zofiQ Demo Speaking Points
Generated: 2026-05-27

Purpose
This demo is split into two clear stories:

1. zofiQ AI Assistant - "AI you see"
   The technician-facing Copilot experience inside ConnectWise and app.zofiq.ai.

2. zofiQ AI Agents - "AI you don't see"
   Background workflow agents that classify, enrich, route, report on, and monitor tickets.


============================================================
PART 1 - zofiQ AI Assistant: AI you see
============================================================

Positioning
- zofiQ Copilot is the AI assistant technicians interact with directly.
- It is designed for MSP service desk workflows, not as a generic chatbot.
- It sits inside the technician's normal ConnectWise ticket workflow and helps answer: "How do I solve this?"
- The goal is to reduce context switching, speed up ticket understanding, and support better decisions while the technician remains in control.

Where it lives
- Embedded inside ConnectWise service tickets as a zofiQ POD.
- Available as a standalone web app at https://app.zofiq.ai.
- The embedded ConnectWise POD is the best place to show ticket-grounded assistance because it can fetch the current ticket context.
- The standalone app is useful for broader questions, reviewing past conversations, managing workspaces, and using custom assistants.

Setup speaking point
- A ConnectWise administrator performs the initial POD setup once.
- The POD is added to the Service Ticket screen through Manage Hosted API setup.
- Technicians then add the zofiQ POD to their ticket layout through Screen Layout Settings and Pod Configuration.
- First-time authentication should be completed through the ConnectWise Manage web app in Chrome or Edge.

Suggested demo flow
1. Open a ConnectWise service ticket.
2. Point out the zofiQ POD inside the ticket screen.
3. Explain that Copilot is available where technicians already work, so they do not need to leave the PSA.
4. Fetch the ticket context or show AutoFetch if enabled.
5. Ask Copilot for a ticket summary, likely issue, and recommended next steps.
6. Use a follow-up question chip to continue the workflow.
7. Show time entry with #time or the Add Time Entry button.
8. If useful, show #web for current vendor documentation or external error-code research.
9. Upload or reference a screenshot/PDF/CSV to show file and image analysis.
10. Close by showing how managers can create custom AI assistants and knowledge bases.

Core demo points

AutoFetch
- AutoFetch gives technicians immediate context when they open a ticket.
- It can surface ticket metadata, blindspot insights, path-to-resolution guidance, relevant documentation, similar resolved tickets, and time entry options.
- This is a strong demo moment because it shows value before the technician even writes a prompt.
- Speaking line: "Instead of asking the technician to go hunting through the PSA, documentation, and old tickets, zofiQ brings the starting context into the ticket."

Blindspot Insights
- Blindspot Insights help surface things that are easy to miss during manual triage.
- Examples include similar recent tickets, repeated SLA issues, or signals that a contact or company may be high impact.
- Speaking line: "This is not replacing judgement. It is highlighting context the technician may not have time to manually search for."

Path to Resolution
- Path to Resolution gives directional starting points.
- It is not intended to fully solve the ticket automatically.
- It helps the technician decide what to check first.
- Speaking line: "The assistant gives a structured starting point, so the technician can move from reading the ticket to investigating faster."

Documentation surfacing
- Copilot can surface relevant IT Glue or Hudu documentation.
- For IT Glue, it can link company-level documentation, company-specific documents, relevant documents from other companies, and flexible assets.
- For Hudu, it can surface the top relevant documents based on ticket context.
- Speaking line: "This helps documentation become usable at the moment of need, instead of relying on the technician to know what to search for."

Similar tickets
- Copilot can surface previously resolved tickets that resemble the current issue.
- It can show company-related examples and similar issues from other companies.
- Speaking line: "This turns past resolution history into a reusable troubleshooting asset."

Time entry
- Copilot includes an Add Time Entry button and a #time command.
- The time entry experience mirrors native ConnectWise time entry.
- Technicians can set start/end time, deductions, work type, billable status, work role, note type, recipients, and AI-generated notes.
- Speaking line: "This is about reducing the admin burden after the technical work is done."

