Property AI Report

Property AI Report co-hosts Mal McCallion and Matt Goddard cover a proposed private rental property database and the burden of property-by-property registration, then examine portal competition and AI-led property search. They also discuss Google's latest models, European open-weight AI, a major crypto theft, and Granola's expansion into meeting notes and voice workflows.

Chapters

00:00 Why the rental database workflow worries letting agents
02:08 Property-by-property registration and the case for bulk uploads
04:40 Will regulation and tax push responsible landlords out?
06:01 Domain challenges REA's market-leading claim
07:09 How agents are reducing dependence on property portals
10:13 Strutt & Parker brings natural-language search to property
12:26 Gemini's return and the race among AI model makers
16:39 Could AI answers undermine Google's search-ad business?
18:20 Genie world models and AI-powered creativity
22:08 Why European open-weight AI matters
24:20 The Binance theft and AI's cybersecurity arms race
26:11 Granola, voice-first AI and the value of free tools
29:50 Episode 101, EA Masters and what's next

Key Topics

Mal McCallion and Matt Goddard explain the proposed private rental database workflow, including separate records and agent invitations for each property, and why bulk uploads are a key industry request.
Matt cites portfolio scale examples: Lomond at roughly 57,000 properties, LSL at around 75,000, and Connells Countrywide at about 100,000, illustrating the administrative burden.
The hosts discuss concerns that rental regulation and a reported 2% increase in tax on rental income could weigh on responsible landlords, while failing to deter rogue operators.
Domain's challenge to REA Group's claim that nine in ten buyers find a home through its platform prompts a broader discussion of portal market claims and counterclaims.
Australian agents' efforts to reduce portal dependence include email campaigns, agency websites, social media and selectively withholding listings from portals; Mal also mentions free-listing alternatives and structured data for AI discovery.
Strutt & Parker's natural-language property search illustrates how AI could replace rigid search filters, while Mal stresses the importance of helping users when no listing matches every requirement.
Matt discusses Google's Gemini 4 Carbon, described as an advanced model not yet released, and the hosts weigh model competition against Google's concern that AI answers could reduce search-ad clicks.
Google's Genie world models and open-weight European models prompt discussion of creative possibilities, practical use cases and the value of competition beyond US AI labs.
The hosts discuss the reported $6.3 billion Binance crypto theft, describing cybersecurity as an AI-enabled arms race between attackers and defenders.
Granola's expansion beyond meeting notes leads to a discussion of dictation, spoken context as an input to AI, and questions about what free tools may gain from user data.

Send us Fan Mail

Hosts Mal McCallion and Matt Goddard mark the Property AI Report’s 100th episode with a look at alleged inaccuracies in Rightmove’s property-count claims, the unintended effects of rental legislation, and the government’s projected home-buying reform savings. They also unpack a major PHP security vulnerability, practical cyber-resilience steps, new AI agents, and a humanoid robot built for security and rescue.


 key  topics

Mal reviews the evidence and timeline behind Rightmove’s property-count claims, including the stated 900,000 homes for sale and 200,000 rentals versus an actual figure described as nearer 680,000 in early September 2026.
The hosts consider how portal audience claims may influence consumer choices, competitor visibility, agent costs and market competition, while emphasizing the need for accurate marketing claims.
Matt explains how removing rent-in-advance as an option under the Renters’ Rights Act may affect overseas students and other tenants who previously used upfront payment to meet referencing requirements.
They question the government’s estimate that home-buying reforms could save property firms £116 million a year, comparing it with roughly 24,000 UK estate agencies and a potential £5,000 per agency.
Matt discusses a critical PHP vulnerability, coordinated disclosure and patching, and the risk of exposed services and credentials; Mal shares that attackers used exposed API keys, generating $266 in ChatGPT usage.
Practical cyber-resilience advice includes auditing third-party access, reviewing vendor dependencies, rotating credentials, and having manual or offline processes for service outages.
The hosts discuss Anthropic’s possible IPO, Claude Duet as a personal AI agent, and the fast-closing gap between leading US models and Chinese open models.
Mal highlights Unitree’s Kung Fu-style humanoid robot, while Matt recommends secure password vaults and OnePassword’s agent-focused credential storage.

