August 21, 2026
12 min read
The Agentforce World Tour is Salesforce’s global series of one-day events dedicated to the future of AI, autonomous agents, and the rise of the agentic enterprise. At each stop, customers, partners, and Salesforce leaders come together to explore emerging technologies, share real-world success stories, and discuss what’s next for the ecosystem. For anyone looking to understand where Salesforce AI is headed, these are among the year’s most important events.
To get an inside perspective, I sat down with Kaytlin (Kayt) Parrella of 28 North Consulting and Raul Acosta of Klient. Between them, they’ve attended five Agentforce World Tour events this year, with more still on their calendars, giving them a unique view of the conversations happening across the Salesforce community.
At 28 North, Kayt helps organizations design scalable go-to-market operating models by aligning systems, processes, and technology. Her work focuses on helping companies get more value from their revenue technology stack, with deep expertise in Salesforce.
Raul serves as Director of Global Partnerships at Klient, a Salesforce-native professional services automation platform that helps services organizations manage projects, resources, time tracking, and financials. As AI becomes increasingly embedded in professional services, Klient is focused on enabling teams to work more effectively alongside AI agents.
As longtime partners, 28 North and Klient collaborate closely to help customers maximize their Salesforce investments. I spoke with Kayt and Raul about the biggest themes they encountered across this year’s Agentforce World Tour events, what they’re hearing from customers, and where they believe AI, and the Salesforce ecosystem, is headed next.
Interviewer: Thank you both for talking with me. Between you, you’ve attended five Agentforce World Tour events this year. What themes and conversations are you hearing most often at these events?
Kayt Parrella: The biggest theme they’ve been pushing is that AI is here. Organizations need to leverage it if they want to stay on par with everyone else or get ahead.
But largely, nobody’s ready for it. Everybody’s stuff is a mess. If you put AI into a mess, you’re just going to accelerate your mess.
That’s really the theme I took away. Even the breakout sessions focused on the same thing: How do you clean this up so you can use AI? Everything came back to getting your foundation in order first.
Raul Acosta: I agree with Kaytlin, but also, it’s interesting because the messaging has evolved throughout the year.
At the beginning of the year, a recurring theme was the agentic enterprise. The message was that agents are becoming part of the operating model—they’re not just a bolt-on. They’re becoming infrastructure.
Around April, though, everything shifted. Salesforce started emphasizing what they call Headless 360, and that became the backbone of nearly every keynote after TDX 2026.
Now the conversation is much more about Salesforce as an API, Model Context Protocol (MCP), agent-to-agent orchestration, governance at scale, and the API patterns underneath it all. That message is starting to permeate almost every piece of content Salesforce is publishing.
Kayt: What’s interesting is that you can tell Raul works for a product company because his lens is different than mine.
He’s talking about MCP and how everything connects together. When I’m walking around these events, I’m hearing customers ask, “How do I actually use this? How do I embrace it?”
We definitely have different perspectives, and that’s why I enjoy having these conversations. Raul: I agree, but we’re also learning from customers.
Since the beginning of the year, we’ve been getting constant feedback on how organizations are actually using AI models, where they’re interested in MCP, and where they’re hesitant about Agentforce.
You can tell Salesforce is still figuring out exactly what the message should be. It’s a very fluid situation, which makes it challenging for partners because the conversation keeps evolving.
Interviewer: How are these topics showing up in conversations with your clients and customers?
Kayt: A lot of customers ask, “Why don’t we just build it ourselves?”
We could spend hours talking about that.
Take professional services automation, for example. If a company decides to build its own PSA tool, it doesn’t have years of product experience behind it the way Klient PSA does. They build something that works for the way they operate today instead of incorporating industry best practices.
Then they have to support it themselves.
You end up with an isolated tool that nobody is continually improving, and that becomes difficult to maintain over time.
I’m sure you’re hearing the same conversations, Raul.
Raul: Absolutely.
One of the biggest conversations we’re having right now is around cost forecasting.
Salesforce can demonstrate great Agentforce use cases, but nobody demonstrates a credit consumption forecast.
Even when we sell our own agents, the real bill comes when customers start purchasing Salesforce consumption credits. It’s difficult to model what that consumption will actually look like.
Because of that, there’s a lot of hesitation around investing in Agentforce, especially when companies are already using Claude or other large language models where they understand exactly what they’re spending.
Customers want to know what the return on investment will look like beyond the pilot. Kayt: That’s part of why readiness has become such a big conversation.
Salesforce account executives are pushing Agentforce so heavily that organizations almost don’t have the option to ignore it anymore.
As renewals come up, companies know AI is going to be part of the discussion. That’s forcing them to ask, “How do we get organized so we can actually take advantage of this?”
Interviewer: What surprised you the most about developments at Salesforce and the broader evolution of RevOps and PSA?
Kayt: For me, it’s how quickly the conversation has changed.
As I said before, people know AI is here. They know they need to adopt it. The question isn’t if anymore—it’s how. But most organizations still have foundational work to do before they can really take advantage of it.
There’s still a lot of work around process, governance, and getting the business aligned. Those things aren’t as exciting as talking about AI, but they’re what make AI successful.
Raul: I agree. Everything is moving incredibly fast.
Customers are asking us where they should start. They’re asking how to build an AI roadmap, what use cases add the most value, and where the low-hanging fruit is.
The best opportunities are often the more deterministic use cases—projects that deliver value quickly and help fund the larger AI initiatives that require more investment and strategic planning.
What normally would have taken years to evolve is happening over a matter of months. I think we’ll see another major shift before the end of the year.
