Things I've shipped

The business had 12 years of history spread across disconnected tools: Zoho, spreadsheets and email. Data was siloed, reporting was manual, and there was no single source of truth for sales, marketing or operations.

Implemented Salesforce as the central CRM, migrating core data and rebuilding sales processes in Sales Cloud. Connected Rightmove, Zoopla and Xero so leads, stock and financial data flowed into one place.

Standard lead conversion couldn't model the reality of the business — a single property has purchase, sale and investor sides that all need to hang off one parent record. Trying to force it through the out-of-the-box flow left data fragmented and reporting unreliable.

Designed a custom conversion process that turns a qualified Lead into a Property record, then automatically creates and links the Purchase, Sale and Investor Opportunities against it — each with the right record type, stage and ownership.

Property profitability was being tracked across spreadsheets. Leadership had no real-time view of budget vs actual, gross profit or performance across the pipeline, and finance was burning hours reconciling numbers manually.

Modelled Property as the parent record with child Opportunities, Costs, Overheads and Budgets. Automated gross profit calculations tie listing, purchase and sale prices together, and everything rolls up into dashboards covering pipeline, budget vs actual, and negotiator performance.

The Property record had grown to depend on ~25 declarative rollups. Realtime calculations were breaching the SOQL 101 governor limit while full recalcs ran fine, meaning the numbers were correct in theory but unreliable in day-to-day use.

Rather than piling on more automation, stepped back and redesigned the rollup architecture — consolidating fields, rethinking which values genuinely needed to be realtime, and shifting the rest to scheduled recalculations to stay well inside platform limits.

Property data lived in Salesforce but listing it to the investor app meant manual re-entry. Offers, reservations and subscription payments coming back from the platform had no automated route into the CRM, creating admin overhead and visibility gaps.

A custom Apex integration layer serialises property records and pushes them to the platform's API. Inbound webhooks land offers, reservations and Stripe subscription/reservation payment events back on Salesforce records, keeping both systems in sync in real time.

Finance was manually re-keying invoice data from Salesforce into Xero every time a deal closed. Slow, repetitive, and a regular source of errors.

Built a Salesforce-to-Xero integration that passes contact and line-item data through automatically, creates the invoice in Xero, and updates the CRM record with the result.

Marketing comms sat disconnected from the CRM. Sends were manual, timing was inconsistent, and there was no clean way to trigger the right message off a real Salesforce event.

Configured Marketing Cloud journeys to trigger from Salesforce events, using the Contact (Who) relationship as the link. Records flow into the right journey automatically based on CRM stage and activity.

Sales reps were spending a significant portion of their day manually calling prospects and logging call notes. Follow-up tasks were inconsistent and dependent on individual discipline, leading to dropped leads and slow pipeline progression.

Connected an AI outbound calling tool to Salesforce using the record ID as the anchor. Once a call completed, the integration pushed the full transcription back to the Salesforce record, applied AI-generated sentiment scoring, and created follow-up tasks based on the content of the conversation — all without manual input.

Inbound calls from prospects and vendors were being handled manually, with notes logged inconsistently or not at all. There was no structured way to capture what was discussed or ensure the right follow-up happened afterwards.

The inbound AI system matched callers to Salesforce records using their ID or phone number, handled the call, and on completion pushed a full transcript back to the record. Sentiment was scored and follow-up tasks were automatically generated based on what was said during the call.

As the number of integrations grew, ad-hoc auth and shared credentials became a real risk. Each new integration needed a proper, auditable path in and out of Salesforce.

Set up Connected Apps with scoped OAuth, built custom REST endpoints for inbound calls, provisioned dedicated integration users with least-privilege permission sets, and locked down access with domain verification and IP allowlisting.