AI for field service that works with your agents
Use the AI built into Forz, or connect the agent your team already uses: Claude, ChatGPT, Grok, Hermes Agent or your own. Either way, your jobs, schedule and invoices stay in one field service management platform.
Two ways to use AI in Forz
Forz gives contractors two ways to put AI to work. Start with the AI built into Forz if you want nothing to configure, or connect your own agent if your team already works with one. You can use both.
Forz integrated AI
AI that runs inside Forz, on your Forz data. Your office team uses it from the screens they already work in.
- Ranked schedule suggestions for unscheduled jobs
- Report filters written from a plain-language request
- Call transcripts and summaries for calls on Forz telephony (priced separately)
Bring your own agent
Connect the agent your team already works with and let it read and update your Forz records.
- Claude, ChatGPT, Grok, Hermes Agent or an agent you built
- Connects through the Forz CLI or REST API today
- Your AI provider bills you directly, with no Forz markup
Forz integrated AI: nothing to set up
If you want AI dispatch software without wiring up an agent, start here. It runs on the data already in Forz and shows up where your dispatchers and office staff already work.
Schedule suggestions
For an unscheduled job, Forz looks up to 14 days ahead and returns up to 5 ranked time slots and technicians, each with a score breakdown. The dispatcher still makes the call.
Scheduling & dispatch →Reports from a sentence
Describe the report you want, like “all overdue invoices over $500”, and Forz writes the filter. You can still write or adjust filters by hand.
Reporting & dashboards →Call summaries
Calls recorded through Forz telephony are transcribed and summarized automatically, and the summary is saved as a note on the matched contact, so nobody has to replay the recording. Telephony is priced separately.
Bring your own AI agent
If your team already works with an AI agent, connect it to Forz instead of learning a new one. Bring your own LLM as well: the connection doesn't depend on the model.
Claude
Anthropic's models, for example in Claude Code. The Forz CLI package includes a ready-made Claude Code skill.
ChatGPT
OpenAI's models, for example in Codex. The package includes an AGENTS.md file that Codex and other AGENTS.md-compatible agents read.
Grok
xAI's models, in any agent that can run command-line tools.
Hermes Agent
A free, open-source AI agent from Nous Research that works with any model provider.
Your own agent
Build on the Forz CLI, or have your agent call the Forz REST API directly.
How it connects today
Agents connect through the Forz CLI, a command-line tool published on npm as forz-cli, or through the Forz REST API it runs on. Through the CLI, an agent can list, look up, create, update and delete customers, sites, contacts, jobs, estimates, invoices, sales orders, items, tasks, leads, deals and projects, and read lookups such as tax rates, payment terms, job types and statuses.
A Forz server for the Model Context Protocol (MCP) is on the public roadmap as planned. It isn't available yet.
Things you could ask your agent
- Look up a customer's sites, contacts and open jobs before you call them back.
- Create a job from the details in a customer's email.
- Pull this month's invoices into a spreadsheet for your bookkeeper.
What you pay, in plain terms
Your seats are the only required cost. Schedule suggestions and report filters come with them, your agent's model usage is billed by your AI provider, and any usage-based Forz AI is optional.
Your Forz seats
$50/user/month, billed annually
Or $60/user/month month-to-month. Both have a 10-user minimum, and the core plan comes with a 3-year rate guarantee.
Your own agent
No Forz markup
Forz doesn't mark up or meter your AI provider's usage. Your provider bills you directly, under your own account.
Forz integrated AI
Built in
Schedule suggestions and report filters come with your plan. Call summaries work on calls recorded through Forz telephony, which is priced separately. Any usage-based AI is optional, and you see its price before you turn it on.
Connect your agent
- Check with our team that API access is turned on for your Forz account.
- Create an API key in Forz settings.
- Install the Forz CLI from npm and save your key with
forz login. - Point your agent at the Claude Code skill or the AGENTS.md file that comes with the package.
# install the CLI
npm install -g forz-cli
# save the API key you created in Forz
forz login --token <your-api-key>
# check which account and user the key belongs to
forz whoamiModel-agnostic, with you in control
Forz doesn't tie you to one AI vendor. Switch models whenever you like, and the CLI, your API key and your Forz data stay the same.
It works through an API key you control
A user creates a Forz API key in Forz settings and gives it to the agent. Each key belongs to one user in one company account, and the command forz whoami shows which. You decide which key an agent gets. The CLI saves the key on the machine the agent runs on, in a file only that user account can read.
Changes are live
There's no sandbox or test mode. Whatever an agent creates, updates or deletes changes your real Forz data, so let it look things up before you let it change records.
Guardrails on edits and invoices
Updates and deletes need the record's current version, so an agent can't overwrite a change it hasn't seen. When retrying an invoice or sales-order creation, reuse the same request key to avoid creating a duplicate.
Your data goes to the provider you chose
When your agent reads Forz records, it sends that data to your AI provider under your account and their terms. Choose a provider whose data terms work for your business.
Frequently asked questions
No. Forz AI agents are not a phone receptionist or CSR product, and they don't answer your phones. They work on the records behind your business: customers, jobs, the schedule, estimates and invoices. Forz does transcribe and summarize recorded calls, but a person still takes the call.
Forz doesn't mark up or meter your AI provider's usage. When your agent uses Claude, ChatGPT, Grok or another model, that provider bills you directly under your own account, and Forz isn't part of that bill. Your Forz plan is priced per user: $50/user/month billed annually or $60/user/month month-to-month, with a 10-user minimum.
Schedule suggestions and report filters are built into Forz and come with your per-user plan. Call summaries work on calls recorded through Forz telephony, which is priced separately. Any usage-based AI is optional and priced separately, and you'll see its price before you turn it on. Nothing is added to your bill unless you switch it on. See Forz pricing.
Any agent that can run command-line tools can use the Forz CLI, whichever model it runs on: Claude, ChatGPT, Grok, Hermes Agent (an open-source agent from Nous Research) or one your team built. A custom agent can also call the Forz REST API directly. The CLI package includes a Claude Code skill and an AGENTS.md file for OpenAI Codex and other agents that read AGENTS.md.
Not yet. A Forz MCP server is on the public roadmap as planned. Until it ships, agents connect through the Forz CLI or the Forz REST API.
It works through a Forz API key that belongs to one user in one company account, and you decide which key it gets. Through the CLI it can list, create, update and delete customers, sites, contacts, jobs, estimates, invoices, sales orders, items, tasks, leads, deals and projects. There's no test mode, so its changes go straight to your live Forz data.
See Forz with your AI setup
Book a demo and we'll show you Forz integrated AI and how the agent your team already uses connects to Forz.