Prepared for: Concierge Contact Center (CCC)
Prepared by: Agento AI
Date: July 30, 2026
Executive Summary
Concierge Contact Center has a significant opportunity to expand its use of Retell AI across its client base, reduce the cost of handling routine inbound calls, and create a stronger technology-enabled service offering.
However, the current Retell account has been developed without the operational structure, performance visibility, cost controls, and ongoing optimization required to scale reliably across multiple client accounts.
Our audit identified unnecessarily high per-minute costs, inconsistent agent configurations, excessive prompt length, unnecessary knowledge base usage, and unsuitable LLM selections.
In addition to high latency, insufficient date/time awareness, under-optimized voices, unclear agent status, and no reliable system for identifying poorly performing calls immediately.
Agento AI will take ownership of the Retell side of CCC's operations by:
- Optimizing and rebuilding existing Retell agents.
- Building properly structured Retell agents for new CCC clients.
- Standardizing how agents are configured, tested, deployed, and maintained.
- Reducing avoidable Retell and telephony costs.
- Improving call quality, latency, data collection, and customer experience.
- Implementing automated post-call quality monitoring.
- Manually reviewing calls that indicate issues and implementing performance optimization.
- Providing ongoing management so CCC's internal team does not need to become Retell development specialists.
What We'll Deliver
Existing Retell Agents
For each existing agent selected by CCC, Agento AI will:
- Audit its existing configuration.
- Identify the required client-specific call outcomes.
- Rewrite or restructure prompts.
- Remove unnecessary instructions and duplicated content.
- Remove unnecessary knowledge base usage.
- Select the most appropriate LLM for each use case.
- Optimize the agent's voice and conversational settings.
- Improve interruption behavior, pacing, responsiveness, and latency.
- Configure date, time, business-hours, and timezone awareness where required.
- Review webhooks, post-call analysis, email workflows, and disposition delivery.
- Test each agent thoroughly before approving it for production.
- Implement internal automated performance monitoring.
- Continue manually monitoring and optimizing each agent after go-live.
New Retell Agents
For every new CCC client that elects to use AI, Agento AI will:
- Gather the client's call-handling requirements.
- Determine whether the client requires a messaging-only or intake/scheduling flow before booking hand-off.
- Build the Retell agent from an Agento AI production-tested framework.
- Customize each agent for the individual practice or business.
- Configure client-specific FAQs, data capture, messages, transfer destinations, and workflows.
- Select the appropriate agent architecture, LLM, voice, and settings.
- Configure integrations and post-call outcomes.
- Test the complete flow before it is placed into production.
- Monitor and optimize each agent on an ongoing basis.
Expected Outcomes
Lower Retell cost per minute.
Lower telephony costs.
Faster, more natural conversations.
Accurate collection of names, phone numbers, email addresses, and messages.
Fewer patient and client complaints.
More accurate delivery of dispositions and client messages.
Faster deployment of new agents.
Reduced reliance on CCC's internal employees to learn and manage Retell.
Greater long-term client retention
Stabilizing MRR.
Return On Investment
The business case for improving CCC's Retell operations is based on four primary areas:
1. Lower AI Usage Costs
The audit found individual Retell agents costing between approximately:
- $0.20 per minute at the lower end.
- $0.35 per minute at the higher end.
These higher-cost agents are affected by issues including:
- Prompts that are substantially longer than required.
- Knowledge bases being used for simple use cases that do not require them.
- High-cost or unsuitable LLM models.
- Single-prompt architectures being used for workflows that would be more efficient as conversation-flow agents.
- Instructions being loaded into every call even when they are only relevant to a small percentage of callers.
Depending on the current configuration of each agent, optimization may reduce the AI cost per minute by approximately 10–50% without reducing call quality.
The exact savings will be measured on an agent-by-agent basis after optimization.
2. Lower Telephony Costs
CCC is currently using Twilio numbers purchased through Retell at approximately:
- $2 per number per month.
- $0.015 per telephony minute.
The proposed Telnyx configuration is approximately:
- $1 per number per month.