I’m also excited about what this means for professional services. We talk a lot about hybrid project delivery, where agents handle more of the execution while humans lead strategy, relationships, and decision-making. It’ll be interesting to see what that balance looks like.
Interviewer: Between the two of you, you’ve now attended Agentforce World Tours in New York, Boston, and London, to name a few. Did you notice differences in the messaging, priorities, or conversations depending on the city and country?
Raul: Definitely. The public sector was front and center in London. There were a lot more workshops and conversations around heavily regulated industries, with more customer use cases.
The keynote, led by Salesforce’s UKI President & CEO, Zahra Bahrololoumi, focused less on efficiency and more on growth—how agentic AI is reshaping how companies operate, innovate, and serve customers. It was about driving a different customer experience that goes beyond efficiency and focuses on growth metrics.
Governance was also a much bigger topic than what I’d heard at the U.S. World Tours, which makes sense given GDPR and the heavier regulatory environment. There was still a big focus on Headless 360, but the overall framing centered on growth and governance.
It was also a much larger event than some of the U.S. World Tours, and that changed the feel of the conversations.
Interviewer: Kayt, you attended both New York City and Boston. Did you notice a similar shift?
Kayt: Yeah. Those events were eight weeks apart, so the world had changed a little in two months’ time.
New York City was talking much more about the agentic enterprise and making sure businesses understood where Salesforce was trying to take things.
Boston seemed much more focused on how to get your organization ready to become an agentic enterprise. It felt smaller because it was a smaller audience, and it felt like they were really trying to help people understand what they needed to do to get there.
Interviewer: After attending so many World Tour events, what important topics do you think aren’t getting enough attention?
Kayt: Raul and I are probably going to have very different opinions because of what each of us does.
From what I’m seeing, people know AI is here. They know they need to adopt it. They know they need to get their organizations in order. But nobody is taking into consideration that there’s a cost to doing that.
Nobody is talking about budgeting to get their RevOps house in order and make sure everything is where it needs to be. Everybody thinks it’s just going to happen during an implementation, but it’s something that takes weeks and months to put together and change.
You can put a new process on paper, but then you have to put it into practice. Organizations know they need to make a change, but they’re not talking about how to make that change or what it’s going to mean for their day-to-day business.
That’s what I’ve been seeing.
Raul: That’s a good one.
For me, I keep coming back to cost modeling.
You go to these World Tour events and you can get a demo of Agentforce, but you don’t see a lot of discussion around credit consumption forecasting.
If you’re going to build a business case for AI, especially in a larger organization, you need to understand what it’s going to cost. We’re already seeing large enterprises—even technology companies—scale back on letting teams consume AI credits without understanding the impact.
At Klient, we’re working on models to help customers estimate what that cost could look like in year one. We encourage customers to take a crawl, walk, run approach so they understand their consumption before they scale across the organization.
The other thing that isn’t getting enough attention—and this is where Kayt and 28 North come in beautifully—is change management.
The demos make everything look easy. They make AI look almost plug-and-play. But there’s so much more involved. You need people, process, and a solid data foundation.
Sometimes the best place to start is with deterministic use cases that create efficiencies. But too many organizations assume AI alone is going to solve their problems or drive growth.
At the end of the day, people are still involved.
Whether we’re talking about hybrid project delivery or any other AI initiative, humans still need to enable the data, access the data, and make sure the business processes are right.
It doesn’t matter what AI tools you bring to the table. If you don’t have solid business processes, your outcomes are going to be the same.
For me, change management and cost modeling are the two biggest topics I expect to see more of at these World Tours, and I don’t see enough of either.
Interviewer: Looking ahead a year, what do you expect the biggest themes and conversations will be at the next Agentforce World Tour?
Kayt: A year is a long time.
I don’t know if it’ll take a year to get there, but Salesforce is spending a lot of time answering the question of how organizations get started. Right now they’re encouraging customers to begin on the service side because almost every company has some kind of customer service organization.
Let an agent answer simple questions. Let an agent take a case and route it to the right person. I think that’s going to expand beyond service into areas like sales and marketing.
Raul: I’m going to be very curious to see this year’s Dreamforce.
I think we’re past the point of pie-in-the-sky visions and beautiful demos. Customers are looking for bottom-line results.
Show us the proof.
We’ve been talking about the agentic enterprise for a couple of years. Now show us customers getting outcomes at scale. Show us growth. Show us revenue.
I think the conversation is shifting from demos to the P&L. Show us the numbers. Show us the metrics. Show us adoption.
I think there’s some fatigue around the “art of the possible.” People want to know how other customers are driving business outcomes and improving their KPIs.
If Salesforce is going to continue building momentum next year, that’s what I think we’ll see more of.
After five Agentforce World Tour events, Raul Acosta and Kayt Parrella have seen the conversation around AI evolve from possibility to practicality. Both agreed that the organizations making the most progress aren’t necessarily the ones adopting AI the fastest. They’re the ones investing in the fundamentals first.
For Kayt, that means improving processes, cleaning up data, and helping organizations prepare for change.
For Raul, it means pairing AI enthusiasm with governance, cost planning, and realistic roadmaps that deliver measurable value over time.
As Salesforce’s AI strategy continues to evolve, it’s clear that the conversation is no longer about whether organizations should adopt AI—it’s about how to do it thoughtfully and effectively.
If you’d like to continue the conversation about AI, Agentforce, or the future of the Salesforce ecosystem, connect with the teams at Klient and 28 North Consulting. They’d love to hear what you’re seeing in your own organization.