This week's signal is the proactive agent-assistant: AI stops waiting for instructions and starts initiating. Realtor.com's RealAssist AI watches buyer activity and drafts your follow-ups; SERHANT.'s Dot reads your calendar, email, and CRM and assigns your next steps; Edensign stages your listing photos, then tells you which renovations might pay for themselves. The pattern is identical everywhere: the machine reads first, you approve after. Meanwhile the other fight escalates from threats to filings — CRMLS sued Compass before Compass could sue the MLSs, and the question of where listings are allowed to live heads for federal court. Attention is now automated. Judgment is still yours.
01
Make it shine
Tools that make every listing look irresistible.
2 stories
NewVideo & Visual
Edensign's AI stages your photos, then tells you which reno might pay off
In an October 6 interview with Inman, George Zheng — a Harvard Graduate School of Design-trained architect who previously worked at Perkins&Will — introduced Edensign, an AI startup incubated at Harvard Innovation Labs' Launch Lab X. The tool turns empty or cluttered room photos into staged, listing-ready images in about 15 seconds — but Zheng says staging is only the wedge. The company has expanded into what it calls “listing intelligence”: upload an address and a folder of raw photos, and the platform identifies room groupings, evaluates each room's condition, recommends high-impact improvements and staging choices, analyzes comps and marketing positioning, and suggests where the ROI is. Zheng claims the spatial model is trained to never alter architectural elements — walls, windows, floors — framing that as protection against MLS misrepresentation rules; he points to a Boston client where a suggested $30,000 kitchen update was projected to add $100,000 to the sale price. All capability and ROI figures are founder statements, not independently verified, and this is an early-stage startup with no published user-scale or review data.
Why it matters
Virtual staging is becoming table stakes, so watch what sits on top of it: condition scoring and reno-ROI guidance directly address the pre-listing conversation every agent has with sellers. The compliance framing is the detail to note — with disclosure rules tightening (see NSW in Issue #004, and CA AB 723), tools that promise not to alter architecture are worth shortlisting against cheaper generators that happily widen your windows. Treat the ROI claims as marketing until you see independent results.
Aftershoot launches Photon, a free RAW editor with an AI coach
On October 6, Aftershoot — best known for its AI-powered photo culling and editing for professional photographers — launched Photon, a free desktop RAW editor (macOS and Windows) that requires no credit card. The hook is a built-in AI Coach that tells you what to adjust in each photo and explains why, an educational tilt versus one-click auto-edit; you keep full manual control over every suggestion. It supports RAW and JPEG from recent iPhones, most Canon DSLR and mirrorless bodies, plus Fujifilm, Nikon, Sony, OM System, and Panasonic models, using a custom-built RAW engine with Adobe DNG color profiles. Note the fit: Aftershoot aims at wedding, portrait, and event photographers generally, not real estate specifically — but listing photographers shoot the same cameras and fight the same learning curve, and Photon is a legitimate no-cost alternative to a Lightroom subscription for agents learning to edit their own listing photos.
Why it matters
If your listing photos are shot by you on a decent camera and edited by whoever has time, a free coach-style editor lowers the skill barrier to consistent listing images. The practical move: hand this to the team member who shoots but doesn't edit — the AI Coach framing makes it a training tool, not just a filter button. Don't expect listing-specific presets; this is a general photography tool you're borrowing.
Tools that put every listing in front of the right eyes.
