Working with the virtual technician in LogMeIn Resolve

The LogMeIn Resolve virtual technician automates manual tasks during remote control sessions by executing agent-approved solutions, called plans, by interpreting the remote device's screen and controlling the keyboard and mouse. Plans are written in natural language and can be either prepared in advance and stored in the knowledge base for easy re-use or generated during a support session.

The virtual technician helps automate certain tasks that are either common to an organization or time consuming for an agent. It can also provide training to agents by giving them a visual run-down that highlights the key actions of the support process.

How it works

The virtual technician assists agents in several ways:

Before a virtual technician-based support session, agents can document Standard Operating Procedures (SOPs) for common issues in a LogMeIn Resolve knowledge base article, outlining the steps of a virtual agent plan. Agents can also use the AI-generated session notes at the end of a support session to create an article with the necessary steps for the virtual technician to follow. This plan can then be executed directly from the knowledge base article. The knowledge base in LogMeIn Resolve also integrates with helpdesk ticketing, recommending related articles when viewing a ticket. This provides an easy way to invoke the virtual agent.
During a remote control session, agents can prompt the virtual technician for immediate actions, such as "Clear the cookies of the Edge browser". The virtual technician then creates and executes a step-by-step "plan". Alternatively, agents can use the virtual technician plan library to easily browse and execute virtual technician plans that are already stored in the knowledge base.

For tips on writing effective prompts, see Frequently asked questions about the LogMeIn Resolve virtual technician.

Examples of using the virtual technician

The virtual agent could be used to automate the execution of frequently performed tasks, such as the following:

Clear browser cookies and caches

Clearing browser data solves a range of application issues.

Example prompt: Clear the cookies of the Edge browser.

Run Standard Operating Procedures (SOPs) specific to end-users

Many agents run a set of standard procedures at the beginning of a support session to rule out common issues and get an overview of the system's health. These include installing updates, removing unnecessary files, or checking for specific system settings.

Example prompts:

Install the latest OS and driver updates
Free up space on drive C using the Disk Cleanup tool
Sync the computer clock to the default time server

Note: While you can use script-based Remote Execution for many maintenance tasks, the virtual technician can also automate tasks without using scripts.

Configure and fix company apps, such as office tools and online meeting software

Business communication apps typically produce high volumes of helpdesk tickets that involve standard configuration tasks to fix issues when vendors roll out product updates.

The virtual technician could be tasked with the following, using Microsoft Outlook as an example:

Install add-ins, such as Microsoft Teams, in Outlook
Repair the Outlook application from the Windows system settings
Fix common settings that cause Outlook to disconnect

Create a virtual technician plan in the LogMeIn Resolve knowledge base

You can add troubleshooting steps to a knowledge base article that the virtual technician will perform on a remote device.

A collection of virtual technician steps or instructions is called a plan in LogMeIn Resolve. These are the instructions that the virtual technician will try to follow. Each article can contain any number of plans, and each plan must have at least a goal that captures the plan’s intent and one or more steps to execute on a device. When you add a step that the virtual technician cannot follow during execution, it will try to derive the next logical step in the process based on the intent expressed by the plan and the current screen state of the device, or abort execution with an error.

For beginners

On the Knowledge base page, create a new article.
Name your article.
Select Insert plan on the toolbar.
Add a goal that best describes the task that you want the virtual agent to complete.
For example, if you want to check for available Windows updates on remote devices, type "Check for available Windows updates" as a goal.
Select the play icon on the right of the prompt field.
The AI generates the necessary steps to complete your task.
Review the steps and make changes if necessary.
Save your article.

For experienced agents

In several cases, you can find a solution for your problem online. When you do so, make sure you copy the steps required to resolve an issue.

Search for a resolution for your issue and copy the steps that you want the virtual technician to perform.
On the Knowledge base page, create a new article.
Name your article.
Paste the copied instructions into the article.
Highlight the instructions and select Insert plan on the toolbar to convert the steps into a virtual technician plan.
Enter a goal for the virtual technician plan that summarizes its intent.
Save your article.

