Attending the Salesforce Agentforce World Tour this year offered a brilliant opportunity for me to reconnect with former colleagues and friends from previous delivery projects. As well as networking, the content gave me a definitive look at where Salesforce is heading over the next few years.
I was expecting potentially another round of unbridled AI hype, but the tone of the keynote came as a bit of a surprise. It was refreshingly muted and pragmatic. While the introductory video leaned heavily into a polished, futuristic vision, the on-stage narrative quickly shifted into something far more grounded and pragmatic.
I’m Mike Spencer, Senior Solution Architect at Desynit Limited, a Salesforce Partner based in Bristol. Working with medium to large enterprises, I know first-hand how businesses struggle to manage complex business workflows. The London Salesforce Agentforce World Tour was a great opportunity to understand the direction of Salesforce going forward and the latest tools and techniques to help with this struggle.
Salesforce didn’t shy away from the market’s current frustrations. In fact, a statistic flashed up on the big screen that likely stopped a few business leaders in their tracks.
To answer this value gap, Salesforce has consolidated its strategy into five structural pillars, pitched as the new blueprint for enterprise success:
🔹Context: Powered by Data 360.
🔹Work: Governed by the core CRM (Customer 360).
🔹Agency: Executed via Agentforce.
🔹Engagement: Centred in Slack.
🔹Insight: Delivered through Tableau.
It is a structurally sound framework, even if it feels a bit like a strategic repackaging of the existing ecosystem. The overarching theme was that you need to look holistically at your implementation and your data in order to unlock the potential of AI.
The case studies selected this year were highly relevant, particularly from a UK operational perspective. The standout was Thames Valley Police, which manages roughly 1.4 million contacts annually. Facing severe operational pressure, including an average 26-minute wait time on their non-emergency lines, they implemented Salesforce Service Cloud with Agentforce AI.
The metrics presented were notable: wait times were halved, and CSAT improved by 15%. They achieved this by introducing “Bobby,” an autonomous Agentforce assistant handling non-emergency queries, which now manages 75% of those digital interactions completely autonomously.
Crucially, they insisted that the AI was able to quickly route to a human, on any channel, if it detected that the conversation was actually of an emergency nature.
Other global brands like Canada Goose and Adecco also demonstrated solid use cases, alongside a surprise keynote appearance from the CEO of Formula 1 himself, Stefano Domenicali. This added some high-performance context to the customer engagement discussion.
For those focused on platform design, the most significant announcement wasn’t the out-of-the-box agents but the introduction of Headless 360.
Salesforce is now heavily pushing this as an architectural game-changer and the next big opportunity for businesses and partners. The platform is increasingly decoupling its core business logic and workflows from the standard browser-based UI and moving into Slack, AI tools like Claude and other interfaces.
Common use cases, such as answering automated service requests or intelligently routing complex legal workflows, can now bypass the traditional presentation layer entirely.
This is interesting because the standard Salesforce UI is one of the big selling points of the platform. The fact that by creating the data layer, the presentation layer is provided for free is a big reason why Salesforce implementations are fast and effective at providing businesses what they need. If users are not going to use the interface at all, it seems to me that this minimises the value of Salesforce.
Nonetheless, the long-term goal here is to position Salesforce as the ultimate system of record, governance, and security for both human users and external autonomous systems. For example, developers can now pull live Salesforce data into external LLMs such as Claude or ChatGPT using the Model Context Protocol (MCP) and seamlessly update records back in core Salesforce via Headless 360.
With a Headless 360 rollout expected later this year, Salesforce is clearly moving toward an API-first, agentic infrastructure, for now anyway.
For those of us who have spent years navigating the performance bottlenecks and complex code footprints of traditional Salesforce CPQ, the introduction of Agentforce Revenue Management (ARM) looks like a substantial upgrade.
ARM is built natively on the platform, moving away from legacy object constraints toward high-performance, attribute-based pricing procedures. It supports full commercial-to-technical decomposition and orchestration, meaning it is designed to handle thousands of quote lines without breaking a sweat. The entire architecture is driven by specialised catalogue, quoting, and billing agents, signalling a major evolution in how we will design quote-to-cash lifecycles moving forward.
I must admit, the absence of any Service Cloud Voice demonstrations or real-world implementation stories across the event was quite surprising to me.
Given how foundational contact centres are to the broader Service Cloud value proposition, the complete lack of focus on Voice implementations, such as Amazon Connect Unified Routing, felt like a bit of a missed opportunity.
There have been a number of announcements by Salesforce in this area, including Salesforce Voice or Agentforce Voice. It does leave me wondering whether Salesforce is still undecided how that specific product line is going to be promoted and are undergoing a quiet strategic realignment before they present to us the way forward.
Under the backdrop of a few specific successful implementations, Salesforce kept the broader industry hurdles front and centre.
For many, it was a stark reminder that deployment without strict architectural guardrails, data maturity, and clear business outcomes is a recipe for a failed pilot.
Projects signed off without a clear plan for Return on Investment (ROI) were always going to cause problems, and unfortunately, the hype and fear of missing out (FOMO) generated by AI put many implementations into this position.
Overall, the World Tour showed a Salesforce that is growing up alongside its AI strategy. The shift toward headless architecture and robust revenue tools gives architects a lot more concrete material to work with, provided we can steer our clients past those initial ROI hurdles.
If you like what you have read here, you may also like The Hidden Cost of Your Legacy Voice Setup: 5 Reasons to Move to Unified Routing Now
_______________________________________________________________
Mike Spencer | Salesforce Consultant | Desynit Limited
_______________________________________________________________
Our independent tech team has been servicing enterprise clients for over 15 years from our HQ in Bristol, UK. Let’s see how we can work together and get the most out of your Salesforce implementation.