- $0.0032 per telephony minute, before any additional volume-based pricing.
This represents:
- 50% savings on monthly phone-number fees.
- Approximately 78.7% lower telephony cost per minute.
For every 10,000 telephony minutes:
- Current estimated cost: $150.
- Telnyx estimated cost: $32.
- Estimated saving: $118 per 10,000 minutes.
These savings increase as CCC deploys Retell AI to more clients and call volume grows.
3. Reduced Internal Labour
CCC currently has multiple employees attempting to learn, configure, troubleshoot, and maintain Retell.
This creates internal costs while still producing inconsistent outcomes.
Moving Retell ownership to Agento AI will:
- Reduce internal development and troubleshooting time.
- Remove the need for multiple employees to learn advanced Retell configuration.
- Reduce time spent manually reviewing poorly performing agents.
- Allow Parker and CCC's management team to focus on client growth, operations, and business development.
4. Improved Client Retention
CCC is handling calls on behalf of its clients.
When an AI agent responds slowly, sounds unnatural, repeatedly asks for the same information, fails to capture a name or email, routes a caller incorrectly, delivers an incomplete message, or creates a frustrating patient experience — the impact is not limited to one poor call.
It can result in:
- Complaints from the client.
- Patients abandoning the practice.
- Loss of confidence in CCC.
- Cancellation of CCC's services.
- Lost monthly recurring revenue.
Improved call quality protects CCC's client relationships and helps CCC retain each client for longer.
AGENTO TESTIMONIALS
Advocate Rights Center,
Dallas, Texas.
LEGAL INDUSTRY
ROBERT WILSON - FOUNDER
It’s been a great pleasure working with Agento AI.
They are very knowledgeable and experienced in AI-generated communications.
We consider Agento AI to be a valued & trusted partner.
⭐⭐⭐⭐⭐
Norton Insurance,
United Kingdom.
INSURANCE INDUSTRY
BEN WHITE - HEAD OF OPERATIONS
It's been a pleasure working with Agento AI. We had already started looking into how AI can improve productivity & performance and their solution certainly did both of those things.
We have been able to integrate their voice agent in to our lead chasing which has had a huge positive effect. Getting our prospects to answer the phone was our biggest issue (even when they request call backs on our website).
After introducing the voice sales agent, our contact rate has improved from 40% to over 60% - which has exceeded our hopes and targets!
⭐⭐⭐⭐⭐
Blip Payments,
Ireland.
FINANCIAL SERVICES INDUSTRY
HUGO URIEN - FOUNDER
Agento AI developed several solutions for us that have helped our business become more systematic and efficient.
They take the time to really understand the goal and offer solutions to get there.
We plan to continue to work with Agento AI and would recommend them to anyone who is looking to optimize systems in their business.
⭐⭐⭐⭐⭐
Preferred Fertility Concierge,
Atlanta, Georgia.
HEALTHCARE INDUSTRY
TAKERA MITCHELL - FOUNDER
Agento AI exceeded my every expectation and went above and beyond to come up with a solution for my somewhat complex request.
Not only was the service thorough and timely but the customer service was superb.
Agento ensured my needs were understood and followed up with me consistently throughout the entire process.
I highly recommend them for any of your automation needs!
⭐⭐⭐⭐⭐
1) Current Situation
CCC supports approximately 300 clients across dental, healthcare, and other business categories.
CCC operates 24/7 and primarily provides two types of call-handling service:
Message-Only Clients
For messaging-only clients, CCC:
- Answers the caller.
- Identifies why they are calling.
- Collects the relevant contact information and message.
- Delivers the disposition or message to the client's office.
Scheduling Clients
For scheduling clients, CCC:
- Determines whether the caller is a new or existing patient.
- Collects the relevant intake information.
- Transfers the caller to a CCC human agent.
- The human agent remotely accesses the client's practice management system and books the appointment.
Retell is currently being used as the initial AI layer for a growing number of CCC clients.
Early outreach to clients has shown encouraging adoption, and CCC expects to introduce Retell AI to a larger portion of its client base.