Run a virtual technician plan during a support session

On the Devices page, select a device by its checkbox on the left.
The Manage selected device side panel opens on the right.
Select Virtual technician.
A remote support session starts with the Virtual Technician tab open in the side panel. If the remote session was started via Remote Control and you don't see the panel, select Show More on the toolbar and then select the Virtual Technician tab at the top of the side panel.
To create a new custom plan, type a prompt into the Describe what to do field on the right.
For example, to check for the latest available system updates on a Windows device, type "Check for available Windows updates".
Press Enter or select the play icon next to your prompt to let the virtual technician generate the necessary steps.
If you are not satisfied with the provided solution, rephrase your prompt and regenerate the steps.
Optional: Select the Execution Mode of running the task:
- Auto Clicking: The virtual technician performs the steps one by one and lets both the agent and the end-user watch as it proceeds. This is the default mode.
- Highlighting: The virtual technician only guides the agent to perform the necessary steps by highlighting the part of the screen where the agent has to take action. The virtual technician also provides a tooltip for each step with a short summary of what to do.
Optional: Select what actions should stop the execution of a running plan on the device. You can choose a combination of the following actions:
- Key Press: the end-user or the agent presses a key. Enabled by default.
- Mouse Click: the end-user or the agent clicks the mouse. Enabled by default.
- Mouse Move: the end-user or the agent moves the mouse. Disabled by default.
Optional: Select Enable Diagnostics to retain screen descriptions created by virtual technician for a limited period of time to help us diagnose problems that you may experience. During the open beta period, this is enabled by default.
Select Execute at the bottom.
You can monitor the virtual technician's progress on the right of the support session window. You can stop running the plan any time by selecting Cancel execution at the bottom of the side panel and rerun it later.
Tip: You can also run an existing virtual technician plan by selecting Browse plan library. Choose a plan from the list and then select Execute plan.

Run a virtual technician plan from a knowledge base article

On the Knowledge base page, select an article that is marked as having an executable plan.
- When viewing the articles in table layout, articles with executable plans are marked with a hand icon .
- When viewing the articles in card layout, articles with executable plans are marked with an Executable label.
After opening the article, select Execute on device next to the plan.
Select the remote device where you want the virtual technician to run the plan.
A remote support session starts and the steps of the plan are executed automatically.
Let the virtual technician finish its task.
You can track its progress on the right of the support session window. You can also stop and rerun the process any time.

Results: When the task runs, the virtual technician checks the steps that it has already performed so that you can track its progress. On the left of the support session window, you can watch what the AI is doing in real time.

Enable the virtual technician in GoTo Admin

For LogMeIn Resolve plans that include the virtual technician, an admin can configure the feature for each user who provides remote support for end-users. By default, all virtual technician features are enabled for all users.

Sign in to GoTo Admin.
From People > Users, select the user whose settings you want to change.
From the Settings > Virtual technician tab, limit access to virtual technician features as necessary:
- Virtual technician: Disable this setting to hide all virtual technician features from the user in LogMeIn Resolve. Existing virtual technician plans in knowledge base articles remain visible and can be edited. However, the user cannot insert new plans into articles or run existing plans.
- Execute stored plans: Disable this setting to prevent agents from executing virtual technician plans stored in the knowledge base. The Execute on device button in knowledge base articles will be disabled and the virtual plan library won't be accessible.
- Execute custom plans: Disable this setting to prevent agents from generating and executing custom plans based on prompts from the virtual technician side panel in remote control sessions.

Frequently asked questions about the LogMeIn Resolve virtual technician

Where to reach out with questions and issues?

If you have issues with the virtual agent, please contact your Customer Success Representative.

What is the virtual technician, and what can it do?

What are the license and system requirements?

During the open beta period of virtual technician, the following requirements apply:

The virtual technician feature is available to LogMeIn Resolve Premium or MSP license holders.
By default, the virtual technician is enabled for each user in GoTo Admin.
Executing a plan with the virtual agent requires a Windows device that has been added for unattended access.

Create and execute plans from a knowledge base article

Knowledge base articles can contain multiple virtual technician plans that include reusable automation steps in human-readable format, as well as background information on those steps. This makes it possible to create customer-specific knowledge and domain expertise and is a great choice to automate standard operating procedures that require customer-specific knowledge and domain expertise. For information on running plans, see Run a virtual technician plan from a knowledge base article.

Create and execute plans during a support session

You can run a virtual technician plan during a remote support session started from the Devices page. The agent can generate a plan on the fly based on a custom prompt. Agents can also open the virtual technician plan library to quickly access plans stored in the knowledge base. For more information, see Run a virtual technician plan during a support session.

When does the open beta start and who will have access to the feature?

The virtual technician open beta starts at the end of March, 2025 and is available to every user with a LogMeIn Resolve Premium or MSP license. During the beta period, the feature is available for testing without additional costs, but usage is limited per company account, which is reset every 30 days. To raise your monthly limit, reach out to your Customer Success Representative.

