Transforming Risk Management at Virgin Media O2 with Generative AI

Virgin Media O2 (VMO2), a leading UK telecommunications company under the Telefónica and Liberty Global parent brands, embraced Strategy AI to streamline core risk management processes and operations using trusted data with novel advancements in generative AI. Partnering with Strategy, VMO2 launched a fully-tuned Risk Insights Bot to help risk management practitioners reduce manual workloads, improve data quality, and quickly access and interpret insights.

The AI Risk Bot didn’t just improve efficiency: it transformed how the group risk management team conducted their work. From reducing the time spent on analyzing data to assisting with quality assurance (QA) checks, the bot has become a key part of their workflow. Here are some of the use cases they explored—and their top tips to help other teams replicate their success.

Accelerated Risk Reporting

Risk management is both an art and a science for practitioners, requiring a proactive eye to identifying new risks and due diligence in managing them throughout the lifecycle. Unfortunately, simple tasks like monitoring risk status and reporting on risks by division can be relatively manual and labor intensive without a solution to help teams prioritize their next best actions.

VMO2’s new AI bot helps streamline risk reporting processes by assessing vast amounts of information and using generative AI prompts to refine result formats into the specific answers or desired charts needed for key presentations. Reporting on risks relating to new laws and regulatory compliance concerns is one area in which this has proven particularly helpful, enabling analysts to copy charts they need directly into PowerPoint.

“I can ask questions, refine my answers, and get the right report outputs I need in under 10 minutes with our Risk Insights bot. It used to take 4-5 hours to gather all of the necessary information to create similar reports, the efficiency impacts from our bot have been huge.”
— Sohaib Ejaz, Enterprise-Wide Risk Advisor, Group Risk Management, Virgin Media O2

Enhanced Quality Assurance Workflows

Best practice risk management frameworks often recommend enhanced oversight for risks with potentially higher impact. These quality assurance (QA) practices require that certain types of risks are reviewed upon creation, helping practitioners ensure risks are appropriately captured, receive proper attention, and escalated if needed to ensure impactful resolution.  

Risk quality assurance is another key area where VMO2’s Risk Insights bot adds significant value. The AI bot has significantly transformed the team’s workflow by easily identifying gaps of information in the risk capture during QA checks. In addition to this, simply by asking which risks of a certain type haven’t been updated within a certain interval, the team can proactively identify and escalate any risks that might not hit their requisite update timelines—without working through audit logs manually.

“We conduct internal quality assurance checks for newly created risks to ensure they’re scoped and captured properly in the risk register, which ensures they are monitored and managed appropriately. We used to manually review the risk information accuracy but now by asking a few simple questions, our Risk Insights Bot helps us make sure information like categories, subcategories, and metrics are inputted correctly throughout the process.”
— Katie Lowther, Enterprise Risk Manager, Group Risk Management, Virgin Media O2

Business-Aligned Data Logic

Most business applications and data warehouses are built on models that don’t leverage the everyday terms commonly used by the general workforce. The solution logic makes sense to IT professions and solutions engineers, but business needs usually need to be translated into technical requirements for big technology projects.

As the VMO2 team discovered, this process is much easier with Strategy AI. While the comprehensive data fabric many customers build within the Strategy One platform is often formatted like a typical schema, the AI features allow you to easily define natural language and business terminology equivalents and proactively tune how your AI bots respond.

“Incorporating key business knowledge into the AI layer enabled our teams to interact with the bot naturally in a way that makes sense to them. It’s incredibly powerful to simply ask what you need to know and find the answer so quickly.”
— Paul Scullion, Corporate Finance, Senior Manager, Business Intelligence, Virgin Media O2

AI Bot Impact So Far…

VMO2 built, tuned, and launched their customized Risk Insights Bot in a matter of weeks in collaboration with AI experts from Strategy Consulting. VMO2’s Risk Insights Bot has led to substantial time savings through:  

• Significant reduction in time spent gathering and presenting risk data—from hours to mere minutes 

• Faster ad-hoc risk analysis for divisions that manage hundreds of risks 

• Improved data quality by automating quality checks  

The VMO2 team learned just how quick and easy it is to tune and deploy Strategy AI for business impact—and they’re now building on their success by exploring new use cases in other parts of their business.  

Top Takeaways for AI Bot Success

Here are VMO2’s top 5 recommendations when exploring Gen AI bots for business impact with data-fueled generative AI to help you achieve similar success. 

1. Pick to Prove Value  
   Start with a data-rich use case that combines structured metrics with free-form text to demonstrate value early in your generative AI exploration. This gives the AI bot a variety of data formats to process in showing its full potential. 

2. Humanize Your Language 
   Define your data dictionary with natural language to ensure bots know how your employees actually speak. Explaining equivalent business terms and acronyms enables bot users to ask questions like they’re chatting with a colleague.  

3. Test & Tune for Quality 
   Fine tune bot performance by asking the same question in multiple ways. Using a variety of language during testing ensures that the AI bot delivers consistent, accurate answers even when users phrase questions differently.

4. Socialize Your Success 
   Once an AI bot adds value, share your success with other teams to build momentum and scale use. By demonstrating impact, line of business leaders and IT practitioners can encourage other teams to collaborate on similar projects. 

5. Just Try It Out! 
   You’ll be impressed by how much of the basics the AI bot gets right from the go. Test it out early, then partner with business leads to fine-tune as you go. An agile, iterative approach helps focus your efforts and quickly add value. 

“This process taught us just how easy it is to get started with Strategy AI. I’d encourage others to simply start by putting a bot on top of your data model to see what types of questions can answers you can derive from it. We were surprised by the strength of our initial results, then partnered with the business to finesse the bot’s business acumen in the tuning process to rapidly release our custom bot.”
— Paul Scullion, Corporate Finance, Senior Manager, Business Intelligence, Virgin Mobile O2 

Start Your AI Journey Today