1 story
UpdateMarketing Stack
CRMLS sues Compass first; Unlock MLS refuses — and Compass says it will sue them both
On October 5, one day before Compass's October 6 deadline for MLSs to stop fining agents who publicly market office-exclusive listings, California Regional MLS (CRMLS) filed a 119-page complaint against Compass in the U.S. District Court for the Southern District of New York. CRMLS asks the court to declare its cooperation rules (Rules 7.9/7.9.1 — its version of NAR's Clear Cooperation Policy) lawful under the Sherman Act and California's Cartwright Act, and to permanently bar Compass from suing over them. The complaint also reveals enforcement scale: of 271,000+ listings submitted last year, only 89 Rule 7.9 fines were issued (8 to Compass agents); but in 2026 CRMLS has processed roughly 300 cases, about 75 involving a Compass for-sale property. The same Sept. 8 Compass demand letter went to Austin's Unlock MLS (17,000+ subscribers), which rejected it on October 6 — CEO Emily Girard called it a demand that “strikes at the very structure of the MLS,” noting a seller can already exclude a listing from MLS dissemination by signed certification plus a $100 fine. On October 7, a Compass spokesperson confirmed to Real Estate News it plans to sue both MLSs — CRMLS “in the next few weeks.” Compass's counter-framing: MLSs supporting 350,000+ agents across 12 states already permit public marketing of office exclusives, a company claim, not a verified figure.
Why it matters
This is the distribution fight of the year, and it is now actual litigation, not posturing. CRMLS's own numbers undercut both extremes: 89 fines last year suggests enforcement was rare, while the 2026 surge in Compass-related cases shows where the enforcement energy actually goes. Direct follow-up to Issue #005's ultimatum story. For agents: your MLS's office-exclusive rules are now subject to court outcomes — read your MLS's actual rule text and fine schedule this month, because the landscape under you may be rewritten by judges.
Turn attention into relationships, and relationships into commissions.
2 stories
NewMarketing Stack
Realtor.com puts RealAssist AI inside the Realtor.com+ agent workspace
On October 6, Realtor.com announced that RealAssist AI — its AI assistant, which launched for consumers in June — is now built into Realtor.com+, the portal's collaborative home-search platform offered exclusively through MLSs as a member benefit. The AI onboards clients and co-buyers, keeps buyer preferences and notes, auto-generates tailored home searches from MLS data, and sends agents a Daily Digest: each buyer's chat, search, notification, and price-change activity plus suggested next steps. It can also refine searches using professional-level MLS data fields and draft client messages — but agents must review and approve AI-generated tasks before anything executes. The same launch adds an RPR (Realtors Property Resource) integration with property insights and market-trend reports inside the workspace. Realtor.com says Realtor.com+ is live in 21 markets (four more coming, with Austin's Unlock MLS the newest addition) with 10% agent adoption — both company-reported figures, not independently verified. Notably, this announcement sits inside the Compass/MLS fight: MLS-distributed Realtor.com+ is now positioned as agent-first infrastructure for the same markets where the litigation is landing.
Why it matters
This is the portal inserting itself into the agent's daily workflow rather than the consumer's — a Daily Digest of buyer activity plus pre-drafted follow-ups is a genuine CRM-adjacent workflow, and if your MLS offers Realtor.com+ as a member benefit, it's newly available capacity at no extra stated cost. The practical test is the approval step: the value stands or falls on whether the drafted messages and digests are good enough that you only glance, not rewrite. Watch adoption versus availability — Realtor.com's own 10% figure is the honest baseline.
SERHANT.'s S.MPLE 2.0 starts assigning your day for you
SERHANT. unveiled S.MPLE 2.0 at Inman's Blueprint conference on September 23, and rollout to its agents began October 5, expanding across company markets through Q4. The new assistant, called Dot, is pitched as an “AI chief of staff”: where version 1.0 completed tasks the moment an agent asked, 2.0 reads the agent's business — calendar, email, CRM, pipeline — and assigns the next steps itself. Per the company, one objective can be broken into about 20 steps coordinated across five specialized AI agents, covering CMAs, listing presentations, marketing campaigns, transaction coordination, and client follow-up; compliance screening (Fair Housing Act, RESPA, Sherman Act) runs inside the process. The headline figures are all company-reported and unverified: 54,000+ requests completed for 2,000+ active agents, 86 NPS, nearly 28,000 admin hours returned — and the eyebrow-raising claim that agents who joined SERHANT. and adopted S.MPLE raised gross commission income by an average of 144% year over year versus their prior firm, with some over 500%. SERHANT.-only tool; no public pricing.
