---
title: "Vic Sun’s playbook"
description: "Calling, Field Coach, and Call Center as one operating playbook."
---
> Documentation Index
> Fetch the complete documentation index at: https://guides.channelautomations.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Vic Sun’s playbook
[Open original guide with interactive examples →](/originals/calling-field-coach-call-center/)
## What this playbook covers
## Calling
The rep’s daily workflow in **both the CRM and standalone app**: Call Guide, Power Dialer, Text, Recent Calls, Callbacks, appointments, Team Stats, Second Chance and Visit reps.
[Walk through Calling →](#calling-tour)
## Field Coach
Turn recorded phone calls and field conversations into summaries, evidence, coaching and scorecards. Managers review team performance, process steps and analysis rules.
[Record and review an in-home demo →](#home-demo)
[All Field Coach workflows →](#coach-tour)
## Call Center
The standalone workspace for calling operations. Reps can work outside the main CRM while using the same workspace and shared services. Managers organize queues, review activity and open coaching.
[Walk through Call Center →](#center-tour)
## Choose your starting point
[1\. Understand and sell the whole solution](#sell)[2\. Practise a rep’s working day](#calling-tour)[3\. Run and coach the team](#manager-tour)
**The complete loop:** assign leads → call or text → save the outcome → honour callbacks or book a visit → review results and recordings → coach the team → recover eligible missed opportunities.
## Call Center: enter the right workspace
1. From the CRM dashboard, open **Calling** in the dock or Apps launcher. Choose **Open Call Center** to enter the standalone app, or open your organization’s standalone URL directly and sign in.
2. Confirm the workspace name and your account before choosing a lead. Shared workspace data does not mean permission to see every contact; member roles, queue access and contact visibility still apply.
3. Use the standalone sidebar: **Calling** for the rep workflow, **Call Queues** for queue work within your permissions, and **Field Coach** for conversation review.
4. Use CRM administration where the standalone app links back to it. Calling-hours editing and full analytics do not need duplicate standalone screens.
**Names matter:** “Call Center” here means the standalone operating app. The separate sidebar item “Contact Center” is a different surface; access to Calling does not mean every other product or permission becomes enabled.
> “Your reps get a focused place to work, with the same workspace behind it. Your manager can keep broader administration in the CRM.”
## Calling: practise one rep’s working day
**Where:** CRM dashboard → Calling, or standalone sidebar → Calling. Start with a manager-assigned test lead for any real demonstration.
1. **Pick the right work:** choose your queue in Call Guide. Confirm the selected homeowner, phone, lead status and prior context. Queue membership and access decide which leads are available.
2. **Prepare the conversation:** read the suggested guide and latest contact history. Confirm device readiness before Start call. Power Dialer supports successive queue work; it is not permission to bypass calling restrictions or skip wrap-up.
3. **Choose the next action:** call, or use Text for an appropriate message. Read and explicitly send the intended text; a draft is not delivery.
4. **Handle the homeowner’s answer:** if ready, use Offer times and confirm a feasible visit. If busy, record the agreed callback through the normal wrap-up. If unanswered, record the actual outcome rather than inventing a conversation.
5. **Finish the record:** save the disposition and relevant notes before proceeding. Recent Calls provides history and the contact drawer, recording, summary and available qualification evidence.
6. **Keep promises:** return to Callbacks and work due follow-ups. An unanswered retry does not automatically fulfil a promised callback.
7. **Recover missed opportunities:** use assigned recovery work when there is supporting evidence and the lead is still eligible. Second Chance is one part of this daily workflow, not the whole product.
Vic’s rehearsal: Morgan is busy now but wants a bath visit
Explain your selected queue and contact first. Choose Text if Morgan asks for it, or save an agreed callback. When Morgan is ready, explain Offer times, then the saved booking and wrap-up. Finish by opening Recent Calls to show continuity. Practise the decisions in the feature lab; do not send a real test message without an approved recipient.
