At a Glance
| Industry | Commercial real estate / property management |
| Size | ~55 employees, downtown LA office + on-site building staff |
| Starting point | 7 technology vendors, no unified IT ownership, ungoverned AI use |
| Engagement length | 120 days |
| Headline results | 23% lower all-in tech spend · 7 vendors → 1 · ~65 staff-hours/month recovered |
The Starting Point: Nobody Could Answer One Simple Question
The firm managed a portfolio of commercial properties across greater Los Angeles. Growing company, healthy business — and an IT situation that looked like most companies their size: functional on the surface, expensive underneath.
When we asked leadership one question — “What do you spend on technology per month, all-in?” — nobody could answer it. Not because they weren’t paying attention, but because the answer lived in seven different places:
| Vendor | What They Provided | The Hidden Problem |
| Break/fix IT company | Hourly support | Invoices spiked unpredictably; paid when things break, not to prevent breaking |
| Cybersecurity subscription | Endpoint tool sold after a scare | Overlapped with another tool already in place |
| VoIP provider | Phones | No integration with anything else |
| Copier/print company | Printers + “network stuff” | Another party with admin access, zero accountability |
| ISP contracts (per building) | Connectivity | Nobody tracked renewal terms or pricing |
| Ad hoc Microsoft licensing | Email, Office, Teams | Purchased over years; managed by no one |
| Employee-expensed AI tools | Various | Sanctioned by no one — more on this below |
Seven vendors. Zero accountable for the outcome. When something broke between two systems, the firm’s controller spent her afternoons refereeing vendors who each insisted the problem was on the other side of the fence.
The Diagnosis: What a Four-Week Audit Actually Found
The first thing we did wasn’t install anything. It was discovery: every invoice, every license, every login, every recurring charge on every corporate card.
If you run a company between 30 and 150 people, some of this will sound uncomfortably familiar.
Finding 1 — Paying for 31% more software seats than employees
Departed employees whose licenses were never reclaimed. Three separate e-signature subscriptions running in different departments. Premium tiers purchased for features nobody used. The unused-license line item alone exceeded the cost of a gym membership for every single employee, every month.
Finding 2 — Their “AI adoption” was actually shadow AI sprawl
Fourteen employees were expensing personal AI subscriptions — several different tools, several different security postures, and zero data governance. Lease documents, rent rolls, and tenant correspondence were being pasted into consumer AI tools with no business agreement in place.
Leadership thought they “hadn’t adopted AI yet.” In reality, they’d adopted it fourteen times, badly.
The point most companies get backwards: the risk isn’t that your team will start using AI someday. It’s that they already are — and nobody has decided what data is allowed to go where.
Finding 3 — Manual work was hiding in plain sight (and not where they thought)
Leadership assumed their biggest inefficiency was in accounting. The audit found it elsewhere: roughly 40 hours a month in lease administration, where staff read incoming leases and retyped key dates and dollar amounts into spreadsheets, and another 30 hours tracking vendor certificates of insurance by hand, with Outlook reminders standing in for a system.
Seventy hours a month of skilled employees doing work that a well-configured document AI pipeline handles with human review in a fraction of the time.
Finding 4 — Double-paying to guard the wrong door
Two tools duplicated endpoint protection — both billed monthly. Meanwhile email, the number one wire fraud vector in commercial real estate, had default-tier protection only. In an industry where one compromised email thread can redirect a six-figure wire transfer, they were paying twice to protect the wrong door.
The Fix: Consolidate First, Then Automate
The order of operations matters, and it’s the part most firms get wrong. Bolting AI onto a chaotic stack just automates the chaos. So the engagement ran in two deliberate phases.
Phase 1 — Consolidation (Days 1–45)
- Collapsed seven vendors into a single managed services agreement: one monthly number, one accountable partner
- Reclaimed and right-sized every license; standardized on one collaboration stack instead of two and a half
- Replaced the overlapping security tools with a unified stack that closed the email gap — including payment-fraud-specific protections for wire instructions
- Delivered a full environment map owned by the firm, not the vendor
Phase 2 — Targeted AI, With Guardrails (Days 45–120)
- Rolled out a company-sanctioned AI platform under a business agreement with data controls, paired with a one-page acceptable use policy written in plain English
- Established an internal catalog of approved AI tools and a lightweight intake process for proposing new use cases — governance that lives in a system, not a forgotten PDF
- Deployed document intelligence for the two workflows the audit flagged: lease abstraction and COI tracking. Incoming documents are now read, extracted, and routed automatically; humans verify instead of retype
What we deliberately did not do: three proposed AI use cases were rejected during the pilot because the error-correction time exceeded the time saved. That discipline is the difference between AI optimization and AI theater.
The Outcome, in Numbers a CFO Cares About
| Metric | Before | After |
| Technology vendors | 7 | 1 accountable partner (plus ISPs — nobody escapes those) |
| All-in monthly tech spend | Unknown — literally | Known to the dollar, and 23% lower |
| Software seats | 31% over headcount | Matched to headcount, reviewed quarterly |
| Lease & COI administration | ~70 staff-hrs/month manual | ~65 hrs/month recovered; humans verify, not retype |
| Email / wire fraud protection | Default tier | Dedicated protections on the actual attack vector |
| Shadow AI subscriptions | 14 | 0 — not by banning AI, but by making the sanctioned option better than the workarounds |
Is Your Stack Due for the Same Audit? Five Signs
- No one can state your all-in monthly technology cost within 10% accuracy, in under a minute. You can’t optimize what you can’t total.
- Your vendor count exceeds your department count. Every extra vendor is another gap where accountability goes to die.
- Employees are expensing AI tools individually. That’s not adoption — that’s ungoverned data leaving your building.
- Skilled staff retype information that already exists in a document. Anywhere a human reads a PDF and types what they read, there’s measurable savings on the table.
- When two systems conflict, your staff mediates between vendors. You’re paying your team to do your vendors’ jobs.
The Takeaway
Consolidation is the prerequisite. AI is the payoff.
Companies that skip straight to AI tools on top of fragmented vendors and dirty data get demos, not results. Companies that clean the foundation first get compounding returns from every automation they add afterward.
Wondering what your own audit would find? Crimson IT offers a no-obligation vendor and license assessment for Los Angeles businesses — the same discovery process described above. The findings are yours to keep, whoever you work with next.