However, the Retell operation is not currently structured to support this growth reliably.
2) Problems Identified During the Retell Audit
The current Retell account contains a number of operational and technical issues that need to be addressed before CCC scales further.
Excessively Long Prompts
Many prompts contain significantly more information than is required for the calls, which increases costs.
Unnecessary Knowledge Base Usage
Knowledge bases are being used on some agents where they are not required. This adds cost without providing meaningful value.
Unsuitable LLM Selection
Some agents are using models that are unnecessarily expensive or slow for basic call-handling use cases.
High Latency
CCC has received complaints about delays during conversations. High latency makes the agent feel less natural to callers.
Unnatural Voice Experience
Agent voices have not been optimized for natural delivery. The agents sound overly polished, robotic, or inconsistent.
Poor Data Collection Experience
Callers have experienced difficulty when providing names, email addresses, phone numbers, and other intake details.
Limited Date & Time Awareness
Some agents are not properly aware of the current date, time, day of the week, client business hours, or client timezone. This is resulting in incorrect or confusing answers.
No Reliable Performance Alerts
There is currently no system that automatically identifies when an agent behaves poorly outside the intended workflow. Without alerts, CCC may only discover a problem after a client complains.
No Standardized Performance Tracking
CCC does not currently have a reliable Retell-wide process for tracking call performance outcomes.
Unclear Agent Inventory
The account contains multiple agents with different statuses and configurations. In some cases, it is unclear which agents are live, which are test versions, which are inactive, which phone number belongs to which agent, etc.
Deprecated API Usage
Deprecated API usage was identified in existing email workflows. Deprecated integrations create an avoidable risk that message-delivery workflows may stop functioning when the underlying service is retired.
3) Business Impact
The current issues create three direct risks for CCC.
Rising Costs
Poor prompt architecture, unnecessary knowledge bases, unsuitable LLMs, and expensive telephony increase the cost of delivering the AI service.
Client Churn
A client may tolerate one poor call, but repeated latency, caller frustration, missing information, and failed transfers can lead to complaints and cancellation.
For CCC, a poorly performing AI agent can directly reduce monthly recurring revenue.
Operational Instability
Without monitoring, documentation, and ownership, every additional Retell client increases operational complexity.
Scaling the current setup without correcting the underlying problems would amplify those issues across more client accounts.
4) Scope of Work
Agento AI will manage the Retell side of CCC's AI call-handling operations.
Included within the scope:
- Existing Retell agent audits.
- Existing agent optimization or rebuilding.
- New Retell agent development.
- Prompt and conversation-flow architecture.
- Voice and conversational setting optimization.
- LLM selection.
- Knowledge base review.
- Data collection optimization.
- Date/time and business-hours logic.
- Transfer configuration.
- Webhook and disposition workflow review.
- Internal performance scoring.
- Poor-call alerts.
- Manual call review during ongoing management.
- Retell account organization.
- Telephony migration support.
- Ongoing support and maintenance.
Not included in the current scope:
- Client-facing performance portals or dashboards.
- Detailed monthly reports that CCC can share directly with its clients.
- Full direct appointment-booking integrations with every practice management system.
- Development outside the Retell and associated voice-automation workflows.
- Major custom integrations not identified during the individual client scoping process.
Client-facing reporting and direct calendar booking can be added later as separate projects.
5) Messaging-Only Agent Flow
Each messaging-only agent will be customized for:
- The client's business name.
- Office details.
- Business hours.
- Required intake fields.
- Frequently asked questions.
- Emergency or urgent-call rules.
- Message delivery email.
- Escalation requirements.
- Post-call disposition categories.
Answer using the client's approved greeting.
Identify why the caller is calling.
Capture the caller's name.
Capture and validate the caller's phone number.
Capture other client-specific required fields.
Collect a clear message.
Confirm the message back to the caller.
Apply urgency and intent categories.
Send the disposition to the correct office email or workflow.
Confirm what will happen next.
End the call professionally.
