U.S.-Based
Salesforce Agentforce implementation for service and sales teams
We plan, configure and test Agentforce agents in your Salesforce org: instructions, actions, data access and the handoffs to your people.
$12,000
Starting price of an Agentforce implementation
40
AI projects delivered since 2020
5.0
Clutch rating from 40 reviews

How an Agentforce agent runs
We approach AI and machine learning as engineering disciplines, not experiments. Our process is designed to move from clear business goals to production-ready models that deliver measurable results.
A customer asks about an order
The question arrives in web chat or Slack.
Agentforce looks it up
An action reads the order and shipping status from Salesforce and your ERP.
Hard cases go to a rep
Refunds above your limit hand off with the full transcript.
The case is closed
The answer and next steps are logged on the case record.
Agentforce implementation, in short
- Kansas City metro software and AI company that implements Salesforce Agentforce
- Agentforce agents, actions and data access for service and sales teams on Salesforce
- Implementation from $12,000, Agentforce licenses billed by Salesforce
Agents in your org, tested first.
- 01ImplementationFrom $12,000: agent design, actions, data access, testing and launch
- 02LicensesBilled by Salesforce, directly to you
- 03Claude in Salesforce?Connecting Claude to your Salesforce data is a separate integration
- 04After launchMaintenance from $500 a month
What is Agentforce, and how does it differ from a chatbot?
Agentforce is Salesforce's AI agent platform, which Salesforce says delivers 24/7 autonomous support at enterprise scale. Unlike a rigid chatbot, an Agentforce agent answers questions and takes actions on your records, within guardrails you set.
How an Agentforce agent works
Atlas Reasoning Engine
The Atlas Reasoning Engine breaks a request into smaller tasks.
Agent Builder and Agent Script
Admins define subagents, instructions and actions in Agent Builder, and Agent Script gives builders tighter control.
Actions
Actions come from your existing workflows, prompt templates, Apex and APIs.
Systems outside Salesforce
MuleSoft API connectors and MCP support reach systems outside Salesforce.
Agentforce Observability
Observability monitors and tunes agent performance in near real time.
What an Agentforce implementation covers
Salesforce supplies the platform and the licenses. An implementation is the work around them: choosing the first job, writing instructions, building actions, connecting data, testing on past cases and launching in stages.
Types of Agentforce agents we implement
Salesforce offers purpose-built agents for sales, service, HR and more. These are four common starting points, what each handles and what it touches. Each agent starts with one job and a handoff rule; implementation starts at $12,000.
What Agentforce connects to
When Agentforce implementation pays off
Agentforce pays off when Salesforce is your system of record and a job repeats at volume, such as order questions, case updates or lead follow-up. The data the agent needs should already sit in Salesforce or behind an API.
Signs we see before a Salesforce team asks for an implementation:
- Service reps answer the same order, billing or account questions all day.
- Cases wait in a queue while reps look up data in two or three systems.
- Sales reps leave leads unworked because follow-up takes too long.
- Your Salesforce data is clean enough that a person would trust it.
- Leadership has bought Agentforce licenses, and nobody owns the rollout.
When to wait
If most of the work happens outside Salesforce, a custom AI agent is usually a better fit. If your records are incomplete or permissions are messy, fix them first; an AI readiness assessment shows where to start.
What to prepare before the first call
The Salesforce clouds and editions you run and who administers them, the first job you want an agent to take on with its current volume, about 50 real past cases or conversations with the right outcome for each, and handoff rules: what the agent may do alone, and what goes to a person.
Salesforce Agentforce consulting: what we set up
Five parts of every implementation, configured in your org and documented for your admins.
Agent design and instructions
We define each agent's job, subagents, actions and instructions in Agent Builder, with Agent Script where you need tight control. For example, a service agent whose instructions cover returns under 30 days, and hand anything else to a rep.
Actions into your systems
Actions read and update Salesforce records, and MuleSoft API connectors reach systems outside Salesforce. For example, an action that checks shipping status in your ERP through a MuleSoft connector and writes it to the case.
Data and grounding
Agents answer from your records and knowledge, with outside data brought in through Data 360. For example, product manuals and warranty terms used as grounding, so answers match what your reps would say.
Voice, Slack and web chat
We connect agents to the channels your customers and staff use, including Agentforce Voice and Slack. For example, the same order-status agent answering in web chat, in Slack for staff and on the phone.
Testing and monitoring
We test every action and handoff on real past cases, then track results with Agentforce Observability. For example, real past cases run against the agent in a sandbox, with every wrong answer fixed before launch.
What you get when the agent goes live
You get a working agent in your Salesforce org and everything your admins need to run it: instructions, actions, test results and documentation.
