Every acquisitions team, asset manager, and investment firm we talk to is having the same conversation: How should we be using AI?
It's the right question. And the energy behind it is exactly right — the firms leaning into AI now are going to be in a much stronger position over the next few years. We use AI tools like Claude across our entire team every day, and it's made us significantly more productive across engineering, product, and go-to-market.
But after working with hundreds of deal teams and watching how this is playing out across the industry, we've seen a clear pattern: the firms getting the most out of AI aren't choosing between AI tools and purpose-built platforms. They're using both — and the combination is where the real leverage lives.
This isn't a vendor pitch. It's a framework for thinking about where to invest your team's time, budget, and attention.
The instinct to go all-in on AI is correct. The mistake is thinking that means you have to build everything from scratch — or that AI replaces the need for best-in-class tools.
Here's what we've found: AI doesn't replace your best tools. It makes them dramatically more valuable.
We run our entire company on a stack of best-in-class platforms — HubSpot, Gong, Mixpanel, Figma, ClickUp. Since we started weaving AI into our workflows, our usage and perceived value across all of them has exploded. Not because we replaced them with AI, but because AI lets us use the data we're creating and capturing in wholly new ways — and do a lot more with what those tools do best.
Every tool in our stack has a sweet spot — the thing it was purpose-built to do better than anything else. AI doesn't replace that sweet spot. It opens up entirely new value from the same tool.
HubSpot (CRM)
Best in the app: Pipeline management, deal stages, contact records, automated sequences.
Best in AI: Analyzing win/loss patterns across hundreds of deals, drafting personalized outreach sequences, synthesizing activity history into account briefs before a call.
The pattern: HubSpot is the system of record. AI makes the data inside it actionable in ways the UI was never designed for.
Gong (Conversation Intelligence)
Best in the app: Call recording, keyword tracking, deal warnings, talk ratio analytics.
Best in AI: Extracting feature requests across hundreds of calls, building competitive battle cards from prospect objections, summarizing months of deal context for handoffs.
The pattern: Gong captures the conversations. AI turns months of conversations into strategic intelligence you'd never find manually.
Mixpanel (Product Analytics)
Best in the app: Event tracking, funnels, retention curves, real-time user analytics.
Best in AI: Asking natural language questions about usage patterns, correlating feature adoption with retention, building product hypotheses from behavioral data.
The pattern: Mixpanel tracks every click. AI connects the dots between user behavior and business outcomes.
Figma (Design)
Best in the app: Design systems, prototyping, component libraries, collaborative design.
Best in AI: Generating initial layouts from product specs, writing component documentation, brainstorming UX flows from user research.
The pattern: AI accelerates ideation. Figma is where the actual design craft happens. The two together are faster than either alone.
ClickUp (Project Management)
Best in the app: Sprint planning, team workload views, dependencies, project tracking.
Best in AI: Drafting project briefs from strategy docs, breaking epics into tasks, writing release notes from completed tickets.
The pattern: ClickUp manages execution. AI handles the writing and planning work that used to slow execution down.
The through-line: In every case, AI amplifies the tool. It doesn't replace it. The tool captures and structures the data. AI makes that data work harder.
We believe the exact same pattern applies to Archer and CRE. AI is phenomenal at drafting memos, building dashboards, and automating workflows. Archer gives it the structured, verified, benchmark-enriched data that makes those outputs trustworthy for decisions worth tens or hundreds of millions of dollars.
Just as work migrated from the desktop computer to the smartphone — where some things moved and some things stayed — AI will absorb certain CRE workflows while others stay in purpose-built platforms. The key is knowing which is which.
These are workflows where the platform's infrastructure, data depth, and deterministic accuracy deliver value that AI can't replicate:
Parsing T12s and rent rolls with 99%+ accuracy across hundreds of PM formats. Populating your BYOM Excel model in under 60 seconds. Building proprietary comp sets that grow with every deal. Normalizing financials to your Custom Chart of Accounts. Expense benchmarking at the submarket and line-item level. Tracking deal history and institutional memory.
These are workflows where AI's reasoning and generation capabilities shine — especially when connected to verified, structured data:
Drafting IC memos from Archer-structured deal data. Asking natural language questions about your portfolio. Building custom deal screening criteria and dashboards. Synthesizing market research into investment narratives. Comparing deals across your pipeline in conversational format. Generating LP reports from portfolio analytics.
Over the coming years, CRE workflows will increasingly migrate into AI-first environments. Firms may spend more than 50% of their analytical time in tools like Claude, ChatGPT, or Gemini rather than traditional platforms. The platforms that meet clients where they're working — whether that's in the app, in Excel, or in an AI tool — will win. The ones that force you into a single UI won't.
This isn't a prediction. It's already happening.
Everyone talks about how cheap it is to prototype with AI. And they're right — it's never been easier to "vibe code" a quick tool.
What they don't mention: it's still expensive to maintain. But even that isn't the full picture. There are three costs that rarely get discussed:
Maintenance cost. It's cheap to build v1. It's expensive to keep v1 working across 100 different document formats, edge cases, and evolving requirements. The maintenance burden of a custom AI tool is comparable to traditional software — but without a dedicated team behind it.
Opportunity cost. This is the most expensive trade of all. Every hour your acquisitions team spends maintaining a custom parsing script is an hour they're not evaluating deals. Every sprint your engineering team spends building a financial normalization layer is a sprint they're not building the features that win new clients. Time is zero-sum. Spending it on infrastructure that already exists elsewhere is a strategic choice with real consequences.