Follow-up questions
- Copilot suggests clickable follow-up questions after responses.
- These keep the technician moving without having to think of the next prompt from scratch.
- Examples include drilling deeper, drafting a customer response, summarizing findings, or preparing notes for escalation.
- Speaking line: "The assistant keeps momentum in the workflow by suggesting the next useful action."

Special tool calls
- #time opens time entry from the chat.
- #web performs a targeted web search and returns a cited, summarized answer.
- Speaking line: "Technicians can bring up specific tools from the same conversation without changing screens."

Image and file analysis
- Copilot supports PNG, JPG, JPEG, PDF, and CSV files.
- It can interpret screenshots, extract error messages, explain what the content likely means, and suggest next steps.
- Speaking line: "When a user submits only a screenshot, Copilot can help turn that screenshot into an actionable starting point."

RMM device context
- Copilot can connect to NinjaOne and Datto RMM.
- These integrations are read-only.
- A manager configures the integration, and Copilot can then answer device questions when ticket context is available.
- Example questions:
  - "What devices are associated with this company?"
  - "Tell me about all specs for device ABC-LAPTOP-01."
- Speaking line: "This lets technicians access device context from the ticket conversation instead of opening the RMM separately."

Supported RMM examples
- NinjaOne can return device ID, system name, last logged-in user, OS, hardware, storage, IP/MAC details, warranty dates, reboot status, and other inventory details.
- Datto RMM can return device ID, hostname, site, device type, OS, internal/external IP, last logged-in user, patch status, antivirus status, web remote URL, and other device details.

Workspaces, AI Assistants, and Knowledge Bases
- Managers can create custom AI assistants for specific roles, workflows, or response styles.
- Assistants can have behavior prompts, expected output prompts, prompt suggestions, knowledge base attachments, and access rules.
- Knowledge bases can be built from uploaded files, directories, or direct text content.
- Supported knowledge base content includes CSV, Word documents, PDFs, and TXT files.
- OCR is used where readable text is embedded in images inside supported documents.
- Speaking line: "This is where the assistant becomes tailored to how the MSP actually works, instead of staying generic."

Important limitations and caveats for the demo
- Copilot augments technician workflows; it does not replace the technician.
- RMM integrations are read-only.
- AutoFetch works in the ConnectWise POD; the standalone app requires the user to manually fetch a ticket by typing "Fetch" plus the ticket number.
- Manager access or higher is required to configure integrations and workspace assets.
- Chrome or Edge is recommended for the first-time web setup.


============================================================
PART 2 - zofiQ AI Agents: AI you don't see
============================================================

Positioning
- zofiQ AI Agents are background workflow agents.
- They are designed to act on tickets based on configuration, SOPs, historical patterns, and ticket context.
- Unlike Copilot, technicians do not chat with these agents during normal work.
- They run behind the scenes to improve classification, routing, assignment, reporting, and oversight.

High-level value
- Reduce manual ticket triage.
- Standardize ticket quality.
- Improve routing accuracy.
- Reduce dispatch effort.
- Make ticket decisions auditable.
- Give managers visibility into performance and SLA risk.

Suggested demo flow
1. Open agents.zofiQ.ai.
2. Show the Workflows Dashboard.
3. Explain that each row is a purpose-built agent.
4. Point out live/off toggles and Shadow Mode.
5. Open one agent configuration page.
6. Show trigger settings, configuration, SOP field, simulation, and workflow logs.
7. Run or describe simulation against a ticket number.
8. Open Workflow Logs and show reasoning.
9. Explain the rollout path: configure, simulate, run Shadow Mode, review logs, refine SOPs, enable live mode.

Platform navigation
- Managers access the platform through agents.zofiQ.ai.
- The Workflows Dashboard lists available agents.
- Each agent can be opened for configuration.
- Agent status is controlled through toggles.
- Green means live and writing to the PSA.
- Gray means off.
- Shadow Mode is available for agents that would otherwise modify tickets.