Agentforce ROI: a worked example
An Agentforce agent pays back when the rep hours it saves cost more than the implementation and the Salesforce usage. This example uses assumed numbers, not a client result: at these numbers, the implementation pays back in about two and a half months, before Agentforce usage, which Salesforce bills by action, conversation or user. Run your own numbers in the AI ROI calculator. After launch we track resolutions, handoffs and usage, so you can compare the real saving with this estimate.
How our Agentforce consultants run the rollout
Four steps from a readiness review to agents your team trusts.
We check your data, permissions and processes, and pick the first job for an agent. You get a scope, risks and a fixed price.
We write instructions, actions and handoff rules with your service and sales leads. You get an agent design your admins approve.
We configure the agent in a sandbox and test it on real past cases. You get test results on your own cases.
We go live in stages, watch results and tune instructions and actions. You get a live agent and a runbook.
AI work in production
AI builds from our case studies, and the record behind them.
Risks in Agentforce implementation, and how we handle them
Each risk gets a control in your org before launch, not a promise in a slide.
How Agentforce keeps data safe
Trust Layer, plus our rules
Salesforce says the Einstein Trust Layer is a set of features and guardrails that protect the privacy and security of your data. We add your own permissions, handoff rules and testing on top: each action reaches only the records it needs, and every change outside Salesforce can wait for a person to approve it. For company-wide AI rules, see AI governance and compliance.
Trust Layer features
Salesforce lists dynamic grounding, zero data retention and toxicity detection among the Trust Layer's features.
Audit Trail
The Audit Trail feature tracks agent actions and outputs.
Agentforce Guardrails
Guardrails combine safeguards you define with protections Salesforce manages, and can block off-topic or inappropriate conversations.
Testing tools
Agentforce's testing tools generate test cases and show how guardrails are followed before launch.
How Agentforce is priced, and what to budget
Licenses billed by Salesforce
Salesforce bills Agentforce directly, with several models: Flex Credits per action, pay per conversation, or per-user licenses and add-ons. Implementation is a separate cost: ours starts at $12,000 for a scoped first agent.
Buying models include pre-purchase, pre-commit and pay as you go. Salesforce's pricing examples exclude Data 360 credits and other consumption services. Source: salesforce.com/agentforce/pricing, read November 2026.
In the readiness review we estimate actions and conversations from your real volume, so you can compare license options before you sign.
Salesforce options
Flex Credits: per action. Salesforce lists an Agentforce action at 20 Flex Credits and a voice action at 30
Conversations: per conversation with the agent
Licenses and add-ons: Agentforce user licenses, add-ons for Sales, Service and Field Service, and editions
Salesforce Foundations: a way to start with Agentforce for free, which Salesforce offers for any use case

Questions to ask an Agentforce partner before you sign
These questions separate a tested rollout from a demo. A strong partner answers each one in writing.
Questions to ask, and what a good answer includes
- Which job should our first agent take on? One high-volume job with clear rules and a measurable result.
- How will you test actions before go-live? A sandbox run on your real past cases, including handoffs and edge cases.
- What happens when the agent is unsure? A handoff rule that passes the case and transcript to a person.
- What will Salesforce usage cost? An estimate of actions and conversations from your real volume.
- Who owns the configuration after launch? Your team, with full admin access and documentation in your org.
- Do we need data from outside Salesforce? A plan for MuleSoft connectors, MCP or custom APIs, limited to what the agent needs.
- What paperwork comes first? An NDA and a certificate of insurance before the first call.
What an Agentforce implementation costs
A fixed price for the implementation, agreed before work starts. Salesforce prices Agentforce with Flex Credits, Resolutions or per-user licenses, billed to you directly. See AI agent development cost.
Salesforce Agentforce implementation
A scoped first agent
From $12,000
Custom integration or MCP server
For a system outside Salesforce
From $4,000
Maintenance and support
After launch
From $500 a month
Project Discovery
Credited to the build
$490
What to look for in an Agentforce implementation partner
Six checks we suggest for any partner, including us.

- 01
AI agents in production
Ask for agents running today, not slides. We have delivered 40 AI projects since 2020.
- 02
Vendor documents first
Our NDA and certificate of insurance reach your procurement team before the first call.
- 03
You own the setup
Agents, actions, flows and documentation stay in your org, with full admin access.
- 04
Verified client reviews
Ask for reviews a third party checked. SoftDoes holds 5.0 on Clutch from 40 client reviews.
- 05
Depth beyond Salesforce
Agents often need outside data and models. We are a Claude Partner Network member and an OpenAI Registered Partner.
- 06
A team you can reach
Headquartered in Mission, KS, in the Kansas City metro. Call 800-909-2345 to reach the team.
Agentforce by industry
The same agent patterns, set up around each industry's records and rules. Each industry page covers use cases and compliance in detail.