6) Intake & Scheduling Agent Flow
For intake and scheduling clients, the Retell agent will handle the first part of the call before transferring the caller to CCC's human scheduling team.
New or Existing Patient Identification
The agent will ask whether the caller:
- Is a new patient.
- Has previously visited the practice.
- Is calling on behalf of someone else.
Existing Patient Intake
For an existing patient, the agent may collect:
- Full name.
- Phone number.
- Reason for calling.
- Urgency or appointment type.
New Patient Intake
For a new patient, the agent may collect:
- Full name.
- Phone number.
- Street address.
- Email address.
- Reason for appointment.
- Additional client-specific intake fields.
Transfer to CCC Scheduling Team
Once the required intake information has been collected, the AI will:
- Explain that it is connecting the caller to the scheduling team.
- Transfer the caller to the correct CCC team or queue.
- Pass the captured information to the human agent where technically supported.
- Use a fallback path if the transfer fails.
- Log the outcome after the call.
This hybrid model reduces the time CCC's human staff spend completing basic intake while retaining human scheduling for practice-management systems that cannot yet be integrated directly.
7) Existing Agent Optimization
Each existing agent selected by CCC will be individually reviewed.
The optimization process includes:
Agent Audit
- Review current prompts.
- Review agent architecture.
- Review call objectives.
- Review LLM selection.
- Review voice configuration.
- Review knowledge base usage.
- Review phone number and telephony configuration.
- Review post-call analysis.
- Review webhooks and disposition workflows.
- Review recent calls.
Rebuild & Optimization
Where the current foundation is usable, the agent will be optimized. Where the current structure is fundamentally unsuitable, the relevant components will be rebuilt.
The work may include:
- Rewriting the global prompt.
- Converting complex single-prompt agents into conversation-flow agents.
- Removing duplicated or irrelevant instructions.
- Separating rarely used functions into sub-agents.
- Reducing unnecessary token consumption.
- Updating the LLM.
- Updating the voice.
- Adjusting responsiveness and interruption sensitivity.
- Improving data capture.
- Adding current date/time awareness.
- Improving transfer behavior.
- Correcting integrations and webhooks.
- Replacing deprecated API calls.
- Adding automated performance analysis.
Testing
Internal test calls will be completed.
Common call scenarios will be tested.
Caller interruption will be tested.
Data capture will be tested.
Transfer and fallback behavior will be tested.
Webhook and disposition delivery will be tested.
The agent's cost per minute will be reviewed.
8) New Client Retell Agent Builds
Every new agent will be developed using a controlled onboarding process.
Step 1 - Client Requirements
CCC or the client will provide: business name, agent type, greeting, business hours, call-handling objectives, required data fields, FAQs, transfer destination, message delivery destination, emergency rules, and client-specific restrictions.
Step 2 - Agent Architecture
Agento AI will determine: whether the use case should use a single-prompt or conversation-flow agent, whether sub-agents are required, the appropriate LLM, the appropriate voice, whether a knowledge base is required, and which post-call fields and integrations are required.
Step 3 - Build & Configuration
The agent will be built and customized for the individual client.
Step 4 - Testing
Agento AI will test the complete call flow before go-live.
Step 5 - Client Approval
CCC can provide the practice with a test number or sample recordings before activation.
Step 6 - Go-Live & Monitoring
Once approved, the agent will be placed into production and monitored during the initial optimization period.
9) Retell Account Governance
Agento AI will establish a more structured operating model for CCC's Retell account.
Agent Register
The account should maintain a clear record of:
- Client name.
- Agent name.
- Agent type.
- Agent ID.
- Assigned phone number.
- Telephony provider.
- Live/inactive/test status.
- Current cost per minute.
- Transfer destination.
- Disposition workflow.
Naming & Version Control
Agents will use a standardized naming structure to reduce confusion between:
- Live agents.
- Test agents.
- Archived agents.
- Duplicated templates.
- Previous versions.
Account Cleanup
Where approved by CCC:
- Unused agents can be archived.
- Old test phone numbers can be removed.
- Unused knowledge bases can be disconnected.
