Customer Support and Voice AI
AI receptionist for small business: calls, bookings and handoffs
What an AI receptionist does on a real call, the rules that apply, a sample greeting and a pass/fail checklist to launch one safely.
Health & Risk
AI Customer Risk Agent
Accounts showing churn signals flagged with the reason why, and a recommended save action attached.
A red flag without a reason is not much help. This agent watches usage, support and billing signals for every account, flags the ones showing real churn signals, and attaches the reason and a recommended save action to each flag. Your CSMs know which account to call, why it is at risk and what to try first, and they review every action before customers see it.
- At-risk accounts flagged with the reason behind each flag
- A recommended save action attached to every alert
- CSMs review every action before it reaches a customer
$2,900
20–30 working days
Optional maintenance $249/mo
What it is
AI Customer Risk Agent watches usage, support and billing signals for every account, flags the ones showing real churn signals, and attaches the reason and a save action to each flag. It reports to the one channel you choose and alerts you when a pattern crosses your threshold. Your CSMs approve every save action.
Why businesses buy it
Most churn warnings arrive too late or too vague to act on. A dashboard full of red accounts with no explanation leaves your team guessing, and the guessing takes time you do not have. This agent tells your CSMs why an account is at risk and suggests the play that fits, so they can act the same day. The written scope sets out which signals it reads, how thresholds are set and where alerts land before you order.
Problems this solves
Churn warnings arrive without any reason attached
AI Customer Risk Agent attaches the reason to every flag, drawn from usage, support and billing signals. CSMs can see why an account was raised before deciding what to do.
CSMs can see risk but not the next step
Each flag comes with a suggested save action based on your rules. Your CSMs approve every save action, so the decision stays with the person who knows the customer.
Too many false alarms teach teams to ignore risk flags
Flags fire only when a pattern crosses the threshold you set. The logic is checked against three months of your history, so alerts reflect real churn signals in your data.
A good fit if
You know the warning signs of churn in your business, such as falling usage, a support spike or a failed payment, but nobody is watching for them across every account at once.
You want each flagged account to come with a reason, not just a red light, so your team knows what to say on the save call.
You want a recurring report or live dashboard showing at-risk accounts, sent to email, Slack or the web, with alerts when a pattern crosses your threshold.
You can provide read-only access to your product usage, support and billing data, your definitions of a churn signal, and at least three months of historical data.
Not the right fit if
You want retention plays run automatically against the accounts this flags, not just a flag and a recommendation. See AI Churn Prevention, which predicts churn risk and runs targeted retention workflows.
You want a single health number for every account rather than specific risk flags with reasons. See AI Customer Health Score, which calculates account health from usage, support and billing signals.
You need an agent that also monitors the whole account lifecycle and prepares CSM actions beyond risk. See AI Customer Success Agent, which monitors accounts and runs success playbooks.
You need more than three read-only data sources, more than twenty-five defined signals, or a live write-back into your CRM. Add a note at checkout for a Change Quote, or use our Custom AI enquiry form for larger projects.
How it works
- We connect your signals. HubSpot, Salesforce, Zendesk, Stripe, Mixpanel or Amplitude, using read-only access so nothing in your live systems is touched.
- We define what counts as risk with you. Falling usage, a support spike, a failed payment or a stalled renewal conversation, up to twenty-five signals mapped to your own experience.
- We check the flags against your history. Three months of your own data run through the model before launch, so accounts you already know were at risk get flagged correctly.
- The dashboard or recurring report goes live, delivered to the one channel you chose, showing every flagged account with the reason attached.
- Alerts fire when a pattern crosses your threshold. Up to five thresholds you define, so a fast-forming risk reaches you before it shows up in a quarterly review.
- You review the flags and we hand it off, with two rounds of changes and a recorded sixty-minute session for your team.
What you provide within 72 hours
Your portal lists these the moment you order. Delivery starts the working day after everything is in and your access is checked.
- Read-only access to your product usage, support and billing data sources.
- Your own definitions of what counts as a churn signal in your business.
- At least three months of historical data for us to check the flags against.
- The report's recipients and the delivery channel you want to use.
- Admin access and API keys for every connected system, shared by adding our team as a user or creating a key you control, never by email.
- Your business information (company basics, hours and brand voice), shared during intake.
- One decision-maker who can approve the risk definitions and sign off on the report.
- An API key for your own AI model account (for example OpenAI or Google Gemini), created for this build with a spending limit.
The 72-hour rule. Send everything on your checklist, including access, within 72 hours of ordering so your build starts on time.
Timeline
Start
Intake verified, risk signal definitions agreed.
Weeks 1–2
Data connections built, risk model drafted.
Week 3
Dashboard populated with your live flagged accounts.
Weeks 3–4
Flags checked against three months of your historical data.
Week 4
Your review and two rounds of changes.
Go-live
Live, recorded handoff session, support starts.
| Option | Delivery time | Price |
|---|---|---|
| Standard | 20–30 working days | $2,900 |
| Express | 15–25 working days | $3,400 |
| Rush | 10–20 working days | $3,900 |
What you receive at handover
- AI Customer Risk Agent configured and deployed in your environment
- Workflow, prompt and guardrail configuration with your thresholds
- Read-only connections to up to three data sources
- One dashboard or recurring report in your chosen channel, plus up to five alerts
- A test report checked against three months of your historical data
- An admin guide, the recorded 60-minute handoff session and 30 days of post-launch support
Expected benefits and how to measure them
- Helps CSMs act on risk with a clear reason and a ready next step.