Financial services
Account questions, document requests and case updates, with an audit trail for every agent action.
Insurance
Claim status, policy questions and first notice of loss intake, with a person approving payouts.
Healthcare
Appointment and billing questions routed to staff, with patient data rules set first.
Real estate
Lead follow-up, showing requests and tenant questions answered from your records.
Manufacturing
Order status, warranty claims and dealer questions answered from Salesforce and your ERP.
Logistics
Shipment status and delivery exceptions answered, with hard cases handed to a rep.
Questions about Salesforce Agentforce
Scope, pricing, partners and fit. Can't find your answer?
What is Salesforce Agentforce?
Agentforce is Salesforce's AI agent platform, which Salesforce describes as delivering 24/7 autonomous support at enterprise scale. Agents run on the Atlas Reasoning Engine and use actions to read and update records. They can reach other systems through MuleSoft API connectors. Teams configure them in Agent Builder and monitor them with Agentforce Observability.
What does an Agentforce implementation include?
Our implementation covers a readiness review, agent design, actions into Salesforce and other systems, data access, testing on your real past cases and a staged launch. We document every agent, action and handoff rule for your admins. Implementation starts at $12,000, and maintenance starts at $500 a month after launch.
How is Agentforce priced?
Salesforce sells Agentforce with Flex Credits, Resolutions or per-user licensing, and bills you directly. Implementation work is separate: ours starts at $12,000 for a scoped first agent. We estimate your likely usage during the readiness review, so you can compare license options before you sign anything.
What should we look for in an Agentforce partner?
Ask to see AI agents running in production. Check who owns the configuration after launch, how the partner tests actions before go-live and what happens when an agent is unsure. Ask for a fixed price in writing, and for an NDA and certificate of insurance before you share data.
Can Agentforce connect to systems outside Salesforce?
Yes. Salesforce offers MuleSoft API connectors to connect agents to other systems, and Agentforce supports MCP to reach new tools and resources. We build the actions and connections, then limit each one to the records and steps the agent needs. Changes outside Salesforce can wait for a person to approve them.
Should we use Agentforce or a custom AI agent?
Choose Agentforce when Salesforce is your system of record and most of the work happens on Salesforce data. Choose a custom AI agent when the job spans many systems outside Salesforce, or needs models and tools you control yourself. We scope both options in the readiness review.
Is Agentforce just a chatbot?
No. Salesforce says Agentforce replaces traditional chatbots with AI that handles a wide range of service issues without preprogrammed scenarios. An agent answers questions and also takes actions, such as updating a case or checking an order, within guardrails you set. Hard cases still go to a person with the full transcript.
Is Agentforce any good?
It works well when Salesforce is your system of record and the data an agent needs is clean and reachable. Results depend on the setup: clear instructions, tested actions, good knowledge articles and firm handoff rules. We test every agent on your real past cases before launch, so you see how it performs before customers do.
Do we need Data 360 for Agentforce?
It depends on where the agent's data lives. Agents answer from your Salesforce records and knowledge, and Salesforce uses Data 360 to bring in external data. Its pricing examples exclude Data 360 credits, so budget for them if the agent needs outside data. We check this in the readiness review, before you sign anything.
Can Agentforce work in Slack and on the phone?
Yes. Salesforce offers Agentforce Voice to bring AI voice to every channel, and agents can work in Slack and web chat. We connect the channels your customers and staff already use, with the same instructions, actions and handoff rules on each. Every channel is tested on real past conversations before launch.
How do you test an Agentforce agent before launch?
We configure the agent in a sandbox and run it on your real past cases, including the hard ones that should go to a person. Salesforce's testing tools generate test cases and show how guardrails are followed. We fix instructions and actions until results meet the pass mark agreed in the readiness review.
Will Agentforce replace our service reps?
That's not how we set it up. The agent takes the repeat questions and routine updates, and hands harder cases to a rep with the full transcript. Reps spend more of their time on the cases that need judgment. Every handoff rule is agreed with your service leads before launch.
Can you fix an Agentforce rollout that is not working?
Yes. We start with a review of instructions, actions, knowledge articles and handoff rules. Then we test the agent on your real past cases to find where it fails. Fixes go to a sandbox first and launch in stages. Maintenance starts at $500 a month if you want us to keep tuning it after that.
How long does an Agentforce implementation take?
It depends on the job, your data and your release process. We set the scope, price and schedule in a written plan before the build starts. Many teams begin with Project Discovery for $490, which is credited to the build. The plan also lists what we need from you, such as admin access and past cases.
Updated ยท Reviewed by Orest Andrusyshyn, Founder and CEO














































