What do users say?

I really appreciate the AI in Agentforce Sales, which is driving the business faster and improving adaptability. It's nice how salespersons can access it throughout their sales UI, and with integration to MCP, it connects to chats like Teams and Copilot. This connection helps in driving leads and working efficiently with the sales CRM. I also like that it's much less complex for the user because you don't have to log in multiple times; everything can be done in one place. The AI connectivity and lack of unnecessary logins make processes like opportunity creation and approvals really smooth and efficient. Review collected by and hosted on G2.com.

I think Agentforce Sales could be improved by addressing the hallucination issues with the AI and adding more proactive features. It would also be better if the system was more flexible, allowing users to make changes and adjustments beyond standard agendas. Overall, I feel there is room for enhancement, and these changes would make it more user-friendly and adaptable. Review collected by and hosted on G2.com.

I really like the ease of use of Agentforce Sales, as the tool gathers all data into a single view, which greatly simplifies my tasks. The integration of functionalities is excellent, and I don't need advanced technical knowledge to use it. The ability to capture data from different objects and consolidate them into a single view is a great advantage, facilitating customer approach with quick and assertive information. The initial setup process was very simple due to the low need for customization, and this significantly optimized my work. Review collected by and hosted on G2.com.

I think some usage limitations can be improved, especially in the area of relationships. There are interruptions in capturing data from all objects related to the client, without limitations, which makes it difficult to cross all the information. Review collected by and hosted on G2.com.

What I like most about Agentforce Sales is how it brings customer data, sales processes, AI-powered insights, and automation together in one platform. It helps sales teams boost productivity by simplifying lead and opportunity management, offering real-time visibility into the sales pipeline, and supporting more personalized customer engagement. Its AI capabilities also help teams spot opportunities sooner, prioritize the right actions, automate repetitive tasks, and make data-driven decisions that strengthen sales performance and improve customer outcomes. Review collected by and hosted on G2.com.

One drawback of Agentforce Sales is the steep learning curve for new users, largely because of its broad feature set and many customization options. To get the full value from certain workflows, teams may need substantial configuration along with ongoing administrative support. Licensing costs and paid add-ons can also become expensive as an organization’s usage scales. In addition, some users may find the interface complex to navigate, and creating reports or dashboards can be challenging without specialized training. Review collected by and hosted on G2.com.

What I like best about Agentforce Sales is having a centralized, connected view of the customer and sales process. It brings accounts, opportunities, activities, reporting, and customer data together so teams can work from a single source of truth rather than jumping between systems.

I also really like the direction Salesforce is taking with AI. Agentforce can reduce the amount of manual administrative work—such as updating records, researching accounts, managing pipeline, and preparing follow-ups—while giving sellers more actionable insights. That allows teams to spend more time on relationship-building and higher-value work. (Salesforce)

The flexibility and ability to customize Salesforce around an organization’s specific processes is another major strength. It can evolve as business needs change while still maintaining a consistent data foundation. Review collected by and hosted on G2.com.

The biggest challenge with Agentforce Sales is that it can become overly complex. With so many configuration options, integrations, fields, automations, and customization capabilities, it can sometimes be difficult for users to know what information is truly important or what the preferred process should be.

The value of the AI capabilities also depends heavily on having clean, accurate, and well-maintained data. If the underlying data is inconsistent, the insights and recommendations are less useful. There can also be a learning curve for users when new features and AI capabilities are introduced, so strong governance, training, and change management are important. Review collected by and hosted on G2.com.

The human-in-the-loop design is what stands out. Agentforce doesn't act autonomously: it suggests, the rep reviews, and only then does anything get written. That makes client adoption much easier.

We built six assistive capabilities on a single Agent (booking identification, case classification, priority analysis, conversation summaries, 360 customer profiles, and AI-drafted replies), each with its own versioned prompt and guardrails. The Prompt + Flow + Apex stack is clean and flexible.