Why it matters
Notable as the most “agentic” brokerage-owned AI play yet — it is a blueprint other brokerages may copy or buy equivalents of, and the “AI leads the human” framing is the sharpest expression of this issue's pattern. Read the adoption numbers as vendor marketing, not evidence: the interesting part is the architecture (reads your business, proposes your day, you approve), which is exactly what RealAssist AI and Edensign's listing intelligence are doing in their own lanes. If your brokerage builds or buys one of these, the due-diligence question is data access: what exactly is the AI allowed to read inside your CRM.
41% of agents now use AI-generated content — and 75% of them write listing descriptions
NAR's 2026 REALTORS® Technology Report, released in late September and based on 1,200 agents, is the freshest read on what agents actually use. The MLS still rules at 96%, followed by e-signature (79%), showing tools (68%), CMA/pricing (59%), drone photography/video (48%), CRM (46%) — and AI-generated content at 41%, the seventh most-used tech in the stack, ahead of transaction management. Only 12% of agents say they aren't using AI and have no plans to, down from 32% in 2025; among AI users, 55% report a positive business impact. Budgets stay modest: about half spend $50–500 a month on tech, 22% spend more than $500, 18% under $50. What do the AI users ask AI to do? Writing listing descriptions (75%), social media posts (56%), and emails and follow-ups (52%). And the least-used AI applications are the judgment-adjacent ones: recommending next steps in a transaction, pricing or risk insights, auto client updates. Note the date: the report itself was released around September 22, just outside this issue's research window (October 5–8) — it's included because it is the newest primary dataset on agent tech, and it backs this issue's theme with data rather than vendor claims.
41%
of 1,200 agents surveyed use AI-generated content — the 7th most-used tech in their stack
12%
say they aren't using AI and have no plans to, down from 32% in 2025
55%
of agents using AI report it has had a positive business impact
75%
of AI users ask AI to write listing descriptions; 52% use it for emails and follow-ups
“The assistants now multiply. The judgment stays yours.”
This week's three launches are the same move in three lanes. RealAssist AI watches your buyers and drafts the follow-up. SERHANT.'s Dot reads your calendar, CRM, and pipeline and assigns your day. Edensign reads your listing photos and scores the rooms. None of them ask you to start — they start, and you approve. NAR's data backs the shift: 41% of agents already use AI-generated content, and the uses climbing fastest are the routine ones (descriptions, posts, emails). Meanwhile the least-adopted AI jobs are the judgment-adjacent ones — next steps, pricing risk, client updates — which remain stubbornly human. The machines have taken the watching. They have not taken the deciding.
This week, pick the AI tool that drafts or scores something in your business and audit one output before it ships: read the suggested next step, check the staged photo against the real room, verify the drafted message sounds like you. The agents who win the assistant era won't be the ones who delegate most — they'll be the ones whose approval layer is strongest.
Editorial disclosure: NanaStack is independent. Nothing in this issue is sponsored. Vendor and company claims are reported from the cited sources and are not independently tested: Edensign's 15-second staging and the $30,000→$100,000 renovation-ROI example (founder statements); Aftershoot's AI Coach teaching claims and camera-support breadth; CRMLS's characterization of Compass's strategy and Compass's “350,000+ agents across 12 states” framing; Realtor.com's 10% agent adoption and 21-market figures and effectiveness claims; SERHANT.'s adoption figures (54,000+ requests, 2,000+ agents, 86 NPS, 28,000 hours) and the 144% average GCI growth claim. The NAR 2026 Technology Report is primary research by the National Association of REALTORS® (1,200 agents surveyed), released around September 22, 2026 — outside this issue's research window (October 5–8) and included as the freshest available agent-tech dataset, as dated in the text. The Compass/CRMLS quotes are drawn from court filings and company statements as reported by Real Estate News; legal outcomes are unknown. The Photon story is a general photography tool, not real-estate-specific, flagged in the text.
