Matthew McGrory
speaker
331 appearances
2 recordings
1 series
first heard Mar 2025
last heard 7 Jul
Matthew McGrory’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsRecordings per month over the last 12 months — 1 in all, peaking in Jul 2026 with 1.
Appearances
Yeah, yeah, absolutely. I think there's this meta-release, this kind of metaverse world. and there's kind of this platform concept of everything being in the same place, making it really convenient. I don't have to set up an account to check out. I don't have to put my credit card details because they're all kind of stored somewhere.
So the whole pain of selection and transaction and buying, all that pain is very much taken care of. And yeah, I think we're headed for dass diese Plattformen noch stärker werden als sie schon sind, weil sie auf den hohen Weg gehen werden. Der hohe Weg hier in den USA wird immer kleiner und kleiner. Wir haben weniger Retail-Outlets und sicherlich mehr Restaurants.
We joke in the UK about the proliferation of charity shops. All our kind of market towns are filled with seven or eight charity shops, where if we went back 10, 15 years, there might be one or two. Yes, great for the charities, they're getting free retail store, but I think it's a product of the fact that a lot of this stuff is moving online. A lot of the experience of Wie kaufe ich Dinge?
Wie interagiere ich? Wie frage ich Fragen? Es wird ein bisschen mehr real. Und sicherlich spielt die AI-Revolution in das ein, weil ich Dinge mit AI erleben kann, die ich vorher nicht tun konnte. Ich kann mit einem Avatar sprechen in einer Art, die ich vorher nicht tun konnte, und es wird in einer etwas menschlicher Art reagieren.
Ich meine, wir sind noch bei Basis 1, aber hoffentlich kommen wir zu Basis 2 und Basis 3 relativ schnell, so sieht es aus.
Ja, genau. Wir haben hier eine Betriebsstelle, eine Betriebsstelle, die ungefähr 2.400 Bars in den USA besitzt. Und sie wollen Dinge wie, wir haben gerade unser neues Menü in einer unserer Brands geöffnet, das 200 Bars sein könnte. How did people feel about that? What was the sentiment? What was the most talked about thing? Did they like the burgers? Did they like the new ribs that we released?
Did we change the chips? Did anyone complain about the fries? Sorry, they're not chips, fries in your world. Yeah, fries.
I'm in America, yeah, yeah. So that's really where people want. They want that intelligence of what people are saying so they can react to it. That could inform product decisions that they're making. and help them respond to people in a more informed and grown-up way.
You've also got, we work with some sports teams in Formula One, and one of the teams came to us and said, well, we want to know what people are saying every 15 minutes throughout the race, because we want to react to it as the race happens. It's a two-hour race, you know, in hour three the story's dead, but we want to react to it, you know, 30 minutes in, something happens, there's a pit stop,
the lead changes, there's a safety flag, that type of stuff. So that's really, people want to, the creators want to react more quickly in a more authentic way to what the audience is talking about. Some of those channels get Ja, 200, 250.000 Kommentare über ein Rennabend. Also es ist viel Sachen.
Also bringen diese, diese Dinge auf die Oberfläche zu bringen, ist wirklich wichtig, damit sie authentisch in die richtigen Gespräche eintreten können, die mehr Entspannung generieren. Sie können es benutzen, um Denken Sie daran, wie die Pressekonferenzen... Sie können die Fahrer führen, um bestimmte Dinge in der Pressekonferenz zu sagen.
Das könnte sowohl um gute Dinge zu promoten, als auch um schlechte Dinge auszuprobieren. Ja, also es ist wirklich interessant, wie die Leute die Technik, die AI-Technik, in vielen verschiedenen Wegen benutzen, um verschiedene Signale auf Social Media zu holen.
Ja, also die Art und Weise, wie wir das betrachten, ist so, dass du deine Kunden, deine Kunden wirklich verstehen musst, was sie suchen. Also es könnte sehr oft ein PR-relevantes Message sein. Ein guter Beispiel ist, I interviewed someone from a very well-known global drinks brand. And they're very sensitive about people talking about them polluting the oceans with plastic.
So you may or may not be able to guess who I'm talking about. And they spend an awful lot of money cleaning the oceans of very... hunderten von Tonnen Plastik. Sie haben ein Wild-Preservation-Programm und sie wollen Menschen darüber erzählen. Was wir tun können, ist, dass wir die Botschaften nehmen, die sie suchen.
Und mit den Verbesserungen in der KI, speziell in großen Sprachmodellen, können wir synthetische Datensätze kreieren. Der Klient gibt uns
10 source comments and they kind of said we're looking for stuff that sounds like this and whenever we whenever that gets posted we want to lean into that conversation and we want to tell them look we're not polluting the oceans we're actually we're saving the whales we're saving the dolphins and we remove this so so pr and kind of crisis people are using these things to get to very specific content in very specific ways
Es ist ein bisschen technisch und ich werde nicht als Experte darauf sprechen. Aber es gibt eine Technik namens RAG, die Facebook entwickelt hat.
which takes a customer's own data, so their own voice, their brand guidelines, their brand voice, the responses that they give to their call center to talk to their clients, pulls that all together, they couple that with the large language model to generate authentic on-brand voice responses for clients. So they use it inbound, how can we detect
specific messages and they're using it outbound for we want the suggested replies. Normally it's suggested to a human agent and they kind of have the last call and edit it of how we respond to that. That's really come about in the last, I would say, 18 months to two years with where kind of open AI and the whole kind of large language model movement has gone and been able to take that.
We can do lots more than we used to.
Showing 281–300 of 331 · page 15 of 17
← Previous
Next →