What settings are available to admin users?

The virtual technician can be configured for each user in GoTo Admin as follows:

Enable the virtual technician as a whole
Enable plan execution from a knowledge base article
Enable plan execution during a support session

How does the virtual technician work?

The virtual technician captures the device's screen and analyzes its elements. Based on the plan, it determines what actions to take to fulfill the plan, for example, click a button, open a menu, or enter text. It will then try to perform these actions one by one and report the outcome along with a new capture of the resulting screen for analysis. This iterative process continues until the plan is considered completed either successfully, or with a failure.

How does the virtual technician handle unexpected situations and recover from errors?

The AI gathers as much context as possible from the plan, the latest screen content, and the outcome of the last action taken on the device. By combining all this information, the virtual technician can handle unexpected situations and errors not explicitly covered in the plan. It does this by retrying failed actions and considering alternative methods to achieve the desired goal. This allows the virtual technician to adapt effectively to the device's specific state and behavior during execution, providing a level of flexibility that static scripts or traditional automation solutions would struggle to match.

Are there any limitations to the virtual technician?

The virtual technician is a beta feature, currently under development, which is why we recommend to test and supervise execution when testing the feature.
The virtual technician provides flexibility and ease-of-use by relying on plans provided in natural language, but this also implies that execution is not as rigid and deterministic as e.g. programmatic scripts or classic RPA solutions that rely on pre-recorded UI workflows.
The agent or end-user can cancel the execution of plans any time, but currently the execution cannot be paused and resumed.
While the virtual agent can confirm User Access Control (UAC) dialogs, actions that require entering custom credentials are not yet supported. We also recommend not to write secrets in the prompt or plan for security reasons.
The virtual technician works best with English language prompts, plans, and applications. There is no official support for other languages at the moment.

Does a virtual technician plan work for different devices?

It depends on the specific content of the plan. When generating a plan from a prompt, the name and version of the operating system running on the remote device is provided to the AI as context-aware information. For example, the resulting plan may support devices running Windows 10 and 11, provided that the steps required to execute the plan are not too different on these operating systems and the required applications are available on all devices.

How to write good plans?

A good plan for the virtual technician consists of two components: a goal, and a list of steps.

The goal captures the intent of the user request. It should be actionable and sufficiently specific when using the AI feature to generate a list of steps.

Examples of a good goal:
- Clear the cookies of the Google Chrome browser
- Run an antivirus scan using the Windows Security app
- Open PowerShell and run the command 'ping www.goto.com'
Examples of a bad goal:
- Fix the Outlook configuration: it is unclear what the issue is and what action would solve it.
- Create a backup of hard drive C: it is unclear what backup solution to use and where to store the backup.
The list of steps should give guidance to the virtual technician on how to reach the goal:
- Each step should have a clear intent (for example, “Close the application”), but should not be over-specific and limit the AI’s options of how to fulfill the intent (for example, “Press the Alt-F4 keyboard shortcut to close the application”).
- Steps can include customer- or solution-specific requirements to solve an issue in a specific way, such as defining specific applications or URLs.
- The steps serve as guidelines for the virtual technician. While it aims to follow these steps closely, it may deviate when it either finds step impossible or unnecessary to execute, or if it identifies a better alternative to achieve the goal. This flexibility enables the AI to perform actions not detailed exhaustively (for example, navigating through an installation wizard), take actions even if the expected step description differs from the observed device state (for example, pressing an "execute" button when the step mentions a "start" button), and manage unexpected situations and errors not anticipated in the plan (for example, closing a pop-up window).

Does the plan library offer template plans?: The knowledge base has an option to import sample articles that contain virtual technician plans. These plans are available in the plan library after the import.
Can a sequence of plans be started for execution, or only one plan at a time?: Right now, you can only run one plan at a time.
Can the same plan be executed in parallel on multiple devices?: You can run multiple support sessions in parallel and start a virtual technician plan in each session.
Can the virtual technician run diagnostics after executing a plan to ensure that an issue has been fixed?: Diagnostic steps can be executed like any other step in the plan, but they need to be detailed explicitly by the plan.
Can it collect and report information to the human agent, or store it in the ticket or session notes?: Simple results can often times be reported in the success message of the execution using a final step like “Report the min/avg/max ping times to the user”. However, this information is not persisted as part of a session report or ticket.

Article last updated: 31 March, 2025