Chapters

00:00 Episode 100: celebrating a milestone
01:05 Rightmove’s property-count claims come under scrutiny
03:21 The timeline, figures and potential market impact
07:15 Why accurate portal marketing matters
11:07 Rent-in-advance rules and tenants who relied on them
13:27 Missing lettings fees and the £116 million savings estimate
16:42 PHP vulnerability: what happened and why it matters
20:32 Exposed API keys and third-party security risks
22:36 Cyber resilience: access audits and offline fallbacks
26:41 Anthropic, Claude Duet and the personal-agent race
30:15 Open models, humanoid robots and agent password security
34:21 A promised AI 101 for episode 101

Send us Fan Mail

Property AI Report - Renters' Rights, fraud risk, Rightmove's AI search, and the latest AI model shifts

This week, Mal McCallion and Matt Goddard cover the biggest property and AI headlines affecting agents, landlords, and tech teams. The episode looks at how the Renters' Rights Act is changing landlord behaviour, why fraud attempts are rising in lettings, and what Rightmove's new AI search means for property discovery.They also dig into major AI and infrastructure stories, including Microsoft stepping back from an Essex data centre project, the growing pressure on power supply for AI buildouts, new model launches from Meta, security concerns at OpenAI, and a practical tool for fast classification work from Cloudflare.

Key topics

• In this episode, Mal and Matt discuss how the Renters' Rights Act is pushing more landlords toward full management and away from self-management, especially as compliance gets more complex.
• They note a Pegasus Insight figure suggesting landlords are spending an average of 33 hours per month overseeing their portfolios, though they question how representative that number is.
• The pair cover a Paragon report showing buy-to-let yields reaching new highs, suggesting there is still momentum in parts of the lettings market.
• They discuss a Lettings Partnership survey finding that 35 percent of agents had experienced or identified attempted scams linked to their business, especially around financial transactions and client accounts.
• They explain how AI is making fraud easier through fake receipts, altered signatures, spoofed bank detail changes, and email compromise.
• In property tech, they look at Rightmove's new AI search feature, Ask Rightmove, which lets users ask questions and compare multiple properties side by side.
• Mal argues this is another sign that property search is shifting toward conversational discovery and that pay-to-list models may eventually face serious disintermediation.
• They cover Barry Plant, the Australian franchise network, and its Launch Academy training programme, which includes AI tools such as role play for sales conversations, predictive lead identification, and a vendor presentation app.
• On the AI infrastructure side, they discuss Microsoft pulling out of a planned AI data centre project in Essex, driven largely by power constraints and the broader challenge of building and cooling large-scale compute facilities.
• They also touch on the idea that the AI buildout may be bumping into a real-world bottleneck, with power availability and infrastructure timelines lagging well behind demand.
• In model news, they compare Meta's new Muse Wonder and Muse Wonderful releases with other foundation models, focusing less on raw capability and more on model personality and user experience.
• They discuss how model choice is increasingly shaped by tone, speed, and interaction style, with examples like Claude, ChatGPT, Opus, and Haiku behaving quite differently.
• The conversation shifts to security concerns at OpenAI after the FBI arrested an employee accused of planting a recording device in a boardroom.
• They also cover the OpenCode API key breach, where exposed keys could be used not just for access, but for running up costs, mining cryptocurrency, and further abuse.
• Finally, they look at Cloudflare's Clairic model, a fast system-one extraction tool designed for classification tasks like pulling names, order numbers, and descriptions from text quickly and cheaply.

Timestamps

00:00 - Weekly intro and what the episode covers
 00:44 - Renters' Rights Act pushes landlords toward management agents
 02:56 - Buy-to-let yields reaching new highs
 03:21 - Fraud attempts rising across lettings businesses
 06:09 - Rightmove launches Ask Rightmove AI search
 08:03 - Conversational property search and the future of pay-to-list
 09:00 - Barry Plant and Australia's AI-driven Launch Academy
 12:30 - Why repetitive agency tasks are the best automation targets
 13:25 - Microsoft pulls out of Essex AI data centre project
 14:35 - Power, cooling, and the limits of AI infrastructure
 17:08 - Neo clouds, capacity bottlenecks, and the AI buildout
 18:12 - Meta's Muse Wonder and Muse Wonderful models
 19:53 - Model personality as the new UX
 22:16 - Alleged OpenAI boardroom recording device incident
 24:18 - OpenCode API key breach and downstream abuse risk
 26:05 - Why resilience matters when cloud services go down
 27:00 - Elon Musk's proposed share gift for American children
 29:19 - Bank of America hires contractors to fix AI-discovered bugs
 31:02 - Cloudflare's Clairic and fast extraction models
 33:43 - Closing thoughts and sign-off

Notable quotes

Copy

"Consumers are hugely catching onto this idea that they can just vibe search."

