Kreonovo
Kreonovo
Start a project
Case studies

Selected integration and automation projects

Client projects described in enough detail to show what actually changed. Each one connected systems that were never designed to work together, and replaced manual steps with a workflow the business can rely on.

  • Several systems, one flow

    Each project connects platforms that were never designed to work together.

  • Built to keep running

    Retries, monitoring and clear failure behaviour are part of the delivery, not a follow-up.

  • People stay in control

    Automation handles the mechanical steps; decisions stay with the person accountable for them.

WhatsAppAI agentAPIBusiness CentralQuotation
Systems and technologies involved
Microsoft Dynamics 365 Business CentralCustom AL API extensionsWhatsApp customer channelAI enquiry agentCustom orchestration APILive pricing and stock dataAutomated quotation creation
AI Integration

AI-powered WhatsApp quoting integrated with Dynamics 365 Business Central

Ace Hardware, Zimbabwe

A conversational sales channel where customers describe what they need on WhatsApp and receive a quotation built from live Business Central product, pricing and stock data — with the same quotation waiting in the ERP when they walk into the store.

The challenge

Every product enquiry started with a salesperson. Someone had to interpret what the customer was asking for, search the catalogue for a suitable item, confirm pricing and availability, prepare a quotation and send it back. The work was straightforward but endlessly repetitive, and it tied the earliest stage of every sale to whoever happened to be free. Customers waited for answers, and this hardware retailer wanted enquiries and quotations to move at the speed of the conversation instead.

Our solution

We built an AI-assisted WhatsApp sales and quotation platform that sits directly on top of Microsoft Dynamics 365 Business Central. A customer describes what they are looking for in their own words; an intelligent agent interprets the request, identifies the relevant products and suggests appropriate alternatives where the exact item is not the best fit. The platform then reads live product, pricing, availability and stock information from Business Central, prepares the quotation from it, and creates that quotation inside Business Central as well — so an automated conversation still lands in the client’s own sales and operational systems rather than beside them. Where the standard Business Central APIs could not expose what the agent needed, we developed custom AL extensions to close the gap, including stronger product search and structured product information written specifically for an AI agent to read accurately.

Business value

  • Customers get a relevant quotation inside the conversation instead of waiting for a salesperson to become available.
  • Product discovery, pricing and quotation preparation happen without a manual catalogue search.
  • Quotations exist in Business Central, so store staff can retrieve a prepared request and fulfil it without repeating the enquiry.
  • Sales staff spend significantly less of the day on repetitive enquiry handling.
  • The product and stock information customers see comes from the ERP, so it stays current.
  • Enquiry and quoting capacity scales with demand rather than with headcount.

Key outcome

WhatsApp became an automated sales interface rather than another inbox to watch, and the path from a customer’s first message to an in-store purchase runs as one connected process. The project brought a conversational AI agent, custom APIs and an enterprise ERP together into a single automated business workflow.

Credit bureausTransUnionCPBExperianXDSArchiveOrchestrationOne contractCOBOL core
Systems and technologies involved
TransUnionCPBExperianXDSCOBOL core systemCustom API orchestration layerValidation and transformation rulesArchived result and soft-search store
Legacy Integration

Multi-bureau credit check orchestration for a legacy COBOL environment

Railway Furnishers

One integration platform that talks to four credit bureaus, standardises their very different responses, and hands the COBOL core system a single predictable data contract it can act on.

The challenge

Assessing a customer for credit or finance depended on information from several external bureaus, each with its own API, response format, rules and data structures. The organisation’s core system runs on COBOL, and wiring every provider into it directly would have pushed four sets of third-party quirks — plus every future change to them — into the least forgiving part of the estate. What the business needed was one place where that complexity could live, and one shape of answer arriving at the core.

Our solution

We built a custom credit bureau orchestration platform that sits between the bureaus and the COBOL environment, integrating with TransUnion, CPB, Experian and XDS. It submits requests to the appropriate providers, then validates, transforms and enriches whatever comes back into one consistent contract designed for the core system to consume. The platform does more than relay responses: it restructures the information so that downstream credit decisions have less work to do. It also supports soft searches — where suitable archived bureau information already exists, it can be located and reprocessed against current requirements instead of repeating the full external enquiry unnecessarily.

Business value

  • Credit checks are processed automatically, with far less manual handling of bureau results.
  • Consolidated credit information arrives in a decision-ready structure, whichever bureau supplied it.
  • The COBOL system integrates once, against a stable contract, rather than against four moving targets.
  • Adding or replacing a credit bureau is a change in the integration layer, not in the core system.
  • Soft search reuses archived bureau data and avoids unnecessary external searches.
  • Provider-specific failures are handled in one place, so one bureau having a bad day does not stall an assessment.

Key outcome

The client gained a single, maintainable interface to the entire credit bureau landscape, while the legacy platform carried on doing what it does well. It is the pattern we return to often: modernise around the system of record instead of replacing it, and keep the complicated parts — transformation, orchestration, reliability and business rules — where they can be understood and changed safely.

WebsiteAPIStrapiRedisLead portalSales team
Systems and technologies involved
SvelteKitCustom API layerStrapi CMSRedis-backed performance layerProduct search and discoveryAutomated enquiry captureLead-management portal
Digital Platform

A high-performance product platform with automated lead management

Bitvest Afcom

A rebuilt product website with a custom API and caching layer behind it, paired with an internal portal that captures every website enquiry and carries it through a structured follow-up workflow.

