Entering 2025, the Salesforce ecosystem is in a crucial position of working out whether Agentforce is the wildly powerful tool it claims to be or just very well-marketed hype.
Experimenting with different use cases, upskilling in AI, and exploring value realization with clients will all be key factors in determining Agentforce’s worth. It’s only in the early days, but the very first seeds have been firmly planted. Here’s what Salesforce professionals have to say about Agentforce so far.
A+ Marketing or Something More?
Since Agentforce was first unveiled at Dreamforce 2024, Salesforce has been on a steadfast mission to give it the strongest promotional run possible. Fortunately, this has been a great success; the Agentforce hype is very much real, and professionals seem to be excited about what is yet to come from this shiny new platform.
Brian Shea, a Salesforce Architect and Independent Consultant, said that Agentforce has already “exceeded expectations” for him.
Brian, who is working on client projects using Agentforce already, admitted that he was skeptical at first, but now feels very positive about the platform.
“I think the big picture is that Agentforce is an effective enterprise AI offering and it is impressive how far the product has come since Dreamforce,” he said. “More importantly, based on the attention and resources that Salesforce is putting behind Agentforce, I expect the product will continue to improve at a similar pace in 2025.”
Deepanshu Soni, another Salesforce Architect, shares Brian’s sentiments.
“I’ve been using the Agentforce for Developers plugin; it’s really helpful and fast-tracks the build time,” he said. “I’ve also configured a couple of agents, but I need to get more hands-on to get optimized results, and I feel like this is where pre-built agents will come into place soon.”
For those in the early stages of adoption and implementation, it’s clear that there is a lot of curiosity. Phillip Poynton, a Salesforce recruiter, says that he thinks that Agentforce has been a long time coming and that we’re now finally seeing it “come to life.”
However, he mentioned that the $2 per conversation pricing model remains confusing, which has been a common complaint among professionals in the ecosystem ever since the model was announced. Rachel Shumway, a Salesforce Design Lead, shared her own confusion, asking whether the $2 per conversation cost was the “price of secure data” or just because it was something that Salesforce could charge.
It’s also something that Salesforce has been trying to make clearer (with mixed success), with many hoping more answers will be provided in the new year.
“Cautious Optimism Is the Short Version”
Despite Agentforce’s hype, certain professionals still remain skeptical. Anand B Narasimhan, the CTO of S-Docs, said that he was approaching the situation with “cautious optimism.”
He claims that there will need to be a reality check to distinguish between “experimentation” and “production-ready”, and that he struggles to see how Agentforce’s pricing model will generate good ROI.
Pietro Piga, a Salesforce Business Architect and AI enthusiast says that the famous Spiderman quote, in this instance, needs to be adjusted from “with great power comes great responsibility” to “with great promises comes great responsibility.”
“Salesforce’s campaign around Agentforce emphasizes ‘easy to set up’ rather than ‘it can significantly enhance employee performance’ – there’s a critical distinction.”
This is definitely something to note, as even on the main Salesforce Agentforce page, you have to scroll at least halfway down before you see any real mention of how Agentforce will work harmoniously with humans in the “Humans + AI” section.
Additionally, although a lot of the marketing messaging is focused on easy setup, this still isn’t enough to convince some professionals.
Joseph Monroe, a Senior Salesforce Consultant, says that he’ll “remain pessimistic” until he sees an actual implementation demonstrated.
“I’d like to see something custom and follow someone’s journey through that – not demo orgs or prebuilt agents that have most of the hard work done in a narrow focus,” he said.
This is why it’s crucial to share experiences and learn from others in the ecosystem when it comes to Agentforce in this experimentation phase. Fortunately, this has already been observed by professionals in the ecosystem, so the continuation of this will hopefully help diminish doubts and inspire others.
What Are the Concerns?
Concerns about Agentforce can be split into four main key areas: too much hype, not recognizing appropriate use cases, a lack of understanding of the amount of preparation needed, and a prominent AI skills gap.
Too Much Hype
Both Pietro and Kendra Von Achen, a Salesforce Consulting Director, said that the current excitement could be clouding judgments.
“My prediction is that Agentforce will generate a lot of initial excitement but will burn out quickly for many,” Pietro said. “I believe that 90% of stakeholders – including both customers and consultancies – are unprepared.”
Kendra shares similar thoughts, warning against the action of just jumping “on the next shiny new object”.
“Things like this should be thought out and planned. Clients constantly say ‘I want AI’, but what does that actually mean to them? What exactly do they want it to do? Have they planned for the infrastructure to support the solution once it’s implemented and live?”