Copy

"The new UX of these things is the personality."

Copy

"If you don't need a wealth of text coming out of the back end of your request and you just want a probability on something happening then these are much more efficient."

Send us Fan Mail

Property AI Reports - Weekly Digest on Property and AI

Mal McCallion and Matt Goddard cover the latest property market headlines alongside the newest AI product updates, with a strong focus on how automation, compliance, and visual AI tools are reshaping the property sector. The episode moves between industry news, practical AI use cases, and a cautionary look at model safety and regulation. It is especially useful for property professionals trying to understand where AI can create operational advantage without losing sight of security and compliance.

Key topics

• In this episode, Mal and Matt discuss Moveli's AI push for self-employed agents at the top end of the market, and how tools for valuations and follow-up are becoming a competitive wedge in prime agency.
• They compare that move with Knight Frank's associate self-employed model and the growing battle for high-performing agents who work in the luxury segment.
• We discuss the impact of the new Awaab's Law in private rentals, with Matt highlighting the added compliance burden around damp, mould, and property safety.
• Mal and Matt look at how AI and automation could help agents manage rising regulatory workloads without simply adding more staff.
• They revisit Purplebricks, Sir Charles Dunstone's continued funding, and the company's growing losses, including the challenge of keeping the Bruce brothers' strategy on track.
• In AI news, they break down ChatGPT 6's new intelligent UI, which creates more visual, interactive responses with graphs, checklists, and image-led explanations rather than plain text alone.
• They explore OpenAI revenue figures being compared and potentially overstated in investor materials, and the ongoing race with Anthropic and other major labs.
• They cover Claude's new dashboard capabilities, including data visualisation through business tools like Salesforce and CRM integrations such as Reapit.
• They discuss Satya Nadella's warning that AI models need an emergency brake, focusing on the need for tamper-proof, human-readable evidence and stronger controls.
• They examine Anthropic's Claude Haiku 5.5, including its 90 percent API price reduction and the practical role of smaller models for fast, lower-stakes tasks.
• They wrap with Meta's Muse Wonder AI models, including the commercial push around a paid version and the wider hardware ecosystem being built around it.
• The Weird AI of the Week segment covers an Anthropic model accidentally submitting a false tip to the Philadelphia Police Department while testing websites.

Timestamps

00:00 - Welcome to the Property AI Reports and this week's agenda
 00:40 - Moveli's AI tool for self-employed prime-market agents
 01:09 - Knight Frank's associate self-employed model and talent competition
 02:38 - Why agencies are building proprietary AI tools in-house
 04:18 - Awaab's law and the rising compliance load in private rentals
 05:44 - Using AI to manage regulation without adding headcount
 06:14 - Security risks when automating property operations
 06:44 - Purplebricks losses, Sir Charles Dunstone, and the Bruce brothers' return
 08:01 - How much money has gone into the Purplebricks venture
 09:54 - ChatGPT 6 and the new intelligent visual interface
 10:42 - Why visual AI responses matter more than plain text
 12:19 - Property's need for image-led AI workflows
 14:54 - OpenAI revenue figures and investor comparison noise
 17:39 - Claude dashboards and business data visualisation
 18:44 - Connecting CRM data into Claude for live reporting
 20:42 - Why AI interfaces are moving beyond text generation
 21:11 - Satya Nadella's call for an AI emergency brake
 22:17 - Tamper-proof human-readable evidence and model safety
 24:01 - Claude Haiku 5.5 and cheap fast-task automation
 25:13 - Using smaller models for research and extraction tasks
 26:45 - Meta's Muse Wonder AI models and monetisation
 28:53 - Weird AI of the Week: Anthropic model sends false police tip
 31:33 - Tool of the Week: why GPT 6 is worth trying
 32:28 - Episode 101 will be an AI 101 starter episode
 33:04 - Closing remarks and what's coming next

Send us Fan Mail

Property AI Report — Weekly digest of property and AI news

Mal McCallion and Matt Goddard cover a busy week across property and AI: new rental regulation, rising mortgage rates, a major investment in conversational AI, and announcements from OpenAI, Apple, Google, and Microsoft.