[Practise Text, Offer times and Callbacks →](#feature-lab)
## In-home sales demos: the Field Agent use case Vic asked about
**Yes—this is Field Coach.** A field salesperson records the conversation at the homeowner’s house or uploads an existing recording. Field Coach transcribes it, summarizes the demo and evaluates the conversation against the published sales scorecard and process steps. The manager can listen to the recording and review the evidence behind missed or partially followed steps.
> “For the HomeFix pilot, we can review what happened during an in-home sales demo: what the homeowner needed, how the rep handled objections, whether the agreed sales process was followed, and what to coach next.”
**What to promise:** recorded-conversation review and coaching after processing. This is not a claim of live manager listening, real-time prompts during the visit, or full feature parity with Siro. Script adherence means evaluating configured rubric and observable process steps; it is not automatic word-for-word verification of any script someone has not configured.
### Before the first visit — manager procedure
1. Open **Field Coach → Team** in the correct workspace. Confirm the workspace AI account is ready for transcription and analysis and that the rep has access.
2. Translate the company’s demo process into the **Scorecard** and **Process steps**. Use observable steps: introduce the agenda, confirm decision makers, discover the problem, explain options, address objections, ask for a commitment and agree the next step. Use the company’s actual process rather than treating these examples as mandatory.
3. Publish the new scorecard version before the pilot. Tell reviewers which version defines the expected process. Publishing does not silently rewrite every old analysis.
4. Agree the recording procedure and obtain the required permission before recording. The app asks the rep to confirm consent; that checkbox does not itself obtain the homeowner’s consent.
5. Run a short approved practice recording on the exact device/browser the rep will use. Confirm both voices are audible, upload completes and the result opens before relying on it for a full visit.
### At the home — rep procedure
1. Open standalone **Field Coach → Uploads** or the CRM Field Coach capture surface. Confirm the workspace, enter a recognizable visit title and associate the correct contact when available.
2. Explain and obtain recording permission, then confirm consent in the app. Place the device where it can capture both voices without obstructing the conversation.
3. Start browser recording and allow microphone access. Check that the recording timer advances. Conduct the demo normally.
4. **Keep the browser in the foreground and the device awake.** Reliable screen-locked/background or offline recording is not established by the current web recorder. Do not promise “leave the phone locked in your pocket.” Alternatively, use a company-approved recording method and upload its supported audio file afterward.
5. Stop recording at the end. Submit the captured file through the upload action; stopping alone does not upload or analyze it. If using an existing file, select it and submit with the required consent confirmation.
6. Wait for upload completion and the queued/processing status. Do not close the page before the captured file has uploaded. Check for Completed or a failure that needs attention before expecting a score.
### After the visit — manager and rep review
1. Open the completed conversation and **Play recording**. Use **Summary** for needs, objections and next actions, then **Transcript** to check the actual dialogue.
2. Open **Scorecard → Your process**. Review followed, partial and missed step verdicts with their explanations. Cross-check the relevant recording and transcript before concluding that the rep omitted a step—poor or incomplete audio can leave evidence missing.
3. Use **Coaching** to identify one specific improvement. Separate presentation quality from process completion: a strong pitch can still end without an agreed next step.
4. Review any suggested **Follow-up** draft for accuracy. It is not automatically sent. Confirm promises with the rep before using the normal messaging workflow.
5. Use **Team** for the appropriate period and compare sufficiently complete, compatible results. Agree one improvement with the rep, then inspect a later demo for that behavior. Use QA review where appropriate if an automated finding appears wrong.
How Auto-analyze rules relate to an in-home demo
Submitting a field recording queues that conversation for processing. Auto-analyze rules primarily select existing recorded phone calls for automatic import and analysis. They are not a microphone switch and do not start recording at a customer’s home. The same phone-call analysis policy is exposed in Second Chance’s Automatic recovery coverage.
