TLDR: Anything an operator can do in Symbol, an agent can now do too. We opened the product — API, CLI, and a skill that teaches the agent how to use them — so the recurring work of a security awareness program does not live in the console. Start at Symbol for agents.
Monday used to start in the console.
Open a child company. Filter the overdue list. Export a CSV. Switch tenants. Do it again. By the time the book of business is walked, the interesting question — what was supposed to run this month, and what actually went out — is still unanswered, and the person who should be answering it is tired of tables.
That is not a training problem. It is an access problem. The work was already in the product. The operator just could not hand it to anything that could do the walking.
So we opened the door.
Agents are how this work gets done now
We have spent years making security awareness something an operator can run for dozens of clients without standing up a new team for each one. The console got better. The dashboard got wider. None of that changed the shape of Monday.
Agents did.
Not a chatbot pasted into the corner of the admin UI. An agent that can act. Look through the book of business. Summarize what is overdue. Assign the training. Collect the policy signatures. Stand up the simulation. Triage the threat. Then come back with the number a quarterly review can actually use.
Anything you can do in Symbol, an agent can now do too.
That is the launch. Symbol for agents is the page. CLI if you want the official command line. A skill if you want Claude, Codex, OpenCode, Cursor, Grok, or whatever harness you already run to be taught how Symbol wants to be used. The Admin API if you would rather build the harness yourself. Same product. Same permissions a person would have. A faster way in.
What “do too” actually means
This is not a slogan. These are the jobs operators already staff.
- Provision users and child companies without a spreadsheet ritual
- Assign training and see who is still overdue
- Collect policy signatures
- Run phishing simulations
- Triage cyber threats
- Operate the MSP book: companies, programs, what was scheduled, what was sent
An agent that can do those things is not a demo. It is a second pair of hands on the work that used to scale linearly with client count.
The old way was possible. The API has been there. A careful integration, a script, a person who knew which endpoint to call. Almost nobody ran their program that way, because building that plumbing was a project, and the console was right there. Agents collapse that project. You do not staff an integration team to ask “who is overdue across these twelve companies.” You ask the agent, and it uses Symbol the way a sharp operator would.
Why we built it this way
The mission is to accelerate operators. Not to add another screen they have to hire for.
The last few posts are that argument, written one job at a time. They are not a reading list we attached at the end. They are the reason this exists.
Why APIs matter when you are running dozens of programs: a multi-tenant dashboard lets you see the clients. It does not scale the work. The API is the operating system. Symbol for agents is how that operating system shows up in the tool the operator already has open.
Automating client onboarding is the highest-leverage hour we know. The signed statement of work should not become a day of clicking before the program is live. An agent that can provision, assign, and report is how that hour shrinks.
Repeatable playbooks are how a small team runs the same program the same way for every client. An agent without a playbook is a fast intern. An agent pointed at a playbook is how the service stays the service when the book grows.
Packaging security awareness as a service only holds if delivery is repeatable enough to price. Acceleration is a margin decision. If every new client still costs an analyst’s Monday, the tier you sold is a story.
Planned versus actuals is the question that ends the quarterly review or saves it. A calendar is a promise. Actuals are evidence. An agent that can ask both, across tenants, is how an operator walks into that room with an answer instead of a dashboard screenshot.
Read them as one line. Operators win when provision, assign, simulate, report, and prove-what-ran do not consume a person for every client. We built the API so that work could be automated. We built Symbol for agents so the operator does not have to become the automation.
The operator does not disappear
Accelerate does not mean the judgment leaves the building.
Which clients are in scope. What the quarterly review is allowed to claim. When a program should pause. When a completion rate is theater. Those calls stay with the person who owns the book. What goes away is earning that judgment one export at a time.
That is the split. The agent does the walking. The operator does the deciding. The posts above are the work we want off the analyst’s plate. Symbol for agents is where you point the agent you already use and start.
If you want a walkthrough on your book of business, the page will book one. If you already live in a coding agent, or you are about to, because that is where this is going — jump in. The puck is over there. We built the door to meet it.