Portal Panel Punchup
- Mal McCallion
- 2 hours ago
- 3 min read

I spent Friday afternoon on stage at The Negotiator Conference, chewing over “AI – how is it shaping agency?” as part of a panel with Gary Barker (Canopy), Rameen Sorkhabi (Lightwork AI) and Zoopla’s Rich Hayes. By chance, Rich and I were seated side by side in a Question Time set‑up, which added a little extra spice to proceedings.
First, credit where it’s due. You never see Rightmove on these panels. The £400m‑revenue, agency-killing beast prefers sponsored, soft‑ball events to unscripted scrutiny. Rich fronted up, took questions squarely and showed genuine passion for what he’s building at Zoopla. In a business openly being priced for sale, that matters. I suspect one reason Silver Lake hasn’t pressed the emergency ejector seat is precisely because Hayes and his team are still pushing to create value.
And there is value. Zoopla has quietly rolled out some cool AI features; millions of homeowners have claimed their properties over the past 18 years; they’re comfortably number two in a market where number one is widely disliked. But to defend the traditional portal model against AI’s rising tide, Rich offered two claims - and avoided one question - that deserve proper scrutiny.
"Filters are what people prefer"
The notion that buyers love ticking price, bedrooms and location, then wading through pages of “near misses” because it lets them “consider more” ... that notion belongs to a different era. People rarely know precisely what they want at the start. That’s why conversational AI changes the game: describe the school run, the dog, the need for a studio over a garage, the two cafés you actually frequent and get back options that match the life, not the checkbox. Then iterate. “Show me the same vibe but fifteen minutes nearer the Elizabeth line.” This is more human, faster and – crucially – not restricted to the subset that maximises portal profitability. Hayes' argument is from the ‘newspapers‑said‑we’d‑always‑prefer‑print’ playbook, circa 1999.
"Voice search has been tried and failed"
Dismissing 'voice discovery' because Alexa in 2015 couldn’t hack it, as Hayes did, misses the point. AI understands voice, interprets it, responds to it - it doesn't just give you the weather or time your spuds. You can have a nuanced conversation with ChatGPT or Gemini while driving back from the shops, refine priorities on a dog walk, then pick up the thread on desktop with maps, listings and schools layered in. If Big Portal truly believes “nobody uses voice”, their decline will accelerate. Considering your life’s next chapter is an ever-evolving, mobile, moments‑based journey, voice is the most natural interface for many of those moments.
" .... "
"Why should I pay for a portal, if AI can find every property online"? This was a killer question from the floor. Rich attempted to swerve it but couldn't find an answer. If discovery no longer depends on a paid-for portal – and AI does the discovery – what, precisely, are agents funding when they pay a whole‑portfolio subscription? The answer can’t be “because we always have”. Pay‑per‑performance, unique content, off‑market exclusives, data services and audience extensions are viable futures. Blanket subs for commoditised listings are not.
So what do agents need to know after the debate?
- Build your moat with proprietary data: valuations that never listed, sales pre‑1995, hyper‑local knowledge. Train your own AI on it.
- Make conversational search live on your site first. Use an AI chatbot that can behave as you would; train it to be as knowledgeable and helpful as you on a great day.
- Prepare for portal shift. Demand free listings or remove stock from underperforming or outdated sites; really question the value of all‑you‑can‑eat portal subs.
Zoopla, to its credit, is at least in the room for this debate. Rich answered what he could and took the heat. Will Rightmove ever step up to equivalent interrogation? Agents should keep asking. When the discovery layer is AI‑led and impartial, the old answers won’t do.
