Akamai Swarms Instead of Escalates

Last updated
Save as PDF

Akamai improves employee experience and response time by using Intelligent Swarming in a complex environment.

Introduction

The Technical Support organization of Akamai Technologies Inc. services a diverse set of customers who depend on Akamai’s technology for content delivery networks, cybersecurity, and cloud services that provide web and Internet security. This is not a simplistic environment! Leveraging a mature Knowledge-Centered Service (KCS®) framework, implemented in 2017, Akamai has seen tremendous benefits in capturing and delivering known answers through self-service. However, leadership felt the escalation process of moving new issues from one level to the next level needed attention. Intelligent Swarming was investigated as a potential new way of working.

Can We Optimize Escalation?

Akamai has a great customer-first mentality and many aspects of the Support organization were running very well based on traditional support thinking and measures. However, with fast growth taking place, the complexity of many issues, and the potential impact on customer productivity, there was a growing sense that the traditional, tier-based model was starting to show its cracks. As part of a deep assessment, segments of the data actually looked very good. A high percentage of cases were being solved at the first tier and Customer Satisfaction indicators were high. However, a wider look at the data showed that as customer issues (cases) moved along the escalation path, there was an exponential increase in time to resolve, an increase in cost, and a decrease in customer satisfaction. And while the cases moving required greater investigation and troubleshooting by higher-skilled people, the increases in time and cost did not correlate simply to increases in complexity of issue. The inherent inefficiency in moving work was leading to greater backlogs, slower responses, and duplication of efforts. On top of this, Akamai’s growing technology stack was making it harder and harder for single resources to solve the most challenging customer issues.

Pilot Group Pre Implementation Data

It left the operations team with challenges and questions such as:

Can we find a way to help engineers leverage each other’s expertise in real time?
Can we achieve faster time to value for new hires?
How can we accelerate learning and time to resolve?
Can we improve cross group collaboration and free up Tier 2?
Can we move towards a flat organizational structure?
We are strong in KCS, but can we improve on knowledge sharing?
How do we manage a large growing product technology stack or underlying technologies that don’t have a clear escalation path?
We are growing, but we can’t grow support at the same rate as our customer volume, can Intelligent Swarming help us scale?

The leadership team discovered that the benefits of Intelligent Swarming would address their challenges, while also offering a way to socialize and scale the structure and process of this new approach.

Swarm Instead of Escalate

In February of 2020 Akamai’s operations team and support leaders engaged the Consortium for Service Innovation to facilitate an Intelligent Swarming Workshop which was held onsite in Cambridge, MA. A cross functional design team was brought together under the leadership of the operations team. The design team included knowledge workers (generalists and experts), support managers, software product engineers, and resources from operations that could do data analysis, CRM design, and program management. The design team focused on developing a plan that would address people, process, and technology as it related to each of the Intelligent Swarming Practices: Connect, Collaborate, and Recognize.

The design team also understood and embraced that implementing a change like Intelligent Swarming was a journey that would take time. The design session focused on ‘good enough to get started’ with a long term road map. One driving factor in becoming successful with the adoption, getting people engaged, and avoiding a major delay was to architect “Collaborate” using the tools they currently had. Though the requirements for Recognize and Connect would require more bandwidth and technology, the designs were architected to be implemented as subsequent phases. The design team also spent a lot of time focusing on organizational change management, which was key in driving behavioral changes.

Results

Intelligent Swarming took off like wildfire with the pilot teams embracing this new collaborative environment immediately. Less than two quarters after the kick off, the pilot team saw the following:

20% lower Level of Effort (LOE) for cases swarmed versus cases that didn’t swarm and were escalated to Tier 2 following the traditional model
7% reduction in transitions to Product Engineering. When issues were sent to engineering, they were more accurately qualified after swarming
68% of knowledge workers said they were learning from following swarm posts (upskilling)
75% of knowledge workers said they were watching to see if they could offer help
The average response time for a request for help was 16 minutes! A big difference compared to the acceptance (not response) time of an Escalation Request (ER) which averaged near 2 hours

One of the benefits clearly seen was the fact anyone could ask for or provide help. Participation, both requesting and responding to swarms, was seen coming from various teams, locations, and seniority of knowledge workers. It quickly became an easy way to get help when KCS knowledge articles and other resources were exhausted. As an opt-in model, the design team wondered if swarm requests might go unanswered. That question never became a concern; people were more than willing to help. Many found it to be highly satisfying and some enjoyed it so much they helped provide answers using their mobile device when they were off shift!

The employee experience also improved, as evidenced by survey results and informal feedback. Some of the captured comments from engineers and managers include:

“It's pretty amazing. You can pick different brains from different teams such as T3, or people in different geos. It's way easier than opening a T2 ER and getting just one person's opinion or what you know within your team.”
“Quick responses that expedite investigation; visibility to a lot of other issues that peers might be working on and that improve my understanding of the topic.”
“It is very helpful for complex queries”
“I hope other teams (outside of support & services) join the swarm soon”
“My Technical Support Engineer is very happy with Intelligent Swarming and likes this way of working; it's helping her.”
“I love it!”