### Practise Vic’s HomeFix demonstration
**Fictional example:** Jordan gives a clear bath presentation and answers a price objection. The recording ends with “I’ll leave the brochure,” without confirming decision makers or agreeing the next step. A polished presentation score alone does not establish a complete sales process.
What should the manager do?
**Check the recording and process-step evidence, then coach the next-step question** **Assume a good overall score means the demo followed every step**
Choose an answer. This practice does not upload audio or run analysis.
### HomeFix pilot acceptance
Before calling the pilot successful, demonstrate one approved in-home-style recording from capture to completed transcript, summary and scorecard; play it back; check the agreed process-step verdicts against the audio; confirm the rep and manager see only authorized records; and rehearse a failed upload/analysis recovery. This guide does not claim that HomeFix has already completed that acceptance.
## Field Coach: learn from the actual conversation
**Where:** standalone sidebar → Field Coach. In CRM, open Field Coach from the dock or Apps launcher when available to your account.
1. **Choose the evidence:** use Uploads for a field conversation, or Calls to find an existing human or Voice AI phone recording. Confirm the right workspace, conversation and contact. Follow the recording-consent requirement before capture or upload.
2. **Wait for a real result:** upload or use the available analysis action within your permissions. Processing is asynchronous. A recording being present does not mean analysis succeeded; review failure status and retry only after resolving its cause.
3. **Read beyond the score:** open the completed conversation. Review Summary, Coaching, Scorecard and Transcript. Use the timestamped evidence to check what was said and choose one improvement.
4. **Review follow-up drafts:** generated SMS/email suggestions are editable drafts, not sent messages. Confirm accuracy and use the approved messaging workflow to contact the homeowner.
5. **Manager review:** open Team, select a reporting period and review completed/failed work, scores, process adherence and QA where available. A missing score is not a zero; limited samples or different rubric versions can prevent a fair comparison.
6. **Set the standard:** edit scorecard sections and observable Process steps with the authorized role, then publish a new version. Set Auto-analyze rules deliberately. This policy is also exposed by Second Chance’s Automatic recovery coverage; changing one changes the shared policy.
Vic’s rehearsal: a polished call with no agreed next step
Open a fictional example in your explanation: the rep handles objections well but never asks for a visit or callback. Show why a high presentation score does not establish process completion. Use transcript evidence to coach one specific behavior: agree the next step before ending the conversation. Do not present an AI estimate as a verified sale.
[Practise Process steps and Auto-analyze rules →](#feature-lab)
## Manager’s day: organize, measure, coach, recover
1. **Before the shift:** confirm queue criteria and rep access, due callbacks, visitor territories/hours/capacity and the analysis policy. A calling rep and visiting salesperson can be different people.
2. **During the shift:** use Team Stats to distinguish dials, carrier connections, reached people and appointments. Check Wrap-up coverage before judging conversion. Use Per agent for targeted review rather than ranking people on tiny samples.
3. **Investigate results:** compare lead sources and project types, Caller IDs and By attempt within one date range. Respect Unknown values and withheld rates. Sold outcomes require the appropriate CRM outcome configuration.
4. **Coach with evidence:** open a relevant recorded conversation in Field Coach; review the transcript and missed process step; agree one behavior to improve. Publish rubric changes deliberately and understand the effect on future comparisons.
5. **Close the loop:** review Second Chance ownership and eligibility, then Appointment follow-ups for unfinished work. Retry a failed follow-up after correcting the cause; do not create a duplicate booking.
[Detailed recovery and visit setup](#setup) · [Recovery states and follow-up operations](#manage) · [Practise the manager’s decisions](#feature-lab)
## How Vic should explain it
> “We connect the work of calling leads with the work of improving your sales team. Reps work assigned queues in your CRM or the standalone Call Center, call or text with context, keep callbacks and book suitable visits. Managers see the outcomes, use Field Coach to review real conversations and coach the process, then recover interested leads that still need an appropriate next step.”