- Can reduce time spent second-guessing which accounts to worry about.
- Flagged accounts saved
- Track how many flagged accounts renew after a save action, over the first two renewal cycles.
- Flags that proved accurate
- Each quarter, compare flagged accounts with actual cancellations to see how often flags were right.
- Gross churn rate
- Compare cancellations in billing for the two quarters after launch with the same quarters last year.
Before you start: Record your current churn rate and list accounts lost last quarter. Note what warning signs, if any, were seen.
These are things to track, not promised results. Outcomes depend on your data, your team and your market.
Running costs
The build price is one-time. Ongoing costs sit in your own accounts once the agent is live: AI model usage plus a small hosting or scheduler fee, usually $20 to $150 a month regardless of account count.
AI usage during build and testing is billed the same way, to your accounts, so you own every provider relationship from day one. Nothing runs through us.
Ask about running costsCustom requirements
If more data sources, more risk signals, or a live write-back into your CRM turns out to be what you need, flag it at checkout. A written Change Quote follows, and nothing starts on it until you say yes. Your original order is unaffected and keeps to its normal timeline. Larger or bespoke projects go through the Custom AI request form instead.
Request a custom versionKeep it running
Monthly maintenance covers monitoring, fixes, updates, a health report and change hours: a new risk signal, a shifted threshold, a redesigned chart. The hours scale with your build. Maintenance begins once the thirty days of support are up, bills monthly, and can be cancelled whenever you choose.
$249/mo (the minimum fee, includes 2 change hours)
Optional. Starts after the post-launch support period. Cancel any month.
How maintenance worksHow this compares
| Do it yourself with a SaaS tool | Hire a custom agency | Upmark Labs | |
|---|---|---|---|
| Price | A CS platform's risk flags are bundled into a per-seat fee you already pay | A custom agency quotes after discovery calls | Ours is one fixed price, shown before you buy |
| Reasons | A platform often flags risk without explaining why | An agency agrees scope during the project | Ours attaches the reason and a recommended save action |
| Delivery | A platform's flags are available immediately but rarely trusted at first | An agency usually takes several weeks, set per project | Ours has a fixed delivery time shown before you buy, with faster options |
| Validation | A platform's flags are rarely checked against accounts you already know | An agency build depends on what was agreed | Ours is checked against three months of your real history |
| Alerts | A platform typically shows a list you have to remember to check | An agency build depends on what was agreed | Ours alerts on up to five thresholds you choose |
Questions about this product
How much does AI Customer Risk Agent cost?
Each delivery speed has one fixed price. The price panel shows standard, express and rush side by side before you pay, and the cart confirms your delivery date. The price is fixed once your risk definitions and data sources are agreed, with no per-account pricing to track. Running costs of roughly $20 to $150 a month are separate and paid to your own providers.
What does AI Customer Risk Agent cost to run each month?
Roughly $20 to $150 a month, covering AI model usage and the hosting or scheduler that keeps flags refreshing on schedule. You pay that directly to your own providers, in your own accounts. Monthly maintenance is optional if you want us managing changes.
How fast can AI Customer Risk Agent go live?
The clock starts once your intake is complete and verified. Your fixed delivery date for standard, express or rush is shown before you buy, and faster speeds keep the same scope, testing and handoff. Most of the time goes into agreeing what counts as a risk signal and checking it against three months of history.
What access does AI Customer Risk Agent need to my systems?
Read-only access to your product usage, support and billing data, shared by adding our team as a user or creating a key you control, never by email. Nothing in your live systems is changed. We also need at least three months of historical data so flags can be checked against accounts you already know were at risk.
Can I change what counts as a risk signal after launch?
Yes. Two rounds of changes are included before launch, based on the test report. After that, a new risk signal or a moved alert threshold fits inside the monthly change hours that come with maintenance. A new data source or extra signal is a separate add-on.
Does AI Customer Risk Agent cost anything once the 30 days of support end?
No. The 30 days of post-launch support are included in the build price with nothing extra to pay. After that, monthly maintenance is optional, not required, and you can cancel it in any month once you trust the flags without further tuning.
Who owns the AI Customer Risk Agent setup once it is built?
You do. It runs in your own CRM, helpdesk, billing and AI provider accounts, so nothing is tied to us to keep it working. If you ever move away from Upmark Labs, the risk model and every connection stay in your accounts, not ours, exactly as delivered.
Guides on this topic
All guides Know why an account is at risk, not just that it is.
Order AI Customer Risk Agent and at-risk accounts are flagged with a clear reason and a recommended save action, so your team knows exactly who to call and what to say.
Works well with
Health & Risk
AI Customer Health Score for Intercom
Calculates account health from usage, support, billing and sentiment signals, connected natively to Intercom so it reads and updates conversations, contacts, companies…
Health & Risk
AI Customer Health Score for Salesforce
Calculates account health from usage, support, billing and sentiment signals, connected natively to Salesforce so it reads and updates accounts, contacts, opportunities,…
Health & Risk
AI Customer Health Score for HubSpot
Calculates account health from usage, support, billing and sentiment signals, connected natively to HubSpot so it reads and updates contacts, companies, deals,…