AI is rapidly redefining customer support, making service faster, smarter, and available 24/7. At Unless, we believe that the future of customer experience is built on intelligent automation that’s easy to implement and delivers measurable value. But how do you truly measure the return on investment (ROI) of AI in customer success and support? Let’s break it down, covering the most important metrics, supporting indicators, and the cost elements you need to consider when using a SaaS provider like Unless.
We'll top it off by showing you a hypothetical example for ROI calculation, based on our experience with real-life clients. Once you are ready to calculate your own potential savings, you can either make a copy of this spreadsheet or contact us to help you!
Measuring success: primary metrics
To understand the positive impact of AI on your support operations, start by tracking these primary metrics, many of which you will find in your Unless dashboard already or very soon:
Deflection Rate The percentage of customer conversations handled by AI that never reach a human agent. High deflection means your AI is effectively reducing the workload for your team and speeding up responses.
AI-Handled Rate The share of conversations the AI manages meaningfully, without any escalation attempt in it, and with its last question being understood (so that the AI did not use any of its defaults there). This metric shows how often your AI is not just responding, but actually helping.
Automated Resolution Rate An automated resolution is counted whenever a conversation is handled by the AI agent and also successfully passes our verification mechanisms. This is a strong indicator of your AI’s real-world impact.
Out-of-Hours Support The number or percentage of customer conversations handled by AI outside regular business hours. This highlights the value your AI brings when your human team is offline.
Customer Satisfaction (CSAT) and Effort Scores (CES) Scores from customer surveys after AI interactions, measuring satisfaction and how easy it was to get help. These scores reflect the quality of the customer experience.
First Contact Resolution (FCR) The percentage of issues resolved by AI in the first interaction. High FCR means faster solutions and happier customers.
Average Resolution Time (support team) Assuming that you use the Unless Team Assistant as well, the ticket handling time equals the average time it takes for your support team members to fully solve a customer’s problem and close the ticket. Lower handling times signal greater efficiency and cost savings.
Measuring success: secondary metrics
Beyond the core metrics, these secondary indicators offer additional insights:
- Average Handling Time (AHT AI): The average time AI spends on each issue (not to be confused with human ticket handling time).
- Average Resolution Time (ART AI): How quickly the AI resolves customer issues .
- Escalation Rate: How often the AI needs to hand off conversations to human agents.
- Total Number of Sessions: The volume of support requests handled by AI, indicating engagement and demand.
- Resource Engagement Metrics: How often customers use self-service resources like help articles or FAQs.
Tracking these helps you fine-tune your AI and maximize its value.
The cost side: using a SaaS AI solution
When you choose a SaaS provider like Unless, your cost structure is clear and predictable. Here’s how it breaks down:
What Unless covers
- Software and Platform: We build, maintain, and update the AI platform, so you always have access to the latest features and models.
- Infrastructure and Hosting: All servers, storage, and backend infrastructure are managed for you.
- Security and Compliance: Your data is protected and meets European and global standards, like GDPR and the EU AI Act.
- Maintenance and Upgrades: Regular improvements, bug fixes, and new features are included.
- Standard Integrations: Easy connections to popular tools, CRMs, and helpdesks.
- Additional Custom Integrations: Depending on your subscription plan, extra connections may be covered by us.
- Data cleaning and restructuring: Upon ingestion, our system will strip the data of PII and restructure the content for readability.
- Knowledge gap detection: Based on unresolved customer inquiries and incoming support tickets, our system will point out where the gaps are in your knowledge base.
- Content Suggestions: The system will suggest updates for your content, based on ongoing learnings.
- Staff Training: Teaching your team to use and manage the platform.
- Technical Support: Our team is here if you run into platform issues.
What you manage
- Subscription or Usage Fees: Regular payments for access, typically monthly or annually.
- Initial Setup and Customization: Time spent configuring workflows, branding, and importing your knowledge base.
- Additional Custom Integrations: If your subscription plan doesn't cover these, there may be extra costs if you need to connect the AI to unique or internal systems.
- Data Selection: Gathering your help content and FAQs for the AI to use, and connecting the right sources.
- Ongoing Oversight: Monitoring AI performance and handling escalations.
- Content Updates: Keeping your knowledge base and responses up to date.
Calculating ROI: a step-by-step guide
To calculate ROI for AI in customer support, focus on ticket deflection (for anonymous visitors and logged-in users) and handling time improvements from internal co-pilots. Here’s how, in a hypothetical example based on what we see in our daily business:
1. Ticket deflection for prospects
What it measures: The percentage of website visitors who resolve issues via AI (e.g., chatbots) without submitting a ticket.