### The rep’s problem
“Who should I call next, what happened last time, and which visit can I confidently offer?”
**Benefit:** an actionable queue, the original evidence and explainable appointment options.
### The manager’s problem
“Which interested leads slipped through, who owns them, and can my field team fulfil the bookings?”
**Benefit:** visible recovery decisions, ownership, visit rules and unfinished follow-ups.
### Discovery questions for a prospect
1. How do you currently find interested callers who never booked?
2. Who owns that second attempt, and how do you prevent duplicate callbacks?
3. Does the calling rep know each salesperson’s territory and daily visit limit?
4. How do cancellations become a follow-up opportunity?
5. How much of your recorded calling is analyzed, and who owns that policy?
### A complete ten-minute demonstration
1. **Minutes 1–2 — Calling:** start in CRM Calling, show the selected queue/contact and Open Call Center. Explain the same working context in standalone.
2. **Minutes 3–4 — Rep actions:** show Text, a callback promise, Offer times and saved wrap-up. Explain what needs an explicit human action.
3. **Minutes 5–6 — Manager visibility:** show Team dialing, coverage, Per agent and source/attempt breakdowns. Explain one missing rate honestly.
4. **Minutes 7–8 — Field Coach:** open a completed recording, inspect evidence and a process step, then explain the shared analysis rules.
5. **Minutes 9–10 — Follow-through:** show Second Chance, Visit reps and Appointment follow-ups. Use the fictional simulator for booking changes until live acceptance is confirmed.
**Do not promise:** guaranteed bookings, automatic calls or texts from candidate creation, every call being analyzed regardless of policy, globally optimal routing, live traffic certainty, or cancellation of an appointment in an external calendar through a local-only change.
## Recovery and visit setup: prepare the workspace
1. **Access:** sign in and select the correct workspace. In the CRM, open **Calling** from the dock/app launcher. In standalone, open its Calling screen. Your workspace owner/admin/manager role is required for manager panels. Calling, Number Rotation and Field Coach are available to every existing and new workspace without a rollout flag. Open them from the CRM dock or Apps launcher. Workspace permissions, recording consent, configured phone numbers and manager-controlled automation settings still apply.
2. **Analysis coverage:** open **Second Chance → Automatic recovery coverage**. These are the existing Field Coach workspace rules, shared between apps. A workspace AI account must be configured. Enable only under your company’s recording policy.
3. **Set realistic coverage:** select Human and Voice AI recorded calls for both source types. Sample 100% for broad coverage within your minimum duration and daily cap. Lower sampling means some established-rep calls are intentionally not analyzed. New-rep rules still respect the duration floor and cap. Save analysis rules; processing runs hourly, followed by recovery scanning. Saving does not immediately backfill all older calls.
4. **Configure visitors:** open **Visit reps**. Choose a workspace member, display name, base address and timezone. Set eligible states/postcodes, project types and appointment types. A restriction is a filter; inconsistent restrictions can exclude a rep entirely.
5. **Set visit hours and capacity:** enter working windows, daily visit limit, buffers and maximum drive time. These are visiting hours, separate from calling hours in CRM settings. Verify the base address has usable coordinates. Enable and save each visiting profile.
6. **Prepare leads:** confirm contact address/location, product and territory fields. Configure an enabled appointment type with its correct duration. Types requiring payment or confirmation still use their existing booking-page workflow.
7. **Route recovery work:** in the CRM’s Agent Command Center call-queue setup, enable the Second Chance/recovery criterion on the intended queue and configure its rep access and other criteria. Keep existing cadence limits. Candidate assignment and queue access must both let the intended rep see the lead.
8. **Test one controlled lead:** use your approved test contact. Check quote, visibility, assignment, date/timezone, visiting rep, save result, reminder processing and cancellation. Complete the go-live checklist below before training the team.