Key topics

• New rental regulations: The hosts discuss the PRS rental database rollout, rent control proposals, and damp and mould law changes—and what compliance could mean for landlords and letting agents.
• Property Franchise Group invests in Enteka: The company has taken a 25% stake in the AI firm, whose tools support call handling and lead enquiries across voice and WhatsApp.
• Mortgage rates rise: The hosts discuss the disappearance of cheaper fixed-rate deals and the pressure that higher borrowing costs may put on buyers and agents.
• OpenAI Dev Day: They unpack announcements including Dots, sign-in with ChatGPT, Pages, Sites, and new pricing tiers.
• How people use AI: A discussion of paid AI adoption, and the difference between casual use and integrating AI into everyday workflows.
• AI, permissions, and platforms: The conversation explores Apple’s tighter Mac OS access controls, Google’s upcoming model, and AI’s growing ability to use a computer.
• Tool of the week: GitHub Copilot adds computer-use capabilities, prompting a comparison with other coding agents.

Timestamps

00:29 — Welcome to the Property AI Report
 01:09 — Rental database, rent controls, and damp and mould law changes
 02:15 — The cost and burden of property regulation
 03:14 — Rental database rollout begins in the West Midlands
 04:47 — Property Franchise Group takes a 25% stake in Enteka
 06:02 — Conversational AI for estate agency call handling and lead enquiries
 08:02 — Cheaper fixed mortgage deals disappear
 09:37 — How a tougher market could affect estate agents
 10:04 — OpenAI Dev Day announcements
 11:12 — Dots and the idea of multiple personal agents
 12:38 — OpenAI’s sign-in, Pages, and Sites announcements
 13:29 — OpenAI pricing and the new Pro plan
 14:24 — Paid AI adoption and how people use AI
 15:53 — AI and recent maths breakthroughs
 18:48 — The proposal to rebrand AI as “SI”
 22:28 — Apple tightens Mac OS access controls
 23:40 — Why granular permissions matter for AI agents
 28:01 — Google’s upcoming Gemini model
 29:58 — Tool of the week: GitHub Copilot gains computer use
 31:22 — Mal’s take on Copilot, Codex, and Claude Code

Send us Fan Mail

Property AI Report — Weekly Property and AI News Roundup

Mal McCallion and Matt Goddard cover a packed week of property news and AI launches, focusing on what matters for agents, portals, and proptech operators. The episode ranges from Purplebricks’ latest funding and HMRC story to new AI models, personal agents, voice tools, and the direction of regulation and licensing in the UK property market.The hosts explore major industry shifts, the growing role of vendor-paid advertising, and how to assess AI tools by benchmark, use case, and workflow—not hype.

Key topics

• Purplebricks faces fresh scrutiny after a report about an HMRC winding-up order, followed by a denial and renewed focus on its financial position.
• The business has reportedly received another £34 million lifeline from Sir Charles Dunstone.
• Purplebricks’ accounts were still overdue at the time of the discussion, including filings for the year ending March 2025. The delay was attributed to a new CFO appointment.
• REA Group has taken a 35% stake in Daft, the major Irish property portal—a move the hosts connect to the rise of vendor-paid advertising.
• Vendor-paid advertising could signal a major shift in portal revenue models, with sellers funding marketing directly, as is common in Australia.
• The UK government has confirmed plans for mandatory licensing for agents. The hosts discuss how regulation might target rogue operators without unfairly burdening reputable lettings businesses.
• A 2.5% Help to Buy-style initiative is discussed as a possible stimulus in a flat, cautious market.
• The episode covers new AI model releases and the trend toward more capable intelligence becoming available at lower cost.
• Model personality is emerging as a form of user experience, including why some users switched away from Opus 5.
• Personal agents are developing into lightweight task runners for inbox curation, LinkedIn triage, shopping, and other repeatable jobs.
• As AI systems become more capable and autonomous, the hosts consider calls from AI leaders for coordination and regulation.
• 11Labs v4 is part of the broader voice AI wave. The hosts argue that good architecture matters more than simply plugging in a voice API.
• The practical tool of the week: use voice input with large language models, especially for brain dumps and drafting.

