AI doesn’t suck. You suck at AI.

The recent Telcos & AI DSP Leaders Report shows 61% of executives don’t believe artificial intelligence (AI) is definitely making their jobs easier. They call it “overhyped,” and “underwhelming.” My personal favorite: “It’s just easier to do it myself.”

Time for some tough love: It’s not the AI that sucks. 

It’s you and your crappy prompts.

The gap between mediocre and game-changing AI outputs isn’t in the technology. It’s in how you’re using it. And unlike most enterprise technology problems, this one has a five-minute fix.

AI isn’t magic. It’s an intern with a perfect memory.

Think of AI as the world’s most brilliant intern who has read everything ever published but zero real-world experience. Without your guidance, AI will deliver generic, amateur mush. With your expertise shaping its work, you’ll get multi-million dollar, actionable insights.

It’s especially true for telecom executives. Your decades of network operations experience, regulatory knowledge, and customer behavior insights are precisely what AI models lack. You understand the real-world constraints of spectrum licenses, infrastructure investments, and legacy system integration—context that transforms generic AI outputs into actionable telecom strategy.

When a telco exec asks, “How can we increase ARPU?” they get obvious, generic answers. But when they ask, “As a tier-2 mobile operator with 2.3M subscribers, 22% prepaid, 70% data utilization during peak hours, and limited 5G spectrum, what three service bundles would maximize ARPU for our postpaid family plans without increasing churn risk over the next two quarters?”, the difference in the quality of the response is mindblowing.

Here’s what most AI discussions miss:  The ROI of AI doesn’t scale linearly with better prompts. It compounds exponentially when executive expertise meets effective prompting. When a junior analyst uses good prompting techniques, they might get 2x better results. But when a C-suite executive with 25 years of industry experience masters these same techniques, the impact is 10-50x because they’re unlocking insights at the intersection of deep human expertise, computational intelligence, and organizational influence—a multiplier effect no algorithm alone can achieve.

How you frame the question is the difference between wasting money on AI and getting truly valuable insights from it. Here are five prompting techniques every executive needs to master:

1. Be ruthlessly specific

Amateur prompt: “How can we reduce network congestion?”

Telco-exec prompt: “We’re experiencing 32% packet loss during peak hours (7-9PM) in our downtown cell sites serving 180,000 subscribers. Our infrastructure includes Nokia RAN equipment, Cisco core network, with 65% of traffic being video streaming. We have $3.2M CapEx available this quarter and need to maintain 99.9% SLA for enterprise customers. What are the three most cost-effective congestion management tactics we could implement within 45 days?”

See the difference? The second prompt gives the AI actual parameters that mirror your business reality. Stop asking vague questions and expecting specific answers.

2. Use expert language

If you want expert-level output, speak like an expert in your input.

Amateur prompt: “How can we make our customers happier?”

Telco-exec prompt: “Using the Net Promoter Score (NPS) framework, analyze potential experience improvements for our premium account tier that has a 32 NPS versus 42 for the industry benchmark. Focus on digital touchpoints and self-service capabilities that could improve satisfaction without adding headcount. Include expected NPS impact and implementation complexity for each recommendation.”

Leverage your expertise. By using professional frameworks and metrics, you force the AI to operate within sophisticated paradigms rather than spitting out advice from Business 101.

3. Force comparative analysis

When evaluating options, make the AI show its work across multiple approaches.

Amateur prompt: “Tell me about cloud migration.”

Telco-exec prompt: “Compare big-bang versus phased approaches to migrating our telco’s 500+ on-premise BSS/OSS applications to AWS. Analyze these approaches across: initial CapEx requirements, subscriber service disruption risk, time to achieve 30% OpEx reduction, security vulnerabilities during transition, and impact on existing SLAs. Assume we’re a tier-2 operator with 18 million subscribers, 4 decades of customer data, 200+ vendor integrations, and regulatory requirements for 99.999% service availability. Include a migration sequence recommendation for our most critical systems: billing, provisioning, and customer management.”

This approach forces the AI to identify nuances between different approaches—exactly what you need for executive decision-making.

4. Demand structured outputs

Tell the AI exactly how you want information organized so it’s immediately actionable.

Amateur prompt: “Which vendor should we choose for our 5G core upgrade?”

Telco-exec prompt: “I’ve uploaded RFP responses from Ericsson, Nokia, and Huawei for our 5G core network upgrade. Create a weighted decision matrix using the criteria and weights from our uploaded evaluation rubric: TCO (30%), technical capabilities (25%), integration with existing 4G infrastructure (20%), security compliance (15%), and implementation timeline (10%). For each vendor, score each criterion on a 1-5 scale with evidence-based justification. Present results as a formatted table showing individual criterion scores, weighted scores, and final rankings. Then create a separate risk assessment matrix identifying the top three implementation risks for each vendor, scoring each on probability (1-5) and impact (1-5), with proposed mitigation strategies for any risk scoring 16+.”

By specifying exactly how you want information structured, you get decision-ready insights. For telecom executives evaluating complex vendor proposals, this approach can compress weeks of RFP evaluation into hours while actually improving decision quality. Rather than having your team manually extract and compare data points across hundreds of pages, you’re getting precisely formatted analyses that highlight the critical differences that impact your business.

5. Channel industry experts

Leverage AI’s knowledge of telecom thought leaders by asking it to analyze problems through their perspectives.

Amateur prompt: “What’s the best approach for Open RAN adoption?”

Telco-exec prompt: “Compare and contrast Open RAN deployment approaches as if you were Santiago Tenorio (Verizon’s SVP of Network Architecture, former Vodafone Network Architecture Director and TIP Chairman) versus Neil McRae (BT’s Chief Architect and Juniper Networks’ CTO). Analyze their likely recommendations for our tier-2 operator with 2,500 cell sites and $40M annual RAN CapEx. Focus on their probable disagreements regarding: implementation timeline, vendor diversity requirements, in-house skills needed, and integration with our existing Nokia RAN. Present as a side-by-side comparison with a ‘key takeaways’ summary highlighting where they would strongly agree and disagree.”

By channeling recognized experts with different viewpoints, you gain access to diverse perspectives informed by years of published thinking and industry influence. Think of it as asking “What would Santiago do?” without the $10,000/day consulting fee to actually hire him. You’re leveraging AI’s analysis of leaders’ public statements, published works, and industry positions to simulate their expert perspectives on your specific business challenges.

From generic inputs to game-changing outputs

Leveling up your prompting game creates an entirely new competitive dynamic: the exponential advantage.

While your competitors might be implementing AI through their technical teams, you have the opportunity to create a multiplicative effect by putting these prompting skills directly in the hands of your most experienced, seasoned, and expert executives. When your CTO, CMO, and CFO all become proficient in these techniques, they’re not just getting better AI outputs—they’re creating an unbridgeable competitive moat.

AI will not replace executives who master these prompting techniques. But executives who master these techniques will absolutely replace those who don’t.

This is just the tip of the AI-prompting iceberg. Want to learn more? I’ll be discussing these strategies and more at TelecomTV’s DSP Leaders World Forum in Windsor, June 2-4. Come find me there, and bring your toughest business challenges. We’ll transform your kindergarten-level prompts into million-dollar ones that deliver 100x returns for forward-thinking telecom executives.

Until then, stop blaming the AI and start fixing your prompts.

Recent Posts

Get my FREE insider newsletter, delivered every two weeks, with curated content to help telco execs across the globe move to the public cloud.