Selecting a candidate’s calling rep does not assign that person to the home visit. Visit reps are configured separately; the booking retains who booked it and who is visiting.
## Recovery procedure: work an eligible missed opportunity
1. Open Calling → **Call Guide** and choose your recovery-enabled queue. Review the selected contact’s name and phone before taking an action.
2. Read the recovery reason and original transcript quote. Treat them as context, not a guarantee that the customer still wants an appointment. Check the latest conversation and any existing callback or booking.
3. If the lead should wait, use Follow up later and choose the agreed future time. If recovery is inappropriate, dismiss it with a useful note. Do not reset exhausted cadence simply because the lead appears interested.
4. Use **Start call** when your device and calling checks are ready. For Text, open the contact’s messaging action, review the conversation, write the intended message and explicitly send it. Recovery itself does not send anything.
5. Confirm the homeowner’s project, address, decision-makers and availability. Say: “I can check visits for that date. Let me confirm who can reach you and still make their next appointment.”
6. Open **Offer times**. Choose the appointment type and date; request matching times. Check the displayed timezone, rep, workload and drive-in/onward explanation. The top result favours remaining capacity, then lower combined travel; it is not necessarily the closest rep.
7. Select a time agreed with the homeowner and confirm. An offer is temporary, not a reservation. If the calendar changed or the offer expired, refresh and reconfirm the new option.
8. Read the saved result. If follow-ups are pending, the appointment is still saved. Do not create a second booking to “fix” pending reminders.
9. Save the call’s wrap-up/disposition through the normal Calling workflow. Record any relevant promises or special instructions for the next rep and visitor.
What if the connection drops during confirmation?
Use **Retry same request**. The app keeps the original request so the backend can return the original result. Do not change the appointment details and create another request until the first outcome is known.
How do I change a saved visit?
Open Offer times and select the existing internal appointment. Choose Reschedule, find and confirm a new time, or choose Cancel, enter the reason and confirm cancellation. The appointment ID is preserved on reschedule. If another user changed it, refresh before trying again. Appointments owned by external calendars must be changed at their source.
## Manager routine
### Start of day
Review Second Chance → Ready. Check that the reason is supported by the quote, assign visible leads, and distinguish review status from current queue eligibility. Verify visitor availability/capacity before reps offer visits.
### During the day
Review deferred and dismissed decisions when needed. Assignment-only saves preserve the follow-up decision. Use status filters and Load more to review the complete list. A stale-screen warning means someone else changed the row; refresh.
### Booking follow-ups
Open Appointment follow-ups. Queued/Processing items need time. After correcting an underlying issue, use Retry on Needs review items. Completed steps remain completed and no second appointment is created. Refresh to verify progress.
### Understand the states
| State | Meaning and next action |
| --- | --- |
| Ready | Reviewed for recovery; current contact, booking, callback, queue and cadence checks still determine whether it can be dialed. |
| Deferred | Do not follow up before the chosen time. The system preserves this decision when newer evidence is scanned. |
| Dismissed | Excluded from recovery until a manager explicitly changes the decision. New scan evidence does not silently undo dismissal. |
| Resolved | A booking has addressed this opportunity. A later qualifying call can create a fresh recovery opportunity if there is no current booking. |
| Ineligible | Current contact state prevents recovery, such as a callback already owning the follow-up or a do-not-call status. Fix the actual condition; do not bypass safeguards. |
## Practice with a fictional homeowner
No customer data is used. All drive times below are invented training values, not a map-provider result.
ScenarioHomeowner’s territoryAlex is at daily capacityManager previously dismissed recovery
**Reset scenario** **Evaluate recovery** **Simulate reschedule**
### Morgan Ellis · bath project
**Original fictional quote:** “We still want to replace the bath. My partner wasn’t here to agree on a time. Please check with us again.”
The next-day visit requires 60 minutes. The caller who books is Jordan; the visiting rep will be selected separately.
