A Whitepaper by Nadi Ulhaque · Founder, REMI
The Institutional
Intelligence Thesis
Why the next decade of real estate investment belongs to firms that compound intelligence — not headcount.
Your best people only use a fraction of the data available to them. That's not a talent problem. It's a human one.
I spent years as a data product person inside financial institutions. The stakeholders I served were smart. Capable. Experienced. But they only used a subset of the data that was actually available to them.
Not because they were lazy. Because their methods were ingrained. Their workflows were habits. The way they screened deals, built models, and reported to LPs was the way they'd always done it.
AI has no such bias. It doesn't have a "usual way" of doing things. It goes the route that's most accurate. And it handles exponentially more data than any human ever could.
That's the blind spot. Not in the data. In the habits.
0+
Deals screened per fund per year — but dozens more never make it past the inbox.
Source: Industry benchmarks
0%
Average close rate on screened deals. 97% of the work is filtering.
Source: CBRE / JLL
AI doesn't replace your team. It makes every person on it 10x more dangerous.
The centaur model comes from chess. After Deep Blue, the best players weren't humans or machines — they were humans working with machines. The combination crushed both.
AI Handles
Data ingestion at scale
Pattern matching across 1000s of deals
24/7 monitoring
First-draft analysis
Historical context retrieval
Centaur Output
Machine speed. Human judgment. Institutional context.
Humans Own
Final investment decisions
Relationship management
Regulatory sign-off
Strategic overrides
LP communications
In a regulated environment like asset management, human oversight isn't a limitation — it's a feature. The firms that win will be the ones where AI handles throughput and humans handle judgment.
REMI doesn't just process your data. It remembers everything your firm has ever done.
Most AI tools start from zero every session. You paste in context. You explain your firm. You re-teach it the same things over and over.
REMI's context graph is a persistent, evolving map of your firm's entire institutional memory — every deal screened, every decision made, every override, every LP preference.
It fits each firm like a glove. What's feeding REMI is your data. What REMI outputs are faster, more accurate processes based on that data. And it gets smarter every single day.
Compound Intelligence
Every deal screened, every memo written, every correction made feeds back into the graph. REMI doesn't just learn — it compounds.
Firm-Specific Memory
Your buy-box. Your thesis. Your LP preferences. Your risk tolerance. REMI knows your firm the way a 10-year veteran does — except it never leaves.
Historical Replay
Query past deals. Compare current opportunities against historical comps from your own portfolio. Simulate outcomes based on what actually happened.
Three forces are converging. The window is open. It won't stay open forever.
AI Infrastructure Matured
Context windows went from 4K tokens to 200K+. Agent frameworks emerged. Reasoning models can now follow multi-step financial logic without hallucinating. This wasn't possible 18 months ago.
4K → 200K+
Context window growth
Cost Collapsed
The cost per token has dropped roughly 100x since early 2023. What used to be economically insane — running AI agents across every deal in your pipeline — is now viable at scale.
~100x
Cost reduction since 2023
LP Pressure Intensified
Higher rates. Compressed timelines. LPs demanding more transparency and faster reporting. Firms can't hire their way out of this — the talent market is too tight and the margins don't support it.
2–3 wks
Avg. quarterly LP report time
A $14.6 trillion industry. Still running on spreadsheets.
Alternative assets under management have exploded. The operational infrastructure hasn't kept up. That's the gap.
$0.0T
Global alternative assets under management.
Source: Preqin 2024
0K+
Mid-market RE and PE firms globally.
Source: Industry estimates
$0.0T
Projected alt asset AUM by 2028.
Source: Preqin forecast
$0K
Average annual cost per departing analyst — institutional knowledge included.
Source: Robert Half 2024
Four workflows. One platform. Human-in-the-loop at every stage.
REMI isn't a black box. Every output has a source. Every decision has a checkpoint. Your team stays in control — they just move faster.
Deal Screening
Ranked pipeline with risk scores, comp overlays, and thesis-alignment flags.
Agent Action
Screening agents ingest OMs, comps, and market data. Every deal is scored against your buy-box — not a generic one.
Human Checkpoint
Your team reviews flagged opportunities. Override, approve, or adjust the scoring criteria. REMI learns from each decision.
Due Diligence
Complete diligence packages with source citations and variance analysis.
Agent Action
Diligence agents abstract leases, cross-reference rent rolls, flag inconsistencies, and pull historical context from your prior deals.
Human Checkpoint
Analysts review abstractions and approve findings. Corrections feed back into the context graph.
IC Memos
Institutional-grade IC memos generated in hours, not weeks.
Agent Action
Memo agents assemble investment committee packages that cite your prior decisions, not just market data. They reference why you passed on similar deals last quarter.
Human Checkpoint
Investment leads review, edit, and sign off. The memo reflects human judgment, accelerated by machine throughput.
LP Reporting
Polished LP reports delivered in days — with full audit trail.
Agent Action
Reporting agents pull portfolio data, generate commentary, and compile quarterly narratives that match your LP communication style.
Human Checkpoint
Portfolio managers review and finalize. No more last-two-weeks-of-the-quarter fire drills.
The firms that win the next decade won't be the biggest. They'll be the ones that remember everything.
Think about what happens when a firm has been running REMI for two years. Every deal screened. Every decision logged. Every outcome tracked. Every LP preference internalized.
That firm can screen a new deal in minutes — and score it against two years of compounding institutional memory. Not market data. Their data. Their patterns. Their thesis.
That's not an incremental improvement. That's a structural advantage that compounds over time. And it's one that a competitor can't replicate by hiring more analysts.
"The best firms don't just have better processes — they have better memory. We're building the memory layer for institutional real estate."
Nadi Ulhaque
Founder & CEO, REMI
Key Takeaways
AI in institutional finance is the greatest tool to multiply the power, output, and quality of people.
Human-in-the-loop in a regulated environment is a feature, not a limitation.
The context graph — firm-specific, compounding, irreplaceable — is the moat.
Context windows, cost collapse, and LP pressure converged. The window is now.
The firms that start compounding intelligence today will be unreachable in 3 years.
REMI is built by operators who lived the problem. Not consultants who studied it.
Your firm learns once. REMI remembers forever.