Olivetree Consulting — CRM & AI integration for growing businesses

Case study — real, still running

Fifteen months rebuilding one CRM, end to end

This is the build history of a live CRM for a multi-brand hospitality & travel business — reconstructed from real commit data, not a pitch deck. Every system below is in production today, handling real customers and real revenue.

Monthly build volume hover a column for the exact count

Foundation & hardening

Dormant plateau

Relaunch & consolidation

What the audits actually found

Real bugs, real stakes — not a features list

Anyone can list integrations. What actually matters is what happens when the data is wrong and nobody's noticed yet. A sample of what surfaced during ongoing ownership of this system:

Attribution audit

Phantom conversions inflating reported revenue

A mistagged click-trigger was logging every "Book Now" click as a full-price purchase — over a thousand phantom conversions feeding the ad platform's bidding algorithm, which had spent weeks optimising toward clickers instead of buyers. Traced across four separate platforms, deduplicated, and fixed at the source.

Root-cause debugging

A metric that wouldn't stay fixed

A key spend-pace figure kept freezing on stale data — fixed once, then found stuck again after redeploy. The real cause was a scheduled sync job that only backfilled a full year once, silently skipping the current year on every run after. Confirmed fixed by checking the actual scheduled-job logs, not just the code.

Trust in the small things

Conversations that looked out of order

Customers saw their own message threads reordering themselves, and replies from the CRM arrived looking like brand-new emails with no history attached. Both traced to timestamp and threading-header bugs invisible in code review — the kind of thing that quietly erodes a customer's trust in the whole system.

Blind-spot audit

Hot leads the system didn't know had already bought

A newer, self-serve booking flow wasn't wired into lead-scoring, follow-up, or ad-spend logic — customers who'd already paid were still being chased as cold leads and counted against marketing ROI. Found by deliberately auditing every service against every data source, not just the obvious ones.

How I work

Audit first. Then build. Then hand over something you actually own.

The most expensive CRM mistakes are invisible until someone checks. That's where every engagement starts.

Discovery & audit

I trace your actual data end to end — what your CRM thinks is true versus what's really happening in your inbox, your ad accounts, and your bank account. Most engagements surface at least one costly surprise here.

Build & integrate

Whether that's selecting and configuring a platform, wiring AI into your existing system, or fixing what's broken — every change is made against your real data, with a working system at every step, never a big-bang rewrite.

Handover & ongoing support

You get documentation your next hire can actually read, not a black box. Available for fractional, ongoing ownership if you'd rather not think about your CRM again.

Services

Four ways to work together

CRM selection & implementation

Choosing and configuring the right platform for how your business actually sells — Sugar, Zoho, HubSpot, Salesforce, or a fully custom build when off-the-shelf doesn't fit.

AI agent integration

Reply drafting, lead scoring, hot-lead detection, and follow-up sequences that run without a human remembering to send them.

Attribution & analytics wiring

Ad platforms, GA4, and your CRM made to agree with each other — so your marketing spend decisions are based on what's real.

Fractional CRM ownership

Ongoing support for a system that keeps working after launch — monitored, maintained, and improved as your business changes.

The CRM your business outgrew. The AI layer that actually does something.