Big Carriers Aren't Benefitting from AI - But You Can
- May 7
- 4 min read
Today, large carriers are investing huge sums of money into AI - but it doesn’t seem to be paying off.
Most well-known providers have seen a drop in customer satisfaction scores over the past two years, and with them, higher churn rates. Throwing millions at AI hasn’t increased their growth or boosted customer retention. Still, big carrier business analysts - and some clients - are pushing towards “contactless” customer service operations.
However, when applied strategically, AI tools can produce measurable benefits. From automating tasks to synthesizing complex data, AI can help streamline operations and solve difficult problems.
So, in an age where there is pressure to use AI or be left behind, how should you approach it with a limited budget? And how can you ensure it actually contributes to your organization’s success? Find the answers using a seven-step approach.
Inventory your existing tools and applications
Start by taking stock of all your tools, from customer relationship management to network monitoring. Do any of them have built-in AI features? Do they offer APIs or data exports that allow for integration with other AI tools? Understanding the capabilities of your current systems helps reveal where AI may be ineffective or redundant, and where it has potential.
Determine if the data is trustworthy
Whether due to problems in a system or human error, data isn’t always reliable. Evaluating its accuracy, completeness, and consistency within and between your platforms - as well as determining if it’s real-time, close to real-time or delayed - is key to understanding whether AI will enhance incorrect facts and assumptions. If that’s the case, some automated tools may do more harm than good.
To ensure its reliability, you may want to standardize the formatting of data, ensure that all records are as up to date as possible, and establish a primary source for key datasets.
Identify bottlenecks in your existing processes
What are the operational issues your organization faces today? Are you dealing with fragmented tools, a backlog of trouble tickets, tedious manual tasks, or something else? Collect employee feedback and hard data to pinpoint processes that are sucking up money and drawing out tasks.
If you don’t know where to start, pick one small and specific challenge your organization is facing, bring it to a chatbot like Claude, Gemini, or ChatGPT and begin to break down an issue, explore its causes, and try to find a resolution. This serves as a low-risk way to test-drive AI tools and discover their capabilities while working towards greater operational efficiency.
Evaluate whether a bottleneck can be addressed by AI
AI isn’t the magic solution to all efficiency-related problems within processes. As such, it’s important to identify the root cause of an issue before deciding how to address it.
With the ability to quickly crawl and analyze large amounts of information, AI models generally excel at pattern recognition, prediction, synthesis, and automation. This makes them useful for solving issues stemming from manual tasks and complex or overwhelming data.
If an issue stems from the design of underlying processes and architecture, however, AI may not be an effective fix. Adjusting workflows and integrating systems can serve as more straightforward solutions that support long-term efficiency.
AI also cannot address underlying personnel issues such as a lack of training, poor communication, or misaligned expectations. If these are a cause of inefficiencies, improving team structure, accountability, or skill development will have a far greater impact than implementing new technology alone.
In general, AI tools are less useful when it comes to addressing bottlenecks related to gray areas, subjective tasks, and anomalies. Without historical data and well-defined guidelines or processes to draw from, they become less useful and more likely to cause problems that you have to clean up. In these cases, too, creating or redesigning processes and guidelines may be a better use of resources.
Agree on whether processes actually need improvement
While it can be tempting to immediately buy into a flashy AI tool that promises streamlined operations, the truth is, it’s not always needed. Individuals on different teams or in different roles may have varying perspectives on which processes could be more efficient and whether AI is the key to making that happen. It’s important to discuss these viewpoints and reach an agreement on whether improvements are needed. If an AI tool is looking for problems rather than addressing a specific bottleneck, it will be much less effective.
Determine whether individuals are willing to utilize AI
When evaluating whether to implement AI into workflows, it’s important to know if the individuals who would use it will buy in. Are teams willing to adopt a new and possibly unfamiliar technology? Do they trust its outputs? Are they afraid of being completely replaced? Gauging these sentiments can determine if AI will actually be useful for your organization.
If there are some reservations about using AI - and even if there aren’t - it’s important to be clear and communicative about what a proposed tool is used for. Emphasizing that AI should be used for augmentation can soothe very real anxieties about job security and encourage individuals to give it a chance. In-depth trainings, clearly outlined use cases, and discussions about how to employ AI responsibly and critically can also increase comfortability with tools.
Evaluate which AI models you can trust
Before sharing data with AI models, it’s important to evaluate how much they can be trusted with it. While the free plans for mainstream tools like ChatGPT, Gemini, and Claude are fine for querying about public information, industry research, and templates, if you plan to have models crawl internal documents and integrate with your systems, it’s critical to choose a paid plan that protects your data.
Further, if you plan to give AI access to customers’ personal information, finance details, and contracts under NDA, purchasing an enterprise plan that follows advanced security and compliance protocols is critical.
AI’s capabilities are game changers for network operations, but not the sole answer to the puzzle. The tools should be employed with purpose in the correct contexts. Through careful evaluation and thoughtful discussions, your company can benefit from AI.
Want more insights from 7Sigma? Follow us on LinkedIn.
