There will continue to be downward pressure until the next market slow down. When homes aren't selling but you have sellers that need to sell, they'll happily pay 6% for an agent who will produce results. After the 2008 crash agents who could show a track record of successful results when nothing was moving were pushing 7-8% commissions. The barrier to entry is far too low in residential and as touched on before allows for too many part-timers to enter the market and eat away at market share. It's counter-intuitive but I think that if the barrier to entry was greater, commissions might actually soften a bit. When full-time, professional agents have to spend more time and money fighting for market share, what they can afford to charge at a minimum goes up as their volume goes down. This is a business cost that few consumers ever consider when factoring where the commission goes to - client acquisition. If it were as simple as "Get License, Sell Houses" then no one would ever flunk out of the business, but a large expense for agents is their personal marketing, not actual property marketing and as commissions are the only stream of income, that cost has to be baked in along with actual property marketing costs, overhead, broker splits etc.
Door Dallas is a VC funded brokerage that hires freshly licensed agents to work for a salary, if inexperience is what you want, you might save some money, you might not. I assume their end goal is to be bought by a Zillow or other large umbrella wanting to get in the retail side of the business. But in order for their model to be profitable each salaried agent will have to cover a large amount of business, and obviously service and quality suffers. Based on how much advertising they do on the radio, their client acquisition costs must be insanely high right now, helps to be VC funded, but I'd l like to see their P&L statement.
Opendoor uses a code based lock system to allow anyone to tour their houses at anytime, but they can get away with that because they are the actual owner of the property and they are vacant. While they have a camera system in the homes usually, i imagine they've gotten knocked off a few times at this point. As mentioned there was a group about 2-3 years ago that was trying to Uber-ize the house-showing activity but it never got off the ground, my guess is due to Texas' confusing and strict agency laws. As soon as a principal sues because they thought they were represented by the person opening the door...and then subsequent suits are brought, it kills the whole plan. So maybe blame lawyers there.
I do think it makes more sense for a buyer to actually hire and pay their own broker, but it adds additional expense to a buyer already shelling out tens of thousands of dollars, and might require them to actually start making some payments along the way of the process. Few industries do so much work completely on contingency - would an attorney, CPA, architect, contractor etc work for 3-6+months and never send a bill? If consumers would be more willing to pay-as-they-go, commissions would also soften.
In my opinion the intricacies and nuances of individual neighborhoods and the homes therein are too great for a large scale flat-fee, automated or AI based listing service to provide quality representation at this point - if all you care about is cheap, you can definitely get that. If anything I can see, and I am trying to create within my own business, a hybrid model of flat-fee geographic experts. If you are positioned as the regional (more like neighborhoods, or city quadrants, not entire metropolises) and have a flat fee system, I think that would be attractive to sellers. But it's an uphill battle and few of us have the funding.
If anything the commercial leasing side of things seems most prime for an automated showing disruption, and the elimination of buyer's reps, especially when a comprehensive commercial MLS happens.