When I talk with partners about AI, I hear a lot of questions. I also see the same blind spots repeat. Most of the problems are not technical. They are about how partners think about risk, ownership, and time. If you get those wrong at the top, it almost does not matter which AI tools you pick. The initiative will stall, or worse, quietly create risk without any durable advantage. Here are the patterns I see most often in CRE partnerships and leadership teams. 1. Treating AI As A One‑Time Software Purchase Partners love to frame AI as: “We just need the right vendor.” “Let’s pilot a tool and see the ROI.” That mindset works for point solutions like a new CRM plugin. It does not work for AI. AI adoption in CRE is closer to: Training a new class of analysts who work differently Re‑wiring how information flows through your firm You do not “install” that and move on. You invest in: Iterating prompts and workflows Capturing what works into playbooks Updating those playbooks as the market and tools change When a partner says “We’ll try a tool for 90 days and decide,” what they often mean is “we are not budgeting time or attention for the learning curve.” That is not a pilot. That is theater. 2. Delegating AI To IT Or Innovation, Not To The Deal Team Another pattern: AI “lives” with IT, a data group, or a single innovation champion The deal team watches from the sidelines On an org chart, that looks tidy. In reality: IT does not live inside your IC meetings The data team does not underwrite deals or negotiate LOIs The innovation person has no authority to change how partners actually work So you get: Cool demos A few internal decks No real change in underwriting speed, IC quality, or portfolio decisions The firms that make real progress put a partner and a working deal lead in charge, with IT and data in a supporting role, not the other way around. 3. Underestimating The Cost Of Doing Nothing Partners are good at weighing explicit costs: License fees Headcount Consultant hours They are much worse at pricing the opportunity cost of waiting, even when the evidence is in front of them. Examples of “hidden” costs: The 20–30 percent of analyst time still spent on tasks that could be 80 percent automated The missed deals that never get a serious look because your pipeline capacity is capped The weaker negotiating position when the other side is using AI for prep and you are not The compounding advantage competitors build as they capture and recycle their deal data On a single quarterly P&L, those costs are invisible. Across three to five years, they show up as “we just seem to be losing more close calls” or “our people leave for firms that run sharper processes.” If you are not explicitly tracking those costs, you will keep telling yourself that “waiting to see” is free. It is not. 4. Assuming AI Is “For Juniors,” Not For Partners I hear this one a lot: “Our analysts should be all over this.” “The kids know this stuff.” That is true and incomplete. If AI only lives at the analyst level, you get: Faster decks Cleaner models No change in partner behavior The real leverage comes when partners use AI for: Reading and summarizing key docs before IC Running alternative deal narratives in parallel Stress testing assumptions against actual history Preparing for negotiations and investor conversations The best firms I see have partners who: Ask AI blunt questions they would usually save for late‑stage debates Use AI output live in meetings as a neutral “third voice” that forces clarity Model that behavior for the rest of the team 5. Chasing Big Bang Use Cases Instead Of Boring Repetition Partners love a marquee use case: “Can we use AI to underwrite deals automatically?” “Can we predict the next breakout submarket?” Those are interesting questions, but they are not where the early returns live. The best returns come from boring work that repeats every day: Intake and triage of new deals First‑pass OM and lease abstraction Drafting and formatting IC memos Standard DD requests and follow‑up email flows If you insist on a “transformational” use case before you invest, you skip the part where AI proves itself on: Time saved Fewer errors Higher deal throughput You also deprive your team of the on‑ramp they need to trust AI on more complex work later. 6. Ignoring The Plumbing: Data, Templates, And Governance Many partners want to talk strategy, not plumbing. They ask: “What can AI do for our investment thesis?” and skip: “Where do we actually store models, memos, and outcomes?” “Are our templates consistent enough for AI to read them?” “Who is allowed to plug what data into which tools?” Without plumbing, you get: Random tools pulling from half‑broken folders Sensitive data pasted into public models by mistake No way to reuse or audit what AI produced Partners underestimate how much structure they need to provide: One clean folder or system of record for deals Standardized templates for IC, DD, and asset reports Simple rules for what