Timestamps

00:00 — Weekly intro and episode overview
 00:52 — Purplebricks winding-up order story and the HMRC confusion
 01:22 — Another £34 million lifeline for Purplebricks
 02:44 — Overdue accounts and the new CFO explanation
 03:13 — REA Group buys into Daft in Ireland
 04:12 — Why vendor-paid advertising matters for portals
 05:35 — Portals, revenue diversification, and charging consumers more
 06:33 — Mandatory licensing for agents announced
 08:22 — Why regulation is welcome, but blanket industry bashing is not
 08:51 — The 2.5% Help to Buy-style initiative
 09:45 — Major AI releases and the pace of model launches
 10:12 — ChatGPT 6, Sol, Luna, and cheaper intelligence
 11:05 — Why model personality is becoming the new UX
 12:28 — Anthropic’s focus on consistency and model personality
 13:21 — Personal benchmarks for choosing the right AI tool
 14:17 — Testing models against property-writing tasks
 15:15 — Transition into personal agents
 15:44 — What personal agents are and how they differ from workhorse models
 16:39 — Using agents to manage LinkedIn and inbox noise
 17:36 — OpenClaus, Grokbot, and the trade-off between flexibility and convenience
 18:34 — Practical use cases like shopping and scheduled tasks
 20:29 — AI safety concerns from the AI “godfathers”
 21:24 — 11Labs v4 and the rise of usable voice AI
 22:21 — Why shallow AI voice integrations fail
 23:15 — Architecture, latency, and the uncanny valley problem
 24:14 — Meta Connect, Muse, and Zuckerberg’s latest AI push
 25:33 — The week’s practical tool: use voice input instead of typing
 26:02 — Wrap-up and closing remarks

Notable quotes

• “Intelligence is getting better and is getting cheaper.”
• “The personality of the model is actually the new UX.”
• “Start using voice rather than trying to type everything.”

Send us Fan Mail

In this episode, we explore the latest developments in property portals, AI innovations, and their impact on the real estate industry. Join us as we analyse market shifts, new technologies, and future trends shaping the property landscape.

Chapters

00:00 Introduction to property market shifts and portal valuations
01:59 Shareholder activism and valuation changes in property portals
03:57 Data ownership, trust, and the future of property portals
06:06 Launch of MyPorta and market competition
07:54 Changing agent behaviours and portal dependency
09:59 Legal and contractual shifts in Australian property portals
11:49 AI's positive impact on healthcare and science
15:50 Emerging AI tools and personal assistants
20:00 Concerns about AI language and communication
24:01 Innovations in AI classification and response systems
27:47 Whimsical AI and language development
31:01 AI safety, IPOs, and industry outlook
33:10 Upcoming events and closing remarks

Send us Fan Mail

In this episode, we explore the latest developments in property and AI, including cyber security risks for estate agents, rising rents, AI safety concerns, and exciting tech innovations like foldable phones and advanced math problem-solving. Join us for insights into how these trends impact the property industry and the future of AI.

keywords

property news, AI safety, cyber security, rent increases, foldable phones, math problems, AI regulation

key topics

• Cyber incidents and insurance for estate agents
• Rising rents and government regulation impact
• AI safety concerns and recursive self-improvement
• Latest tech innovations including foldable phones
• Mathematics breakthroughs with AI

Sound Bites

• "We need to slow AI development to ensure safety"
• "AI models are now talking in bytes, not language"
• "There's a 10% chance AI could kill all humans in a decade"

Chapters

00:00
Introduction to Weekly Property and AI News

02:29
Cyber Security Risks for Estate Agents

04:57
Impact of Rent Regulations on the Market

07:19
Government Impact on Private Rented Sector

09:24
AI Frontier Pacing and Safety Concerns

13:18
AI Self-Improvement and Neural Language Models

18:36
AI Escaping Control and Ethical Risks

20:13
Exciting Tech: Foldable Phones and Future Devices

21:57
Mathematical Breakthroughs and AI Problem Solving

25:52
Introduction to Agents API and AI Tools

Send us Fan Mail

Property AI Report

This weekly roundup covers the biggest moves in property and AI, with a strong focus on portal competition, CRM investment, and the latest model releases. Mal and Matt break down what the changes mean for agents, vendors, and the future of property tech. It is a fast-moving update for anyone tracking where leads, data, and automation are heading next.

