Our work

AI automation case studies and use cases

Explore FusionETA client deployments, pilots, demos and in-house workflows across CRM, accounting, procurement and HR.

These examples come from our project experience. Each status distinguishes a live deployment from a pilot, demo, product or in-house workflow. Client names are withheld.

Subscription & services businessLIVE

Renewal agent in Bitrix24 chat

Account managers kept exporting reports to find which renewals were due, overdue or churned.

  1. Staff DM the agent in Bitrix24 Messenger in plain language
  2. It queries the live renewal pipeline and replies with a tidy summary and tap-to-act buttons
  3. It can mark deals renewed or lost, move stages and add timeline notes — after the user confirms

Outcome: renewal status on demand for the whole team, with every change logged on the deal.

Bitrix24 imbotLLM intent parsingPHP webhook
Read the case study →
Multi-company retail & wellness groupLIVE

Accounting migration & reconciliation agent

Moving several group companies from a legacy desktop ledger to Xero meant thousands of lines to re-map and reconcile.

  1. Claude parses legacy ledger, aging and journal exports and maps them to the group chart of accounts
  2. It builds Xero-ready import files for opening balances, invoices, bills and manual journals
  3. Connected to Xero through our own MCP server, it reads trial balances and ties bank, AR and AP back to source

Outcome: 57 journals / 669 lines rebuilt and balanced for one entity; bank, AR and AP reconciled to source.

ClaudeXero MCPPythonSST
Read the case study →
Engineering & fabricationPILOT

AI procurement: WhatsApp request to Xero PO

Site teams sent material requests as WhatsApp photos and voice notes; purchasing retyped every one.

  1. The agent reads requests and quotation images sent on WhatsApp using AI vision OCR
  2. It builds a structured material requisition, matching items and suppliers
  3. On approval, it pushes a draft purchase order or bill straight into Xero

Outcome: no retyping between site request and PO; approvers stay in control of spend.

Vision OCRWhatsAppXero APIStarship
Read the case study →
Structural steel contractorLIVE

Tender-screening & lead-generation agent

Business development spent hours scanning public tenders and project databases for jobs that fit.

  1. The agent reviews new public tenders and construction-project listings across the region
  2. It scores each one for fit against the company's steelwork capabilities
  3. Qualified projects become enriched contacts and leads in Bitrix24, ready for follow-up

Outcome: BD starts the day with a qualified pipeline instead of a search task.

ClaudeWeb researchBitrix24 CRM
Cross-border logisticsLIVE

Fleet-tracking assistant

"Where's the truck?" calls interrupted drivers and dispatchers throughout the day.

  1. Staff ask the agent for the delivery truck's position in plain language
  2. It reads live GPS from the tracking platform and interprets the location
  3. It posts the update to the right colleague in Bitrix24

Outcome: instant location answers without phone calls to the driver.

ClaudeGPS APIBitrix24
Flooring retailLIVE

AI sales specialist on WhatsApp & web chat

Enquiries on SPC, vinyl, laminate and timber arrived around the clock; sales couldn't reply fast enough.

  1. An AI sales assistant trained on the product range answers and qualifies enquiries
  2. Conversations flow into Bitrix24 Open Channels, linked to leads and deals
  3. The client's team was trained to edit the bot's instructions themselves

Outcome: instant first response, with every chat captured in the CRM.

AI chatbotOpen ChannelsBitrix24
Healthcare clinicsDEMO

Conversational patient & appointment entry

Front-desk staff juggle patient calls while keying records into several screens.

  1. Staff describe the booking in plain language: "Book Mary for a health screening tomorrow 3pm"
  2. The agent extracts patient, service and time
  3. It creates the linked contact and appointment deal in the Bitrix24 healthcare pipeline and returns the link

Outcome: one sentence replaces a multi-screen data-entry routine.

ClaudeBitrix24 CRM
HR teams (product)PRODUCT

HRGenie — HR self-service agent

HR teams answer the same leave, claims and policy questions every week.

  1. Employees ask HRGenie about policies, leave balances and claims
  2. It answers from company policy and HR data
  3. It routes anything needing a decision to the right person

Outcome: HR gets time back from repetitive questions.

AI agentHR data
FusionETA — our own operationsIN-HOUSE

Subscriber onboarding agent

Each new software subscription needed a CRM deal, a draft invoice and a yearly repeating invoice, set up by hand.

  1. We tell Claude the client and plan
  2. It creates the yearly recurring deal in our Bitrix24 sales pipeline
  3. It raises the draft invoice and repeating invoice in Xero with the correct SST rate

Outcome: CRM and accounting stay in step for every subscriber.

ClaudeXeroBitrix24
FusionETA — our own operationsIN-HOUSE

Scheduled QA & content agents

Testing our HR app and publishing marketing content took hours of manual work every week.

  1. A scheduled browser agent runs clock-in / clock-out tests on our HR SaaS and reports results on WhatsApp
  2. A weekly agent turns our delivered work into new service pages for our website
  3. Both run unattended and report back for human review

Outcome: we run agents in our own business before we recommend them in yours.

ClaudePlaywrightWhatsAppScheduled