Cloud is the future. Indeed, Gartner estimates that more than half of IT spending will shift to the cloud by 2025. Yet, that does not mean it is without its faults.
Unfortunately, many of these are cost-related. Businesses get hooked, attempt to “cloudify” everything, and – over time – costs spiral.
Researchers at Andreessen Horowitz delved deeper into this phenomenon. The good news is they found that cloud solutions typically deliver on their promise shortly after the transformation.
Yet, from there, the story often takes a turn for the worse. “The pressure it puts on margins can start to outweigh the benefits as a company scales and growth slows.”
As such, brands shouldn’t view cloud migrations as an all-or-nothing conundrum. After all, businesses must not sacrifice control of their costs.
Cloud In the Contact Center Is a Logical Step, But…
Contact centers deal with incredible spikes in contact volumes. As such, the scalability of the cloud makes migration a forgone conclusion for many.
With the cloud also delivering the possibility of easy integrations to the latest and greatest AI solutions, some may scoff at cloud cynicism.
Yet consider these statistics:
- Upwards of 80 percent of CIOs have yet to achieve the benefits of cloud migration. (Source: McKinsey & Company)
- More than 30 percent of cloud spend in 2021 was wasted or inefficient. (Source: Flexera)
Now, these findings are not contact center specific. Nonetheless, the service environment is not immune to the trend.
Indeed, many are overspending, according to Rebecca Wetteman, Principal at Valoir. In a recent conversation with CX Today, she stated:
“A lot of folks sign into contracts for several years for a fixed number of seats, but those seats aren’t being used. If I’m a CFO, I’m going to start sharpening my pencil a little bit there.”
Fellow industry expert Liz Miller, VP & Principal Analyst at Constellation Research, also noted this trend. “I’m involved in multiple renegotiations at this point for that very reason,” she said.
Of course, this pinpoints the need for better forecasting. However, businesses must realize that CCaaS is more expensive in the long run. There needs to be value to justify that ongoing cost.
To help businesses do that, James Cox, SVP of Marketing at LumenVox, shares the following four best practices to optimize cloud speech processing costs.
1. Don’t Give Up Your Cloud Autonomy
Vendor-locking – where a contact center gets stuck with using one cloud for everything – is one of the more disappointing trends within the CCaaS space.
It involves exclusively signing up for one of the three big cloud players – AWS, Google, and Microsoft. After, the contact center can only access the technologies within one of those clouds for CCaaS, Workforce Optimization (WFO), speech and voice recognition, and more.
As a result, brands often end up paying a premium price for the package rather than shopping around to get the best deal a la carte.
Instead, they are locked in without ready access to APIs and composable components that allow the contact center to easily compose a stack that fits neatly together.
Noting this, Cox adds:
“To get the best of each cloud provider without the headlock risk of sacrificing cost efficiencies, CX components must be cloud agnostic to avoid vendor lock-in.”
On this cost-efficiency front, “locked-in” contact centers can’t readily switch providers if the cloud giant it has aligned with switches up its rates or makes a choice that isn’t best for the business.
For example, once Microsoft acquired Nuance, it encouraged customers to commit to the Azure cloud in the hope of growing the platform.
Now, if its customers want to stay on AWS or Google, they may face difficulties.
Yet, avoiding such vendor lock-in is the right play. Doing so gives the contact center control of its own destiny while allowing it to compare prices and capabilities and find a vendor that best meets its needs at an appropriate price point.
2. Right-Size Cloud Services
When moving to the cloud, many businesses attempt to recreate their legacy systems and look for all the features they have on-premise.
Yet, does the contact center use all of those? It’s unlikely, and by streamlining the cloud solution, there is a significant chance for lowering costs and simplifying processes.
Such an initiative is important, as a CCaaS vendor will charge for idle resources.
Also, think about what the business may want to keep on-premise. After all, it is not tricky to wrap an API around a legacy system and gain interoperability with a cloud platform these days.
Businesses can then work with vendors to create a hybrid, custom agreement, instead of settling for a formulaic pay-as-you-go model – where costs often mount quickly.
Building on this, Cox says:
“Repatriate services to improve performance, control, and cost efficiency as projects evolve and scale. Doing this and leveraging existing on-premises investments is where the cost-benefit is often not as expected.”
Reviewing prices and billing information will help spot the high-cost areas and allow contact centers to determine where they can generate the most saving with this tactic.
3. Target a Composable Service
Composable customer experience combines traditional development, packaged apps, APIs, and no-/low-code. Many consider it the future CX.
By embracing composable CX, businesses may work with cloud-native speech recognition and transcription vendors.
As such, they may benefit from API access, microservices, container orchestrators, auto-scaling, and auto-healing to design the ideal voice infrastructure – with cost and capability benefits.
However, as Cox tells us:
“Crucially, such an architecture deployable anywhere to deliver a truly agile infrastructure and avoid the costs and risks of a single vendor lock-in.”
Opus Research recently presented an excellent webinar that goes into much more detail on how to achieve this. Check it out here.
4. Select Regions Carefully
Storing voice data in the cloud comes with cost risks regarding recording law infringements, such as GDPR. After all, such regulations deem voice to be Personally Identifiable Information (PII).
Other compliance laws take a similar perspective, such as PIPEDA in Canada.
“Even if users do not share their names or credit card information in recordings, voice recordings are still PII,” adds Cox.
As such, businesses must select regions carefully for better management of regulatory and compliance risks – which may result in massive fines.
Indeed, European data regulators issued a record €1.65BN in fines in 2022. That is a 50 percent increase on the previous year, highlighting how these forces are tightening the screws.
Download This Report to Dig Deeper
Just as Big Tech is now reducing its workforce to optimize costs after the pandemic, why wouldn’t you do the same with your cloud workloads?
After all, as Cox concludes: “Continuing to write blank cheques to the cloud providers for speech recognition and transcription is not sustainable.”
“The more firms embrace cloud-computing, the more carefully they must manage their costs to escape the Big Tech headlock and avoid overpaying.”
Delve deeper into how your business can do precisely that by downloading the Speech Recognition Buyer’s Guide
The report walks you through the pros and cons of using big tech for speech recognition – which is an ideal place to start looking for serious savings.
