls expected from a serious business system. The aim was simple: make Nexium Connect feel less like software that must be maintained and more like an operational teammate that helps maintain the business.

That is also what makes the product marketable. The same problem exists in sales teams, service businesses, agencies and growing companies everywhere: valuable information is scattered, follow-ups are missed and staff spend too much time updating tools instead of serving clients. Layla shows how a custom AI assistant can be embedded into a company's own workflow, language, rules and permissions instead of forcing the company into a generic AI product.

Commercial value: fewer missed follow-ups, less manual capture, faster handovers, cleaner pipeline data and a shorter path from instruction to completed action.

The technical build

The product evolved alongside its infrastructure. The interface remains a fast, responsive HTML, CSS and JavaScript application, while the backend now runs as a Node.js service on Render. Supabase provides persistent shared data and authentication, Vercel serves the frontend, and Groq powers Layla's command planning and voice transcription through backend-only endpoints.

System architecture shared, secure and AI-enabled
InterfaceResponsive web appDashboard, outreach, leads, clients, onboarding, directors and Layla.
→
ApplicationNode.js API on RenderAuthentication, validation, business rules, AI planning and confirmed execution.
→
DataSupabasePersistent shared records, users, tasks, approvals and operational data.
AIGroq + LaylaStructured command planning and voice-note transcription.
FrontendVercelFast deployment for desktop and mobile access.
ControlsRoles + confirmationsDirector access, root approvals, rate limits and protected destructive actions.

What had to be solved

Adding a chatbot would have been easy. Giving an AI assistant permission to change real business records required a much stricter design. Layla needed a limited action vocabulary, validated statuses and owners, clear duplicate handling, contextual search and a confirmation boundary around destructive or large-scale changes.

The broader platform also had to remain dependable while becoming more capable. Imports needed to preserve ownership, repeated rows needed to merge safely, converted clients needed linked histories, mobile navigation needed to feel like a proper application and director-only functions needed to stay protected.

The result is an AI feature designed as part of the operating system — not a decorative chat panel. It reduces repetitive administration while keeping the team in control of what is written to the business database.

The result

Nexium Connect now gives BuildsByBuchanan a shared view of sales activity, conversion, client onboarding, revenue and director execution. Every lead can have an owner, status, history and next action. Every converted relationship can continue into a connected client workflow. Every director task and approval can live beside the operational data it affects.

Layla is the feature that changes how the platform feels in daily use. Routine updates can be completed through natural language or voice, pipeline questions can be answered from live records, and multi-step workflows can be prepared without navigating through several forms. The system still remains controlled through confirmations, permissions and backend validation.

What started as a lead tracker has become a market-ready example of practical business AI: a purpose-built operating platform designed around a real company process, strengthened by shared cloud infrastructure and extended by an assistant that does useful work instead of simply talking about it.