With the sheer scale of promotion efforts dedicated to Agentforce, it can be easy to be swayed into trying something new and developing exciting new strategies.
But in order to avoid issues like technical debt, underdeveloped agents, and confused users/customers, then these processes need to be thoroughly thought out.
No Use Cases
Agentforce is only as valuable as its use cases – for organizations who struggle to find appropriate use cases, the tool may seem like an unnecessary expense in terms of both time and money.
Danny Gelfenbaum, a Salesforce Application Architect, says that he feels it’s too early to adopt the tool fully. Timo Kovala, a Salesforce Lead Architect, agrees, saying he’s noticed a hesitation to adopt too:
“Looking at conversations thus far, enterprise-level customers are just coming to grips with generative AI, and now they’re supposed to pivot quickly to agentic. It’s simply not realistic to assume so,” he said.
“When I spar with ops leaders and C-level execs, they’re having a hard time identifying agentic use cases that go beyond the ‘vanilla’ copilot. The pace of development is so fast that businesses can’t keep up.”
Timo also reflected on the importance of not getting ahead of one’s self when it comes to Agentforce – especially if the use cases aren’t there right now.
“We will see a staggered start for Agentforce. Some are getting ready to sprint, while others are tying their running shoes. Some aren’t even at the right stadium.”
Don’t Underestimate the Preparation
Like with any new tool or platform – especially one that relies on your data to perform actions – there is a considerable amount of preparation that needs to take place before you get going with it. Just how in-depth Agentforce’s preparation process is is something that some professionals feel isn’t talked about enough, which can be concerning.
Jodi Hrbek, a Salesforce Coach, says that this lack of conversation means that important issues are left uncovered.
“What makes me nervous is that there is much discussion of how fast you can deploy them, but little mention of the shape your org needs to be in, the importance of having a well-thought-out business process, or the risks – including what an admin should propose to mitigate or monitor.”
Rachel agrees with this, saying that selling Agentforce shouldn’t be a problem, but then the issues creep in from there.
“Training folks how to build it well proves challenging with constant changes to playground capabilities, lack of documentation, and a concise way to identify quality vs. suboptimal use cases.”
Pietro also thinks that there isn’t enough emphasis put on how critical good data quality is for working with Agentforce, which is likely to affect the success of Agentforce implementations.
“Unfortunately, the lack of focus on data quality means many will misdiagnose issues, attributing poor results to Agentforce’s limitations rather than the underlying data,” he said.
“This misunderstanding could negatively impact Salesforce’s reputation, as stakeholders may perceive Agentforce as overly expensive or ineffective when in reality, the root cause lies in their own data shortcomings.”
Tackling the AI Skills Gap
Lastly, it will be extremely difficult to get any Agentforce effort off the ground if professionals aren’t equipped with the necessary skills to get it running effectively.
Anand says that Trailhead is only one small solution here, but that real-life experiences are “few and far between.”
“I still consider prompts, the whole Studio builder experience is quite complex for your average Salesforce administrator,” he said.
The AI skills gap is something that presents significant challenges for everyone in the ecosystem – from juniors all the way to seniors right at the top of their industry.
There will be rising pressure to quickly upskill in AI technologies like machine learning and predictive analytics to stay competitive. Those unable to adapt risk job insecurity and increased competition from AI specialists entering the ecosystem.
Additionally, collaboration gaps may emerge between traditional Salesforce roles and AI experts, while a lack of AI literacy could lead to underutilization or misuse of Salesforce’s AI tools.
Brian says that he believes that Agentforce’s biggest barriers to growth will be surrounding human constraints, and with the AI skills gap positioned as a very real threat, it’s easy to see the logic in this hypothesis.
“At least for 2025, the skills gap may be the biggest limiting factor for Agentforce growth/adoption.”
Final Thoughts: What Can Salesforce Do Next?
It’s evident that Agentforce is a tool that a fair few professionals are already excited about, but the concept of a staggered start is one that seems the most likely to take place.
“There is no doubt that agents will be part of the business landscape,” Ian Gotts, the Founder and CEO of Elements.cloud said.
Xavery Lisinski, who works with AI every day, also shares this view.
“I think there are still a few things to iron out, and it will take some time for people to get familiar with the experience of building and operating agents, but I think 2025 is gonna be the year that Agents take off commercially,” he said.
If Salesforce wants to make that vision a reality, then they must work to make at least three key areas clearer: the range of available use cases, the amount of preparation needed to get an org Agentforce-ready (and what that entails), as well as how they aim to tackle the AI skills gap.
For the skills gap, work has already started with the free AI certifications and hands-on Agentforce sessions at events, but whether that will be enough will definitely be something we find out over the course of this year and next.