Core control model
- Trigger conditions decide when an agent is eligible to run.
- Configuration options decide what the agent can change.
- SOPs provide natural language business rules and guardrails.
- Simulation lets teams test behavior against specific tickets.
- Workflow Logs show the agent's decisions, updates, and reasoning.

Going live
- Agents can be enabled gradually and independently.
- When live, agents process net-new tickets that meet configured boards, filters, and SOP conditions.
- Each qualifying ticket is evaluated once at creation or trigger time.
- Agents do not continuously re-run on the same ticket.
- Tickets created before an agent is enabled are not retroactively modified.
- Most execution cycles complete within a few minutes, subject to PSA API limits.

Safety and auditability
- Workflow Logs capture decision details and reasoning.
- PSA audit trails show applied ticket changes.
- If a human is actively modifying a ticket at the same time an agent attempts a write, the agent halts and does not apply changes.
- This protects against conflicting writes while preserving normal human control.

Shadow Mode
- Shadow Mode lets agents run on real incoming tickets without writing changes back to ConnectWise.
- It records predictions and decisions in Workflow Logs.
- It is best used after an agent has been fully configured.
- Shadow Mode and Live Mode cannot be enabled at the same time.
- There is no automatic promotion from Shadow Mode to Live Mode.
- Speaking line: "Shadow Mode is the confidence-building step. You see what the agent would have done before allowing it to change anything."

Recommended rollout
1. Configure the agent.
2. Validate on sample tickets using simulation.
3. Enable Shadow Mode on real incoming tickets.
4. Review the Shadow Mode tab in Workflow Logs.
5. Refine configuration or SOPs.
6. Enable Live Mode when predictions are consistently correct.

Agent orchestration order
When multiple triage agents are enabled, zofiQ follows an orchestration order:

1. Company Agent
2. Urgency, Impact, and Priority Agent
3. Board, Type, SubType, Item Agent
4. Contact Agent
5. Agreement Agent
6. Dispatch Agent
7. Summary Line Agent

Speaking line: "The order matters because later agents depend on clean context from earlier agents."


Key agents and speaking points

Company Agent
- Ensures the ticket is associated with the correct company.
- Useful when tickets enter under placeholder companies such as Catchall or Unknown.
- Runs early because downstream contact, agreement, dispatch, and routing decisions depend on correct company context.
- Can update only the Company field.
- Important caveat: changing Company in ConnectWise resets the Contact field to the default contact for the new company.
- Important risk note: if downstream workflows send customer information based on company assignment, an incorrect company assignment could expose information to the wrong customer. This should be considered before enabling.
- Speaking line: "This agent establishes the customer context so the rest of the workflow has a clean foundation."

Urgency, Impact, and Priority Agent
- Classifies how critical an issue is, what it affects, and the resulting priority.
- Can update Urgency, Impact, and Priority together.
- Respects ConnectWise priority calculation rules and dependencies.
- Can use Priority First mode for MSPs that do not rely on the standard urgency/impact matrix.
- Speaking line: "This standardizes severity classification early, so SLA expectations and response urgency are more consistent."

Board, Type, SubType, Item Agent
- Assigns or corrects ticket Board, Type, SubType, and Item.
- Helps tickets land in the right queues and follow the right workflow.
- The agent only updates the fields explicitly enabled in configuration.
- It respects native ConnectWise dependency rules between Type, SubType, and Item.
- Speaking line: "This is the automated routing and classification layer that reduces triage rework."

Contact Agent
- Ensures the ticket is linked to the correct requester or end user.
- Runs after Company Agent because valid contacts depend on the correct company.
- Can update only the Contact field.
- Can optionally add an internal note: "Attention: This ticket was triaged by zofiQ AI Agent."
- Contact Creation Mode can be disabled, limited to company domains only, or allowed for all domains.
- Speaking line: "This improves communication accuracy and helps downstream agreement and dispatch logic work with the right requester."

Agreement Agent
- Ensures the correct agreement is applied to the ticket.
- Helps with billing accuracy, SLA alignment, and contract-specific workflows.
- Can update only the Agreement field.
- By default, it does not overwrite an existing agreement.
- Override mode can be enabled when placeholder or default agreements need to be corrected.
- Speaking line: "This is about making sure the commercial and SLA context is right before work is assigned."