Built-in guardrails save credits by blocking calls when input data isn't eligible, and the Trust Layer gives real confidence in regulated industries. Plus, it's fully native to Salesforce: no middleware, no external APIs. Review collected by and hosted on G2.com.

The prompt engineering cycle is slow and opaque. We went through 6 prompt versions on a single use case (AF06) before getting consistent output. There's no built-in A/B testing or version comparison tooling, so iteration is manual and time-consuming.

Calibration with real data is a real gap. The platform works well with synthetic test data, but tuning for production accuracy (e.g., priority analysis, urgency detection) requires large real-world samples that Salesforce doesn't help you structure or benchmark.

Pricing complexity is also a concern. Between Flex Credits, per-conversation billing, and add-on tiers, it's hard to give clients a clear cost forecast, especially when guardrails block calls (good for savings, bad for predictability). Review collected by and hosted on G2.com.

Salesforce gives us a highly configurable platform that can support everything from core CRM and pipeline management to territory planning, sales engagement, analytics, automation, and AI. The biggest advantage is the ability to bring our sales data and workflows into one ecosystem and then tailor the experience to how our sellers and leaders actually work. We’ve been able to build seller-specific experiences, automate lead and opportunity processes, improve forecasting and reporting, and increasingly use AI to surface insights and recommended actions. The platform is extremely powerful when you have the right strategy and governance around it. Review collected by and hosted on G2.com.

The biggest downside is complexity. Salesforce is incredibly capable, but getting the full value often requires significant configuration, administration, data governance, and technical expertise. The product portfolio can also feel fragmented, with capabilities spread across different clouds, add-ons, licenses, and constantly evolving product names. It is easy for an organization to overbuild or create a poor user experience if there is not strong ownership of the platform. I would also like to see more advanced AI capabilities work seamlessly across the core CRM experience without requiring as much additional configuration and integration. Review collected by and hosted on G2.com.

the best thing about this product is entirely conditional on data gravity, not on the AI being good. The SDR agent's edge is that it already has your CRM history, your engagement signals, your deal stage — an SDR tool from Outreach or Clay has to integrate to get that; Agentforce starts with it. If your data is messy or your Salesforce instance is a graveyard of stale fields, this same "advantage" becomes the agent confidently acting on garbage. Review collected by and hosted on G2.com.

Scaling failure rate, implementation timeline, Hidden Data Cloud dependency Review collected by and hosted on G2.com.

Having our data in one centralized database allows us to function more dynamically across work streams. Automation and integrations with business tools provides efficiencies that allow the team more time to focus on selling and less time heads down in a system. Review collected by and hosted on G2.com.

Pricing can be a little on the high side - especially when you factor in the add-ons that enhance the system for deeper efficiency. The mobile app can feel limited when you compare it to the desktop version - tablet users tend to lean towards web browser versions for continuity purposes. Data Storage can be eaten into fairly quickly - admins need to factor in data management to make sure they keep things in check. Review collected by and hosted on G2.com.

I love that Agentforce Sales is evolving beyond a traditional CRM. Bringing customer data, sales, service, and AI together creates the opportunity to spend less time managing the system and more time building relationships and actually selling. Review collected by and hosted on G2.com.

The complexity. Agentforce Sales is incredibly powerful, but that power can come with too much configuration, too many products and add-ons, and an increasingly complicated licensing story. I’d love to see Salesforce make it easier for organizations to understand what they need, get value quickly, and adopt AI without requiring another layer of complexity. Review collected by and hosted on G2.com.

I work mainly as a Salesforce architect, so I see Sales Cloud from the design and implementation side.

I like that we can combine the sales process, automation, data, and now AI agents on the same platform. It gives us many options to design a better experience for sales users without building everything from scratch. Review collected by and hosted on G2.com.

Cost is still a concern, especially when we add more products and AI capabilities. It can also be difficult to explain the ROI before the customer has enough usage data.

From an architect point of view, AI does not remove the need for good data, security, and process design. In some cases, it makes those decisions even more important. Review collected by and hosted on G2.com.