The challenge

Bitvest Afcom needed its website and digital product experience redeveloped from the ground up, with a much better way for visitors to search, discover and explore the product range. Internally, enquiries were handled largely by email, which made it difficult to hold a complete overview of incoming leads: which had arrived, who had replied, what had already been discussed, and which were quietly going cold.

Our solution

We rebuilt the website from the ground up on SvelteKit, supported by a custom API architecture. Structured content and product management run on Strapi, while Redis backs high-performance data access so product browsing and search stay fast as the catalogue grows. Alongside the public site we developed a custom enquiry and lead-management portal. Every website enquiry is captured centrally rather than landing in an individual inbox, and the marketing and sales team can view, assign, process and follow up on leads — with the history of previous interactions attached — from one location.

Business value

  • Visitors find relevant products far more easily, through search and browsing built for a large range.
  • Redis-backed data access keeps product pages and search responsive under load.
  • Every website enquiry is captured automatically into one workspace.
  • Leads move through a visible status workflow, so follow-up is deliberate rather than dependent on someone remembering.
  • Previous enquiries and interactions are on record, instead of being reconstructed from email threads.
  • Handovers between team members no longer lose context, and fewer opportunities are overlooked.

Key outcome

The public site and the internal sales process now work as one platform: interest generated on the website enters a workflow the team can actually manage and report on. It is a clear illustration of how a modern front end, custom APIs, content management, caching and workflow automation combine into a complete digital sales platform rather than a brochure.

Scanner appSubmitServerProcessing
Systems and technologies involved
Flutter applicationScorpio X5 handheld scannersBarcode and item scanningStructured stocktake sessionsOn-device validationElectronic submission APICentral stocktake processing
Process Automation

Paperless stocktaking with a custom Flutter scanner application

TAPE

A purpose-built Flutter application for Scorpio X5 handheld scanners that turns a paper-based stocktake into a structured digital session, submitted straight to the server for processing.

The challenge

Stocktakes were recorded on paper and submitted by hand. Counts were captured manually and then recaptured into another system afterwards, which introduced delays between counting and central processing, left room for incomplete or incorrectly transcribed records, and gave the business little visibility while a stocktake was still in progress. Consolidating paper submissions was an administrative job in its own right.

Our solution

We developed a custom Flutter application for the client’s Scorpio X5 stocktake devices. The stocktaker identifies the relevant store, works through the count on the scanner, and captures everything digitally in a structured stocktake session. Captured information is validated on the device and submitted electronically to the server for central processing. Both the handwritten stock sheet and the second round of manual recapturing came out of the process entirely.

Business value

  • The stocktake is fully paperless, from the first scan to central processing.
  • Stock information is captured once, at the point of work, and never recaptured.
  • Completed stocktakes reach the server for processing as soon as they are submitted.
  • On-device validation lowers the risk of incomplete counts and transcription errors.
  • Stock data arrives in a consistent, structured form every time, improving traceability.
  • Significantly less administrative work once a stocktake closes.

Key outcome

A physical, paper-driven operational routine now feeds the central business system directly, and completed counts are available for processing the moment they are submitted. The application also gives the client a platform for further stock and scanning workflows on hardware their teams already carry.

ProductsEnquiryCapturePortalFollow-upHistory
Systems and technologies involved
Custom websiteCustom API layerStrapi CMSStructured product contentAutomated lead captureInternal lead-processing portalFollow-up and engagement tracking
Lead Automation

A product-focused website with centralised enquiry and lead automation

Buffalo Tapes

A modern product website backed by a custom API and Strapi, with an internal portal that captures every enquiry and gives the team one place to process leads, track engagement and keep the history.

The challenge

Buffalo Tapes needed a modern digital platform where customers could explore the product range and submit enquiries efficiently — and, just as importantly, a structured way to deal with what came back. Handling enquiries mainly through email meant unnecessary manual administration and no reliable view of which leads had been contacted, which were being processed and which had been followed up.

Our solution

We developed a modern, product-focused website supported by a custom API and a Strapi-based content management system, so the client’s team manages site and product content themselves while customers get an effective way to explore the offering. Alongside it we built a custom internal portal that automatically captures enquiries and leads submitted through the website. The portal gives the team a central workspace to review enquiries, process leads, track engagement and keep a history of previous interactions — one place instead of several inboxes.

Business value

  • Prospective customers receive a faster and more consistent response.
  • Every enquiry is captured centrally, so nothing depends on someone noticing an email.
  • The lead pipeline is visible, with clear status and ownership.
  • Continuity holds when more than one team member handles the same enquiry.
  • A complete history of previous leads and customer engagement is kept for reference and reporting.
  • Manual email administration is largely removed from the process.

Key outcome

Website activity now converts into sales opportunities the team can act on and account for, with far less risk of an enquiry being missed. The project shows that the customer-facing site is only half the work — the APIs, portal and automated workflow behind it are what turn interest into a managed pipeline.

Your project

Which systems should be talking?

If one of these looks close to your situation, tell us about yours. We are happy to walk through how we would approach it, including what we would not recommend.

Discuss a project Explore services