- Published on
Voice: Part 3 - Business Matters
- Authors
- Name
- Strategic Machines
Reflection Point
We’ve covered in our prior posts the Design (how to make it work) and the Technology (how to make it fast). Now, we arrive at the only metric that truly counts for the C-Suite: The Business Case.
Let's just take one business function and consider the impact of Voice AI. For decades, Customer Service has been viewed as a cost center—a 'cost of doing business' where the goal was to address "issues and inqueries" while optimizing other metics that are rigidly tracked. Voice AI flips this paradigm on its head. It turns support into a scalable, revenue-generating engine.
Consider this set of legacy metrics for a traditional call center staffed with live agents answering phones or responding to chat messages over the web. No surprise: the impact is on metrics is dramatic by staffing some of the seats with Voice Agents:
| Metric | Description | How It Supports Cost Reduction | AI Voice Agent Impact Insights |
|---|---|---|---|
| Average Handle Time (AHT) | Total time per call, including talk time, hold time, and after-call work (e.g., notes or follow-ups). Typically calculated as (Talk Time + Hold Time + ACW) / Total Calls. | Reducing AHT allows agents to handle more calls per shift, lowering labor costs per interaction and overall operational expenses. | AI voice agents significantly reduce AHT (often by 50-70%) through instant processing, no human delays, and automated resolutions for routine queries, enabling 24/7 operation without fatigue. But more important for this metric, businesses scale infinitely on the number of agents handling calls concurrently, each with a lower unit cost than a live agent. |
| Talk Time | The duration of actual conversation between the agent and customer, excluding holds or post-call tasks. | Shorter talk times boost throughput (more calls per hour), directly cutting costs by maximizing agent utilization without extending shifts. | AI shortens talk time by providing concise, accurate responses via natural language processing, handling complex queries faster than humans, potentially dropping it to under 2 minutes per call. But more important for this metric, the objective may change to engaging customers in more extended (even clever) interactions, to increase loyalty or sales. |
| Hold Time | Time the customer spends on hold during the call (e.g., while the agent researches). | Minimizing hold time improves efficiency and reduces total call duration, preventing agent idle time and lowering per-call costs. | Virtually eliminates hold time (near 0 seconds) as AI accesses data in real-time via integrations, avoiding manual lookups and improving customer satisfaction. But the concept of 'hold time' disappears, where an AI agent can continue an interaction while background processes are used to search databases or explore 'next best actions'. |
| After-Call Work (ACW) | Time spent after the call ends on tasks like documentation or escalation. | Streamlining ACW frees agents faster for the next call, reducing downtime and optimizing payroll expenses. | Automates ACW entirely—AI logs interactions, generates summaries, and escalates via APIs, reducing it to milliseconds and freeing resources for high-value tasks. |
| Occupancy Rate | Percentage of an agent's logged-in time spent on call-related activities (e.g., (Talk Time + Hold Time + ACW) / Total Available Time). Aim for 75-85%. | Higher occupancy means less idle time, improving agent productivity and reducing the need for additional hires to handle volume. | Approaches 100% for AI systems, as they handle unlimited concurrent calls without breaks, dramatically boosting scalability and reducing staffing needs by 80% or more in hybrid setups. |
| Average Speed of Answer (ASA) | Average time a customer waits in queue before an agent answers. | Faster ASA reduces queue buildup and total interaction time, lowering abandonment rates and associated retry costs. | Reduces ASA to near-instant (under 1 second), eliminating queues through infinite parallelism, which cuts abandonment by up to 90% and enhances efficiency in peak times. |
| Calls Answered Per Hour | Number of calls an agent handles in an hour. | Increasing this metric through shorter call times enhances efficiency, spreading fixed costs (e.g., salaries) over more interactions. | Skyrockets this metric (e.g., thousands per "agent" equivalent), as AI scales horizontally without human limits, amortizing infrastructure costs over massive volumes. |
| Cost Per Call (CPC) | Total operational costs divided by the number of calls handled. | Directly ties time metrics to expenses; reducing time per call lowers CPC by amortizing overhead across higher volume. | Slashes CPC by 70-90% via low marginal costs (e.g., cloud compute vs. salaries), with AI handling routine calls and escalating only complex ones to humans. |
This is only a single use case among dozens of functions that Companies should consider in view of the changing landscape with Voice Agents.
The Linear vs. Exponential Trap
Traditional contact centers suffer from linear scaling. If you want to handle double the call volume, you need double the humans. This is expensive, slow to hire, and hard to train.
Voice AI offers infinite elasticity. As PolyAI demonstrated with their "Raven v3" release, a single AI agent can handle thousands of concurrent conversations. Poly is configured in English, but responds in 24+ languages, from Arabic and Mandarin to Japanese and Hindi. Here are a few other results from v3 testing that caught out attention:
- 75% of calls handled autonomously.
- Zero wait times, meaning no call abandonment.
- $7.2M in new revenue generated from high-value calls that humans didn't have time to take.
These types of solutions are not just part of a business transformation project, they accelerate transformation. Consider the potential across a range of critical functions from patient scheduling, client intake, RFQs, order entry, and warranty management. The key is organizing firm policies, process and systems of record to be consumable by the agent - but more on context engineering in a future post.
Time-Based Competition
A few years back we posted our thoughts on the landmark study by the Boston Consulting Group on time-based competition. It's a powerful thesis: "2x time performance over the competition drives 4X financial performance results". We believe the new inflection point with Voice requires reflection among executives on where time-base competitive advantage could be gained by releasing Voice Agents in the business.
Voice is the fastest way to input complex data (3x faster than typing) and the fastest way to resolve issues. By deploying voice agents, you are respecting your customer's time. A "High CSAT" (Customer Satisfaction Score) is no longer just about being polite; it’s about being immediate, and accurate.
Consider the "Flywheel Effect." As the cost of inference drops and the quality of models rises, the companies that adopt Voice now will have a dataset and a customer behavior model that competitors cannot replicate.
From Cost Center to Experience Engine
The ROI is clear.
- Availability: Your business is open 24/7. You don't lose the East Coast customer because your West Coast team went home at 5 PM.
- Consistency: An AI agent never has a bad day, never forgets a compliance script, and never sounds tired.
- Value-Based Pricing: We are seeing a shift in the market where pricing is tied to outcomes (bookings made, issues resolved) rather than just minutes used.
Next Best Action
Voice agents have surpassed the competency threshold. They are no longer "cute demos." They are enterprise-grade employees ready to be hired.
At Strategic Machines, we are actively deploying these agents for reservation desks, sales concierge, and customer support. We invite you to try test agents we deployed. (All you need to do is request a 'one time password' and then select the 'people icon' at the bottom of iPhone to activate an agent. With the Cypress Agents, for example, you can plan a menu with a private chef, find a villa, inquire about amenities, and even reserve a room.)
We are combining the best design, the fastest tech, and the strongest business logic to build the future of conversational commerce.
Give us a call and turn your reflection point into a connection point for your business. We'd welcome the chance to share our insights on navigating this shift, and the remarkable adventure before us in 2026!