Dispatch Agent
- Automatically assigns ownership based on team configuration, technician availability, workload, and historical context.
- Runs after triage is complete.
- Can update ticket owner, ticket team, ticket status, and in one mode create a ConnectWise calendar entry.
- Supports three dispatching modes:
  1. Assign owner and schedule calendar slot.
  2. Assign owner only.
  3. Assign default team only.
- Mode 1 uses calendar availability, working hours, in/out board status, workload density, and historical success where available.
- Mode 2 uses active ticket count and workload thresholds without calendar scheduling.
- Mode 3 routes to the board's default team without selecting an individual technician.
- Speaking line: "Dispatch turns triage context into ownership, using the operating model the MSP already follows."

Summary Line Agent
- Improves vague or poorly formed ticket summary lines.
- Runs last so it can use the final ticket context after earlier agents have completed.
- Can update only the Summary field.
- If the existing summary is already good, it leaves it unchanged.
- Speaking line: "This improves ticket readability and makes queues easier to scan."

Insights Agent
- Read-only reporting agent.
- Analyzes historical ticket activity and emails structured reports.
- Main report runs weekly and provides an organization-wide view.
- Sub-reports can focus on specific boards, customers, teams, or filtered datasets.
- Sub-reports can run daily, weekly, or bi-weekly on custom schedules.
- Uses a fixed seven-day lookback window.
- Does not modify tickets or write back to the PSA.
- Speaking line: "Insights is for trend visibility, not automation. It helps managers understand what changed and where attention is needed."

Oversight Agent
- Read-only monitoring agent focused on SLA risk and potentially orphaned tickets.
- Runs twice daily and can also be triggered on demand.
- Identifies tickets that have breached SLA or are within one hour of breach.
- Uses ConnectWise SLA status as the source of truth.
- Does not modify tickets or write back to the PSA.
- Reports are delivered by email and grouped for actionability.
- Speaking line: "Oversight is the safety net. It helps leaders see what may be missed before it becomes a bigger service issue."


============================================================
COMPARISON: AI you see vs AI you don't see
============================================================

AI you see - Copilot
- Technician-facing.
- Conversational.
- Used inside the ticket workflow.
- Helps with understanding, troubleshooting, documentation, similar tickets, time entries, and device context.
- Best demo moments are AutoFetch, follow-up questions, documentation surfacing, #time, #web, file/image analysis, and custom assistants.

AI you don't see - Agents
- Manager-configured.
- Runs in the background.
- Acts on ticket fields, routes work, dispatches owners, monitors risk, or generates reports.
- Best demo moments are Workflows Dashboard, Simulation, Shadow Mode, Workflow Logs, SOPs, and agent orchestration.

Simple positioning line
"zofiQ has two layers: Copilot is the AI your technicians see and interact with; Agents are the AI workflows operating behind the scenes to keep tickets cleaner, faster, and more consistent."


============================================================
DEMO TALK TRACK
============================================================

Opening
"Today I will show zofiQ in two parts. First, the AI assistant technicians can see and use directly inside ConnectWise. Second, the AI agents they do not see, which run behind the scenes to triage, route, dispatch, report, and monitor tickets."

Transition to Copilot
"Let's start with the technician experience. The key point here is that zofiQ is not asking the technician to leave their workflow. It appears inside the ConnectWise ticket, where the work is already happening."

AutoFetch moment
"When the ticket opens, AutoFetch can bring in the surrounding context: metadata, blindspot insights, a starting path to resolution, documentation, similar tickets, and time entry options. This gives the technician a stronger starting point before they manually search across systems."

Troubleshooting moment
"From here the technician can ask natural language questions. They can ask for a plain-English summary, suggested investigation steps, customer-facing updates, escalation notes, or more detailed technical checks."

Documentation moment
"One of the stronger parts of the workflow is that relevant documentation and similar resolved tickets can be surfaced in context. This helps make existing operational knowledge easier to reuse."