Misapplication cost. This is the one almost nobody discusses. Just because you can use AI for something doesn't mean you should. AI parsing a simple spreadsheet that VLOOKUP handles perfectly. AI drafting a report when a template is faster. AI building a dashboard when your BI tool already has the data. There's a real cost to reaching for AI when a non-AI solution is actually better — slower execution, less reliable results, and the cognitive overhead of maintaining yet another workflow.
The goal isn't to AI-ify everything. The goal is to get more out of everything — and that means using the right tool for each job.
Not every AI project delivers the same return on your time and budget. Here's our honest framework.
IC memo drafts, market research synthesis, LP communications, custom dashboards and visualizations, internal process automation, and comp analysis narratives (when the underlying data is verified). These are excellent AI projects — the technology is mature, the outputs don't require deterministic accuracy, and you can get real value quickly.
Financial document parsing (99%+ accuracy on T12s and rent rolls requires 175K+ trained line items), custom Chart of Accounts normalization (every PM reports differently), proprietary comp sets (200K+ rent comps that compound with usage), BYOM Excel integration (living inside your model without changing it), expense benchmarking (submarket-specific, line-item level), and cross-deal institutional memory (every deal builds your data mountain). These require years of domain-specific training data and infrastructure that compounds with usage.
DIY parsing from scratch (works on the demo deal, breaks on the 10th), building your own comp database (without structured ingestion, you're maintaining a spreadsheet), custom financial normalization (6-12 months to reach 20-35% ceiling), AI-only underwriting (LLMs hallucinate numbers — underwriting requires deterministic accuracy), and maintaining AI tool sprawl (50 custom prompts across 5 tools with no institutional memory).
Generic CRE chatbots (chatting with an OM doesn't build institutional knowledge), point solutions without data moats (thin UI wrappers over public models get commoditized), platforms that force their model (your Excel logic is an asset), and unverified data sources (scraped public data without audit trails is a liability).
Meeting clients where they work isn't new for Archer. It's what we do. This is just the next evolution.
When every proptech company was building their own underwriting model and forcing clients to use it, we made a different bet. Clients had spent years refining their own Excel models — their logic, their assumptions, their structure. We didn't ask them to abandon that. We built BYOM (Bring Your Own Model) so Archer could live inside their model. Their formulas, logic, and outputs remained completely unchanged. We replaced the manual data entry. They kept everything else.
That bet proved prescient. BYOM is now our flagship integration and a core competitive advantage.
The same dynamic is playing out right now with AI tools. Clients are building workflows in Claude, creating custom agents, and developing their own AI capabilities. Just like with Excel — they don't want to abandon those investments. They want to supercharge them.
Our API and MCP strategy is the BYO-M of the AI era: BYO-AI. BYO-Agent. BYO-Dashboard. BYO-Platform. Your tools, our data infrastructure.
The form factor is evolving from Excel to AI. The principle is the same: Archer comes to you.
Every deal that flows through Archer makes the platform smarter — and makes your AI tools more powerful by extension.
Every rent roll you parse generates proprietary rent comps. Every T12 generates expense comps at your custom Chart of Accounts level. Every deal captures the data that powers future benchmarks, screening criteria, and investment analysis.
This is the compounding flywheel that no collection of individual AI tools can replicate: Parse → Map → Analyze → Comp → Underwrite → Share → Repeat. Each cycle builds your firm's proprietary data mountain. Each deal is faster and more accurate than the last. And when you connect AI tools to that data — via API, MCP, or the workflows we're building natively — the intelligence compounds even faster.
The firms that start building that data mountain now will have a structural advantage that widens with every deal. The firms that wait will find themselves trying to catch up to competitors whose AI tools are working with years of institutional memory they don't have.
The largest institutional CRE teams we work with aren't choosing between AI OR Archer. They're rolling out Claude /for their acquisitions teams and asking us about MCP connectivity, API access, and how to pipe structured deal data into their AI workflows.
Using Claude to draft IC memos from Archer-structured deal data. Building custom deal screening agents that query Archer's benchmarks. Automating LP reporting from portfolio analytics. Connecting AI tools to verified financial data via API and MCP.
The AI capabilities are real. The energy is right. The unlock comes from connecting that energy to data infrastructure that compounds.
The beauty of the "Yes, And" approach is that you don't have to choose.
On the AI side: Roll out Claude or similar tools for your team. Start with IC memos, LP communications, and market research synthesis. Build the muscle.
On the platform side: Start building your proprietary data mountain. Parse your deals. Build your comp sets. Map your Chart of Accounts. Every document you touch today becomes part of the institutional intelligence that makes every future deal — and every AI workflow — more powerful.
On the connection: As API and MCP connectivity comes online, the firms that have both capabilities — AI skills AND structured data — will be positioned to combine them in ways that neither can achieve alone.
The firms that win the next five years won't be the ones that chose AI or software. They'll be the ones that connected AI to the right data — and started compounding that advantage early.
Archer is the world's leading real estate intelligence platform for multifamily investment. We help deal teams parse, underwrite, benchmark, and analyze faster — with AI-powered features built on 5+ years of domain-specific training data. Learn more at archer.re.
Questions about how AI and Archer work together? Reach out: thomas.foley@archer.re