The level of excitement seen across the organization was higher than expected. Vice President of Support Services, Scott Lerner recalls, “Because of our success with KCS, when we started the journey of Intelligent Swarming the team had a ton of curiosity and excitement. We were trying to pace the roll out and started with the team that supports our Media product line. The excitement proved too much for those that support our Web Delivery product line, and they couldn’t wait their turn. We caught wind of them trying to implement a skunkworks IS framework. Rather than having them invest unnecessary cycles, we accelerated their onboarding to the formal program ahead of schedule. Definitely the good kind of problem to have when rolling out a new framework.”

Lessons Learned

Timing

When Akamai initially looked into Intelligent Swarming, they were in the middle of their KCS rollout. Timing needs to be right for any major change initiative and considerations for how much any organization can undertake. Amit Singh, Sr. Director and Executive Sponsor of the Intelligent Swarming implementation points out: "First consider timing: is the time right for such a large change? Second, this is a large transformation, so in order to manage it successfully, leverage organizational change management principles, and arm your design team with the knowledge needed for advocacy."

People First

The design team, which was composed of a diverse group of knowledge workers, fully embraced the Consortium’s mentality that change is about people, not technology. Technology is an enabler, but aligning people to the purpose of the organization and allowing them to design and own the Intelligent Swarming processes and communication helped increase adoption. Removing barriers to solving issues was a win everyone embraced: improving the employee experience in turn improved the customer experience.

Build a Good Data Architecture From Day One

When the topic of Swarming is raised, many people think immediately of one of the many available corporate collaboration tools as an easy way to facilitate the “action” of swarming. It will pay dividends to think more broadly about the user workflow and reporting needs that will be critical to building a robust swarming platform. Two of of our key architecture requirements were:

Swarm conversations must be integrated into the case experience
- Trigger swarm requests from within cases
- No cut and paste of information back to cases required
- Searching cases returns swarm results too
Every Swarm action is a healthy action and must be reportable at both the user and case levels (asked, answered, followers, best answers etc..). This data will be critical to your adoption, reputation, balanced scorecard, and KPI impact metrics and will help you tell the story of how Swarming is transforming the way people collaborate to solve customer issues at your company.

Don’t Over-Engineer to Start

There is danger of over-designing your requirements or thinking a tool will “turn on” Intelligent Swarming for you. Early in design, people naturally try to build the perfect solution, but in a major change, learning from experience will lead to greater long term success. The Akamai design team emphasized the importance of helping people understand the Intelligent Swarming guiding Principles and Core Concepts to get started. The initial implementation was void of rules of engagement or ownership and, instead, was built upon a fully “opt-in” model with plenty of visibility for users to see opportunities to help and dashboards that showed the impact that their collaboration was having. This keep-it-simple for users approach proved to be very successful and has not been changed since initial launch.

Akamai’s Recommendations for a Successful Adoption

Study and understand the Intelligent Swarming Practices Guide
Engage with the Consortium for Service Innovation, Consortium Member Companies, and available resources to learn from others
Communicate the “why.” Why will it benefit your organization? Ensure the “why” is understood and clearly articulated
Understand the full scope of the change, and treat the implementation as a long term program with clear milestones
Invest in Change Management
Most important: when in doubt, make it easy to do the right thing

Next Steps

Akamai continues to mature their Intelligent Swarming implementation as they learn more by doing. They have also become a thought leader in many aspects of the Practices and Design Techniques. Currently they are focused on building a value creation data model that will feed a reputation engine. This is a multi-dimensional view of how people are contributing to overall success by triangulating across seven areas: Communication Quality, Case Health, Documentation Quality, Customer Sentiment, Collaboration, Investigation Quality, and Knowledge (KCS). They are also building and training an AI / Machine Learning tool that will help automate work visibility match the right work with the right people.

As of early 2022, both of these advancements are being put into production, which will allow all three Intelligent Swarming Practices (Connect, Collaborate, and Recognize) to be fully operational as the way Akamai supports customers.

About Akamai

Akamai Technologies, Inc. is a global content delivery network, cybersecurity, and cloud service company, providing web and Internet security services. Akamai's content delivery network is one of the world's largest distributed computing platforms, responsible for serving between 15% and 30% of all web traffic.

About the Consortium for Service Innovation

The Consortium for Service Innovation is a non-profit alliance of organizations focused on innovation for the support industry. The Consortium and its members have developed the KCS methodology since 1992 and Intelligent Swarming since 2009, and are committed to developing innovative ways to deliver customer support.

Consortium for Service Innovation™ and the Consortium for Service Innovation logo are trademarks of Consortium for Service Innovation. Intelligent Swarming℠ and KCS® are registered service marks of the Consortium for Service Innovation.

All other company and product names are the property of their respective owners.