- Duplicate or obsolete versions can be documented.
- Deprecated integrations can be replaced.
10) Performance Monitoring & Ongoing Optimization
Although client-facing performance reporting is not included in the current proposal, internal performance monitoring is included within Agento AI's ongoing management.
Automated Post-Call Analysis
Each managed agent can automatically evaluate completed calls against its intended purpose.
Metrics may include:
- Overall call score.
- Whether required information was collected.
- Whether the correct disposition was reached.
- Whether the transfer was successful.
- Whether the caller became frustrated.
- Whether the caller abandoned the call.
- Whether the agent repeated questions unnecessarily.
- Whether the agent followed the correct escalation path.
Poor-Performance Alerts
If a call falls below an agreed score or triggers a serious failure condition, Agento AI can be alerted for review.
Manual Review
Automated scoring is useful but not perfect.
Agento AI will also manually review a sample of calls on a weekly or bi-weekly basis during ongoing management.
Continuous Optimization
Where trends or failures are identified, Agento AI will:
- Review the call.
- Determine the root cause.
- Optimize the relevant prompt, setting, or workflow.
- Test the change.
- Monitor subsequent calls.
This creates a continuous improvement cycle instead of waiting for a client complaint.
11) Telephony Cost Optimization
Agento AI will assist CCC with moving suitable Retell numbers from the current Twilio/Retell telephony arrangement to Telnyx SIP.
Current estimated pricing:
- $2 per phone number per month.
- $0.015 per telephony minute.
Telnyx estimated pricing:
- $1 per phone number per month.
- $0.0032 per telephony minute before volume pricing.
Lower monthly phone-number cost.
Lower telephony
cost per minute.
Estimated savings per 10,000 telephony minutes.
The migration will be completed carefully to avoid unnecessary interruption to active client numbers.
Final pricing remains subject to the selected carrier, number type, jurisdiction, and current carrier rates.
12) Implementation Plan
Phase 1 - Pilot Optimization
(3 Agents)
Agento AI will rebuild and optimize a small group of 3 existing agents. The pilot will confirm the optimization framework, compare cost per minute before and after, validate improved latency and data capture, confirm post-call monitoring, and establish CCC approval standards.
Phase 2 - Account Inventory & Prioritization
Confirm which agents are currently live, which require optimization, which are inactive or test agents, client priority order, and number ownership and telephony mapping.
Phase 3 - Existing Agent Rollout
Once the pilot is approved, CCC can select additional existing agents for rebuilding and optimization in order of priority.
Phase 4 - New Client Onboarding
New agents will be built as clients approve Retell AI deployment (1 business day turnaround).
Phase 5 - Managed Operations
Agento AI will provide ongoing monitoring, troubleshooting, quality control, and optimization for each managed client agent.
Full rollout timeline: Dependent on the number of active agents selected by CCC and the order in which CCC wants them completed.
13) Deliverables
Retell account and agent inventory.
Existing agent rebuilding.
New client agent builds.
Increased agent development standards.
Optimized prompts and conversation flows.
Correct LLM selection.
Optimized voices and call settings.
Remove unnecessary knowledge bases.
Date/time and business-hours configuration.
Webhook and disposition workflow review.
Handling Retell error emails recently received.
Automated internal post-call performance analysis.
Poor-performance alerting.
Manual call review under ongoing management.
Retell account naming and organization standards.
Telnyx migration support for suitable numbers.
Testing and go-live support.
Ongoing managed Retell operations.
14) Pricing & Terms
Notes On Final CCC Partnership Rate
Because CCC has decided to defer the client-facing performance-reporting portal, Agento AI will provide a reduced setup rate of:
$350
One-Time Per Agent:
This rate applies to either:
- Rebuilding of existing CCC Retell agents.
- Development of new Retell agents for each new CCC client.
The $350 one-time fee includes:
- Requirements review.
- Prompt or conversation-flow development.
- Client-specific customization.
- LLM selection.
- Voice and agent-setting configuration.
- Optimizing each agent for low costs and high performance.
- Integration and webhook review.