Review → Match visit → Saved → Cancellation recovery
How this simulation ranks reps
Alex covers North, has 1 of 4 visits booked, and fictional travel of 25 minutes in plus 20 onward. Sam covers North and South, has 2 of 4 visits booked, and travel of 10 minutes in plus 15 onward. Alex ranks first for North while below capacity because the implementation compares booked/capacity first; Sam has shorter travel but more of the day committed. Territory and full capacity exclude a rep before ranking. Production also checks project/type eligibility, working windows, appointment conflicts, buffers and actual provider travel.
## When to stop and resolve the cause
| What the rep sees | Correct action |
| --- | --- |
| No recovery candidates | Check analysis policy, recording availability, sample/cap and processing completion; then current bookings, callbacks, contact eligibility and queue criteria. Empty is not proof that every call was successfully recovered. |
| No matching visit | Check territory/product/type, visitor hours, capacity, buffers and contact/base coordinates. Try another date. Do not promise a time outside the returned options. |
| Road travel unavailable | Resolve configuration/provider availability. The app does not replace missing road travel with straight-line distance. |
| Schedule changed | Refresh offers and confirm the new choice with the customer. A stale offer cannot safely reserve capacity. |
| Follow-ups need review | The appointment is already saved. Have a manager correct the dependency and retry unfinished follow-ups. |
| External calendar appointment | Make the change in the owning calendar/provider, then verify it is reflected in the CRM. A local record change cannot guarantee provider cancellation. |
| Calendar link after cancellation | Current stored links are cleared by the new Calling path. Previously shared links and events someone already added to a personal calendar cannot be recalled. Explicitly confirm the cancellation with the customer through the approved communication workflow. |
## AI outbound caller pools: configure before calling
**Where:** CRM dashboard → dock or Apps launcher → **Number Rotation → AI voice**. Voice Agent Studio must be enabled for the workspace. A workspace owner, admin or manager configures pools; members can review them.
1. **Create a pool.** Give it a recognizable name. Load imported voice numbers, choose a number from the workspace inventory, and enter its callback destination only after verifying that route. This does not buy or import numbers or change inbound assignments.
2. **Set a shared policy.** Select Configure beside a number. Set AI pool dials per day, seconds between calls, minimum classified calls, decline threshold and cooldown hours. Save number policy. A number in two pools shares the same allowance. Existing single-number AI, human and SMS calling keep their existing rules; this is not an account-wide call cap.
3. **Choose distribution.** Balanced uses allowance fairly; Local preference favors an eligible matching area code; Adaptive uses recorded human-answer evidence. Without enough classified outcomes, the system does not invent a performance score.
4. **Attach the pool.** In a draft Voice Agent Studio outbound campaign, choose the pool under caller selection. In a flow’s Voice AI call action, choose Number pool. A flow queues work; it does not immediately dial. Existing single-number selections remain available.
5. **Review before Start.** Confirm the agent, audience, current contact eligibility, calling window, daily campaign cap and line reserve. For the first real test, use the approved owned-phone fixture with one contact and a cap of one. Ask the release owner before Start.
6. **Respond to status.** Pause pool stops new reservations. Configure → Park leaves callbacks intact. Remove from this pool changes membership only. Quarantine requires review; inventory checks and elapsed time do not automatically undo it. Resume is an explicit operator action with a reason.
QueuedEligibility checksReserve number + allowanceOne call requestExact outcome + cost
**When to wait:** no eligible number, stale inventory or exhausted allowance means no fallback call. A timeout or ambiguous engine reply holds the reservation for review, because the phone may already have rung. Do not retry it from another number.
**Health is evidence, not a spam guarantee.** Generic hangups and call duration do not prove a human answered. Unknown reputation means no verified reputation signal. Performance watch and cooldown use minimum samples; a fresh trial follows cooldown. A carrier-spam feed is not connected by this implementation. Do not promise automatic removal of carrier labels.
