Ready or not? How to prepare for AI that lasts beyond the pilot
Many contact centres aren’t ready for AI, and successful adoption is lagging as a result. Joe Bigio, senior vice president of CCaaS, CX, and CAI at CallTower, explains what separates those getting results from those stuck in pilot stage.
We’re fast moving into the AI era, with the promise of many efficiency and automation benefits for the companies that successfully adopt the new generative and agentic tools. And while these upsides are justifiably well documented, their widespread realisation has not kept pace. Earlier this year, Sam Altman, OpenAI CEO, admitted his surprise at how slow the economy has been in embracing the technology that he helped to build.
This is, of course, a striking admission. But in many ways, it is not a surprising one. Companies today weren’t designed to work with AI. Fundamentally they’re built around human processes and timeframes, not AI timescales and agentic processes. There is also a lot of inertia within most companies; making big changes takes a lot of time. What’s more, before you can make a transition to a new technology, you must be ready to change.
In communications and telecoms, we’ve undergone several significant changes over the years, from on-premises phones to VoIP to cloud-based UCaaS and CCaaS platforms. So, when it comes to the appearance and arrival of AI, we know where the similarities are, and appreciate the importance of being prepared. The real question facing businesses today isn’t whether to adopt AI, but how to get ready for it.
The barriers to becoming ready
Any transformation project comes with barriers and risks. And one we’ve commonly seen in organisations is decision paralysis. Some businesses approach AI as if it were a choice they can’t undo – something set in stone or permanent, like a tattoo. Instead, they should look at the technology more flexibly – something that can be tried, tested, and improved on with something better, like a hat or a haircut. Experimenting and refining means that what sticks is more likely to be the best-fit solution, and can speed up an organisation’s journey to successful adoption.
Trying on a few hats takes both nerve and budget, and experimentation must pay off eventually. Another recurring barrier we see is uncertainty about how AI will improve financial performance. So far, AI adoption hasn’t delivered the profound EBITDA changes most companies want to see. According to research from Boston Consulting Group, only 6% of companies today have experienced meaningful value from their AI endeavours. Similarly, data from Gartner shows that 44% of CEOs are struggling to show AI’s direct financial benefit to shareholders.
Naturally, an unclear ROI can be a barrier to properly preparing for and successfully adopting AI. However, the hidden smaller successes in these figures indicate a way to do things properly. From our experience with customers, several things separate those who get results from the rest.
Steps to readiness and adoption
Pick wisely and prioritise. Focus on a limited number of AI projects and do them well. Concentrating on a small number of functions within your organisation and preparing these for AI adoption is what will work best. In the contact centre, this could be focusing on using AI voice to create a self-service channel that replaces routine calls for agents. One US insurance company with millions of policyholders did exactly that, and was able to successfully handle over 70% of its routine calls within this channel.
Focus on and fund ‘day two’. Prepare for production from the start, and think about and budget ahead for the monitoring, tuning, and quality work that follows going live. Tools like agent assist and virtual agents are great examples of AI functionality that can deliver some faster results, but they also benefit from ongoing refinement and tweaking. As products, policies and customer queries change over time, the agility of these systems compounds those benefits too.
Redesign workflows. Bolting AI onto an existing process scales the ways it currently works – including the parts that are glitchy or don’t work at all. A better starting point is to look at the workflow first. In a contact centre, that might mean rethinking how a customer moves between virtual agents and humans, so that someone who texted yesterday – or even moments prior – isn’t asked to repeat everything on the call. It also means connecting the data the AI needs, whether it’s from the CRM, ERP, or billing, while also bringing legal, risk and compliance in before launch so they can understand the downstream consequences of any changes.
Keep customers in mind. Cost savings only count if the customer’s problem is actually solved. The strongest AI examples make experiences better, not just cheaper. Take the hold queue for instance. A caller complains about the music, and the virtual agent offers to change the tune or put it to silence. Similarly, an agent could read into the subtext of someone’s query: rather than just telling them about stock levels in one store, it could tell them about nearby stores and talk through similar products that are available.
An open future
Organisations that are treating AI as a flexible, results-led journey are the organisations targeting longer-term benefits. They start small, plan beyond the pilot, and rebuild workflows around AI. They are ready to fine-tune their deployments and prioritise a smaller number of projects. They aren’t paralysed by needing a future that’s set in stone, they embrace controlled experimentation. And they appreciate that because of AI’s power, the future is more open with opportunities than ever before.
Choose the right partner to leverage AI. CallTower is a strategic advisor supporting businesses from discovery to deployment and beyond, assessing AI readiness and turning pilots into successful projects that deliver results. Visit CallTower for more.