- Internal performance-monitoring configuration.
- Test calls and quality assurance.
- Go-live preparation.
Ongoing Managed Service Fee
For each active agent managed by Agento AI:
- $80 per month for the first 3 months following go-live.
- $30 per month after the first 3 months.
The first three months require more intensive:
- Call monitoring.
- Optimization.
- Bug fixing.
- Prompt adjustments.
- Workflow validation.
- Performance review.
After the initial optimization period, the lower monthly fee covers ongoing monitoring and maintenance within the existing agreed scope.
Similar Client Replacement Rate
If a CCC client leaves within the first 3 months, the existing agent is no longer being used, a new client has a materially similar business and call flow, and the existing agent can reasonably be repurposed - Agento AI can update and redeploy that agent for:
$150 one-time.
This is not a direct duplicate. Client-specific information, prompts, FAQs, emails, phone numbers, transfer destinations, tracking fields, and integrations will still need to be updated and tested.
Client-Facing Reporting
Client-facing dashboards, portals, and reports that CCC can share with each client are not included in the current $350 price. This can be added later as a separate optional service.
Major Scope Changes
Major new functions, custom software integrations, complex practice-management APIs, or materially different call flows will be scoped separately before work begins.
15) Usage & Third-Party Costs
Retell, LLM, voice, telephony, phone number, and other third-party usage charges are not included in the Agento AI development or monthly management fees.
These costs will remain payable by CCC directly to the applicable vendor.
Agento AI will work to reduce these costs by:
- Shortening and restructuring prompts.
- Selecting more appropriate LLMs.
- Removing unnecessary knowledge bases.
- Separating complex call paths where beneficial.
- Optimizing agent architecture.
- Migrating suitable phone numbers to lower-cost telephony.
Third-party rates are subject to change by the relevant providers.
16) Support & Service Levels
Agento AI Support Hours: Monday–Friday, 9am - 5pm Eastern Time via Zoom, Slack, or email.
Response Targets
P1 - Production agent unavailable or critical integration failure:
Response within 2 business hours.
P2 - Degraded performance, transfer issue, or message-delivery issue:
Same business day.
P3 - Minor prompt, wording, or configuration request:
Within 1-2 business days.
Ongoing Management Includes
- Retell system monitoring.
- Prompt and workflow maintenance.
- Poor-call alerts.
- Manual call reviews.
- Minor call-flow improvements.
- Bug fixes.
- Voice and setting adjustments.
- LLM updates where appropriate.
- Integration health checks.
- Support during Retell platform updates.
Major new features or integrations will be quoted separately.
17) Assumptions
CCC provides Agento AI with the required access.
CCC identifies which agents are live and prioritized where possible.
CCC provides client-specific call-handling requirements.
CCC provides access to relevant message-delivery and transfer information.
CCC approves each agent before production launch.
English-language agents are the initial priority unless otherwise agreed.
18) Why Agento AI
Top 5 of +200 Retell Solution Partners.
We are in the top 5 of +200 Retell Solution Partners.
Retell-specific development expertise.
Experienced in building systems that manage thousands of daily calls.
Custom-built +40 Retell agents.
Inbound receptionists, intake agents, transfers, scheduling flows, and support agents.
Deep knowledge of controlling costs.
Controlling LLM, voice, and telephony costs, while keeping performance stable.
Continuous optimization approach.
A continuous optimization approach based on real production conversations rather than test calls alone.
The partnership combines strategic process design, implementation, Retell development, automation, testing, and managed operations.
CCC does not need another vendor that delivers disconnected pieces and leaves the internal team to assemble them.
Next Steps
Review and approve the proposal and pricing structure.
Confirm the initial pilot of 3 existing agents CCC wants Agento AI to rebuild & optimize.
Provide access to the relevant Retell account, webhooks, email workflows, and telephony details.
Schedule a quick kickoff and agent-prioritization call.
Develop and deliver the first optimization of 3 existing agents in one week before expanding further.
CCC Retell AI Operations Optimization & Managed Agent Partnership - Proposal