Admin reduction moment
"Once the technician has done the work, they can use the time entry workflow or #time command to reduce the admin burden of documenting that work."

Workspace moment
"For managers, Workspaces lets the team create custom assistants and knowledge bases. That means the AI can be shaped around internal processes, preferred response formats, and trusted documentation."

Transition to Agents
"That is the AI you see. Now we move to the AI you do not see: background agents that improve the ticket lifecycle without requiring a technician to prompt them."

Agent dashboard moment
"Each agent is purpose built. One handles company assignment, one handles priority, one handles routing classification, one handles contact, one handles agreement, one handles dispatch, and others handle summaries, reporting, and oversight."

Trust and safety moment
"The important part is control. Agents can be configured, tested in simulation, run in Shadow Mode, and reviewed through Workflow Logs before they are allowed to write changes live."

Shadow Mode moment
"Shadow Mode is especially useful during rollout because it shows what the agent would have done on real incoming tickets without changing anything in ConnectWise."

Workflow Logs moment
"Workflow Logs provide the reasoning. This makes the automation auditable: what ticket was evaluated, what field would be changed or was changed, and why the agent made that decision."

Closing
"The outcome is a layered approach: technicians get an assistant that helps them work faster in the moment, while managers get background agents that improve consistency, routing, dispatch, billing context, reporting, and SLA visibility."


============================================================
PRACTICAL CAVEATS TO MENTION IF ASKED
============================================================

- Agents that modify tickets generally run once per ticket and do not automatically retry.
- Live agents process net-new qualifying tickets, not historical backlog.
- Shadow Mode does not write to ConnectWise.
- Read-only agents such as Insights and Oversight do not need Shadow Mode because they do not modify tickets.
- RMM integrations in Copilot are read-only.
- Throughput is currently up to 10 tickets per minute, with actual throughput governed by PSA API rate limits.
- During large ticket bursts, newer tickets are prioritized; some older burst tickets may be skipped and require manual handling.
- If a user is modifying a ticket at the same time an agent attempts to write, the agent halts to avoid overwriting human changes.
- SOPs are powerful guardrails, but they should be tested through simulation and Shadow Mode before live rollout.
- Company assignment automation should be reviewed carefully if downstream workflows send customer-visible information based on company assignment.


============================================================
SOURCE KB PAGES REVIEWED
============================================================

zofiQ AI Assistant / Copilot
- https://support.zofiq.ai/collections/1071162496-zofiq_co_pilot
- https://support.zofiq.ai/articles/1233487640-zofiq-copilot-initial-pod-setup-for-connectwise-psa-manager-only
- https://support.zofiq.ai/articles/8839367876-zofiq-copilot-user-guide-and-operational-overview-connectwise
- https://support.zofiq.ai/articles/3733027760-zofiq-co-pilot-overview
- https://support.zofiq.ai/articles/6671886359-zofiq-copilot-additional-integrations
- https://support.zofiq.ai/articles/5034951514-zofiq-copilot-workspaces-ai-assistants-knowledge-bases

zofiQ AI Agents / Agentic Workflows
- https://support.zofiq.ai/collections/9309902189-zofiq_agentic_workflows
- https://support.zofiq.ai/articles/5236972995-navigating-your-agent-environment-connectwise
- https://support.zofiq.ai/articles/1930851392-connectwise-agents-testing-shadow-mode
- https://support.zofiq.ai/articles/4887017013-company-agent-connectwise
- https://support.zofiq.ai/articles/1402444257-urgency-impacy-priority-agent-connectwise
- https://support.zofiq.ai/articles/9284294753-board-type-subtype-item-agent-connectwise
- https://support.zofiq.ai/articles/1037733857-contact-agent-connectwise
- https://support.zofiq.ai/articles/8924255825-agreement-agent-connectwise
- https://support.zofiq.ai/articles/7389059056-dispatch-agent-connectwise
- https://support.zofiq.ai/articles/5790243210-summary-line-agent-connectwise
- https://support.zofiq.ai/articles/9856869279-insights-agent-connectwise
- https://support.zofiq.ai/articles/2264735043-oversight-agent-connectwise