Key topics

• In this episode, Mal and Matt open with Zoopla’s return to profit, despite lower revenue than last year, and discuss whether that signals a stronger position in the portal market.
• They compare Zoopla and On The Market, with Mal noting that Zoopla’s lead quality seems to be shifting toward vendor leads and homeowners considering a move.
• The conversation turns to Houseful, the parent company above Zoopla, Uswitch, and Alto, and the broader struggle among portals to prove value to agents.
• In property tech, Reapit’s plan to spend £40 million over the next year on becoming an AI dominant CRM is a major theme.
• Matt frames that investment as part of a wider land grab, where CRM providers are trying to lock in large agencies before they switch to AI first alternatives.
• They stress the importance of keeping and integrating client data when moving CRMs, warning that leaving data behind creates long term pain.
• A political property story follows, centred on an Iranian businessman selling high value London properties while allegedly linked to the Supreme Leader of Iran, raising sanctions and compliance questions.
• The AI section starts with ChatGPT 6 Astra, which Mal describes as unusually strong at computer use, with the model actively taking control of his screen during the recording.
• They then cover Claude 5.1, comparing it with Astra and noting that both represent a real step change rather than a minor iteration.
• OpenClaw 2.0 is discussed for its new multiplayer setup, allowing multiple users and teams to work in a shared AI environment.
• World Labs Atlas stands out as the most visually ambitious release, generating and simulating 3D worlds from a few photos, with obvious implications for property walkthroughs and future robotics.
• The week ends with Obliteration.ai, a stripped down model with guardrails removed, which they describe as a terrifying crime large language model.

Timestamps

00:00 - Weekly property and AI news roundup begins
 00:24 - Zoopla returns to profit and pressures portal rivals
 04:37 - Reapit’s £40 million AI push and CRM land grab
 10:22 - Iranian property sale raises sanctions and compliance issues
 12:30 - ChatGPT 6 Astra arrives with powerful computer use
 17:25 - Claude 5.1 and the rise of step change model releases
 19:05 - OpenClaw 2.0 introduces multiplayer AI workflows
 21:08 - World Labs Atlas and the future of 3D property worlds
 25:51 - Obliteration.ai and the dangers of removed guardrails
 27:41 - Tool of the week and why this release cycle matters
 28:32 - What Mal and Matt are planning next week

Notable quotes

• “If anybody's joining another CRM system and getting rid of their data, don't. It's stupid.”
• “It is insane.”
• “People out there, go play, it's great, great fun.”

Send us Fan Mail

Key  topics

Dwelly's acquisition of Paramount and its implications
The $1.7 million investment in lettings technology built on Rex
The future of surveyors and AI's role in streamlining reports
NVIDIA's acquisition of Hugging Face and its impact on open source AI
Salesforce's integration of Claude into its ecosystem and the SaaS market shift


 sound bites

"AI is transforming surveyor reports and industry recruitment."
"NVIDIA's move into open source AI signals a new era."
"Open source models are making AI more accessible and affordable."


Chapters

00:00 Introduction to property tech and AI trends
01:59 Dwelly's strategic acquisitions in the lettings market
04:02 The £1.7 million investment in Rex-based lettings system
07:09 Surveyor shortage and AI's role in report automation
10:05 AI's impact on surveyor recruitment and industry challenges
11:50 The debate over AI-generated journalism and content creation
15:12 The evolution of CRM pricing models and AI integration
17:01 Salesforce's partnership with Anthropic and AI's future in SaaS
22:08 NVIDIA's acquisition of Hugging Face and open source AI advancements
25:59 The rise of open source models and their industry implications

Send us Fan Mail