**Read performance without over-claiming.** Each caller number shows classified human connections and bookings linked to that exact AI call. Each campaign shows Human reached and Bookings attributed. A booking is credited only when the canonical call record carries an appointment ID for the same call ID; matching the same contact near the same time is not enough. Human-answer percentages stay hidden when classification covers less than 80% of resolved calls.
### Try the reservation boundary
This fictional exercise has one number with a cap of 20 and 19 reservations already used. It makes no network requests.
Pool paused
**Try one reservation** **Reset to 19 of 20**
19 of 20 slots used. One slot remains.
## Vic’s feature practice lab
Every item from the Calling and Field Coach review is covered below. Select a feature, follow its real-app procedure, then try a fictional decision. Practice answers stay in this document: they do not save settings or contact customers.
**Entry points:** CRM dashboard → dock **Calling**; standalone sidebar → **Calling**. Coaching configuration starts at standalone sidebar → **Field Coach → Team**. Manager controls require the appropriate workspace role. Full analytics and calling-hours administration can remain in CRM.
Choose a feature to practise
Choose an answer to see its consequence.
**Next feature**
## Vic’s go-live and demo checklist
These checkboxes are personal notes saved only in this browser. Checking them does not run a test or certify deployment.
**Clear personal checklist**
### Close the sales conversation
> “Let’s start with one team and an agreed analysis cap. We’ll verify the recovery queue, visiting territories and one complete booking/change workflow together. Then your manager can expand it with evidence about the actual workload.”
For technical acceptance, use the implementation ledger and testing notes. This guide is an operating and sales aid, not proof that the release gates passed.
## Give the voice assistant lead context before it calls
**Release status:** local implementation under review, September 9, 2026. Not yet deployed. Do not use this section as confirmation that your workspace can run the new behavior.
From the CRM dashboard, open **Voice Agent Studio → select your agent → Reach out → Configure**. For a flow, open **Automation → your flow → Have a voice agent call this contact**. Enable **Give the assistant lead context**.
1. Keep the basic contact information selected to provide name, contact details, address, product and current lead status.
2. In **Additional lead context**, use **Insert field** to select relevant custom fields, such as decision makers, qualification answers or callback instructions. Add labels so the assistant knows what each answer means.
3. Search for a contact by name or phone, choose the contact, and click **Preview lead context**. Review the exact additional brief. Preview does not queue or place a call.
4. Correct fields marked **Not recorded** when they matter to the conversation. The assistant is told that missing values are unknown. Invalid context skips the call with a visible reason. A temporary lookup failure retries up to six times, then skips; correct the cause before re-enrolling the contact.
5. Save the campaign or flow action. A flow enrolment still queues a call; existing suppression, calling hours, line limits and spending limits still apply.
The contact's current values are read just before dialing. The brief is supplied through the existing calling service before the assistant starts speaking. The selected assistant's role and instructions remain in place. This does not automatically load all notes, appointments or conversation history: select the relevant supported fields.
Try a fictional lead brief
This illustration stays in this page and makes no network requests.
Fictional first name
Decision makers What happens between a flow and the call?
Saved flow context template
|
v
Queued contact -- current contact/custom facts --> resolved brief
| |
| lookup fails -> bounded retries; no call
v
Safety gates + reserved call allowance
|
v
Existing calling service receives notes
|
v
Assistant starts with the lead brief already in its prompt
**First live test, after rollout approval:** use your own phone, an audience of one and a daily cap of one. Preview first. Have the developer verify the initial provider prompt and resulting call record. Do not start the existing draft fixture without explicit approval.
Channel Automation · Help · Last updated 2026-09-08.
Questions? Ask your account manager, or open **Help Docs** inside the app — the same guidance lives there.
Source: https://guides.channelautomations.com/coach/calling-field-coach-call-center/index.mdx