Formula:
Deflection Rate = (Tickets deflected / Total visitor inquiries) x 100
ROI impact example:
- Assuming 15,000 monthly visitors generating 1,000 tickets, 40% deflection rate → 400 tickets deflected.
- If each ticket costs €10 to handle (human agent time), monthly savings = 400 × €10 = €4,000.
2. Ticket deflection for existing clients
What it measures: Deflection for logged-in users in your product (e.g., via embedded AI agents in dashboards). Typically, the deflection rate for existing clients differs from the one for prospects.
Formula:
Client Deflection Rate = (Client tickets deflected / Total client inquiries) x 100
ROI impact example:
- Example: 2,000 in-product inquiries/month, 20% deflection → 400 tickets deflected.
- Savings = 400 × €10 = €4,000/month.
3. Handling time with internal co-pilots
What it measures: Time saved when human agents use AI co-pilots to resolve tickets faster.
Formula:
Time saved per ticket = old AHT - new AHT
Monthly savings = tickets handled x time saved x hourly rate for agents
Example:
- Old Average Handling Time (AHT): 10 minutes
- New AHT with AI co-pilot: 6 minutes
- Time saved/ticket: 4 minutes (0.067 hours)
- 2,200 tickets/month, €25/hour agent cost → Savings = 2,200 × 0.067 × €25 = €3,667/month.
4. ROI calculation in the first year
Total monthly savings in year 1
| Metric | Savings |
|---|---|
| Prospect ticket deflection | €4,000 |
| Existing client ticket deflection | €4,000 |
| Handling Time Savings | €3,667 |
| Total | €11,667 |
Costs (using Unless SaaS):
- Subscription fee: €1,699/month
- Setup: €2,000 (for example for an extra integration)
- About 2 days a week from one of your team members: 64 x €25 = €1,600/month
Annual ROI formula for year 1:
ROI = ((annual savings - annual costs) / annual costs) x 100
= ((€11,667 x 12 - (€3,699 x 12 + €2,000)) / (€3,699 x 12 + €2,000)) x 100
= ((€140,000 - €41,588) / €41,588) x 100 = 237% ROI
5. ROI in the years after
While our clients often experience significant benefits within just a few weeks, the real value of AI assistants becomes even more apparent over time. The key advantage is scalability: as your business grows, your AI-powered support grows with you - without the need for additional setup. Imagine your company grows its revenue with 50% the following year. If customer inquiries increase at the same rate, the benefits of AI scale right alongside your growth.
Even better, after the initial setup, there’s no need to repeat the process each year. Ongoing monitoring and optimization typically require only about half of the effort compared to the first year, allowing your team to focus on higher-value tasks while the AI continues to deliver consistent results.
Total Monthly Savings, year 2
| Metric | Savings |
|---|---|
| Total ticket deflection | €27,000 |
| Handling Time Savings | €4,500 |
| Total | €31,500 |
Annual ROI Formula for year 2:
= ((€378,000 - €31,988) / €31,988) x 100 = 1082% ROI
So, that's over 1000% ROI. Extrapolating on that, the third year will leave you even with 2045% ROI.
6. And there is more
There are also valuable side effects of implementing AI-powered support that, while harder to quantify, can have a significant impact:
- Streamlined Documentation Maintenance: Features like article rewriting, knowledge gap detection, missing article suggestions, and automated content updates greatly reduce the ongoing effort required to keep your documentation accurate and up-to-date. This means your support resources remain relevant with far less manual intervention.
- Easier Onboarding and Knowledge Retention: When your team’s expertise is captured and organized within our system, training new team members becomes much simpler and faster. Additionally, if experienced staff leave, their knowledge isn’t lost-it’s securely stored in a semantic database, preserving valuable insights for the future.
- Continuous Improvement: By leveraging AI-driven analytics, you can identify emerging topics and customer pain points, allowing you to proactively address issues and enhance your support content over time.
These benefits not only save time and resources, but also help ensure a consistently high-quality support experience for both your customers and your team.
Ready to see your AI ROI?
With Unless as your SaaS partner:
- No infrastructure costs: We handle hosting, security, and compliance.
- Rapid deployment: Start deflecting tickets in days, not months.
- Continuous updates: New AI features roll out automatically.
Unless makes it easy to track deflection rates, handling times, and cost savings in real time. Book a demo to see how our AI agents and co-pilots can transform your support ROI. Alternatively, use this simple spreadsheet to start playing with the numbers yourself!
Your customers get instant help-your team gets a productivity boost. Win-win.
