Agent Assist Best Practices

Discover use cases, prioritize capabilities, build a roadmap and organize your game plan.
As with any technology project, the streets are littered with failed implementations. Agent Assist, when approached without a sound strategy, is no different. Arguably, agent assistance is more complicated. You have cutting edge technology, from automation to AI, coupled with channel integration and agent adoption. This is not something left to chance, built on a whim or selecting the latest hot start-up that leverages buzzwords like AI and machine learning. In this section we explore the do’s and don’ts to ensure success in implementing agent assist solutions.

Identifying Use Cases and Outcomes for Agent Assistance & Automation

It should be readily apparent that “agent assistance” is more of an umbrella term referring to a set of capabilities that can be deployed for better customer interactions. As such, there is an inherent danger of casually embarking on “agent assistance” solutions without defining the specific use cases and capabilities you need. Starting with the technology or vendors is not setting you up for success.
Instead, work on your outcomes and use cases first, followed by evaluating technology for each of those. While there are many formal methodologies you can utilize (including Jacada’s), you don’t need to over-complicate this. A good starting point is utilizing a simple table such as this:
We want to…
…but our challenge is…
so it would be nice if…
Possible solution
Drive Sales & Protect Revenue
but our challenge is that sales performance varies drastically across our workforce
we can make every agent sell like our best agent
Learn from your best agents and assist new agents, seasonal hires and service agents in real-time to sell like your pros
Improve FCR & Reduce Errors
our call flow is too complex
automate processes and prevent errors
Guide agents in real-time to prevent errors, comply with best practices and deliver FCR
Lower AHT
we have so many apps on the desktop
reduce the apps or allow the agent to interact with fewer apps
Unify the agent desktop, or automate manual tasks so that the agent doesn’t feel the burden of working with disparate applications
Lower Agent Onboarding Time
we handle complex interactions and our agents need to be knowledgeable
enable multi-skilled or universal agents who can handle diverse and complex interactions without extensive training
Automate the process using guided workflows, answer questions using virtual assistants and improve agent adoption using analytics
While one size does not fit all by any means, here are popular automation use cases from 2020, as contact centers sought to simplify the work-from-home agent experience with automation and real-time assistance.
Attended RPA Use Cases

Top RPA Use Cases to Empower Work-from-Home Contact Center Workforce

From that initial definition, you should now select some discrete areas to design and implement first. It is imperative that you do not attempt to boil-the-ocean. Instead, drill-down from your initial definition and choose either low-hanging fruit or areas where you can make a big impact.
For example, in the above context for improving FCR, you may develop a heat-map to see where the complexity lies, perhaps even for a specific call type:
Call Time
Stage
Notes
00:00 to 00:30
Identification and Verification
Our agents need to touch 3 systems to complete this process. It would be so much easier if all this information was surfaced in one window. We estimate we could save 18 seconds off of each call by doing this. Maybe this can be automated?
00:30 to 01:30
Intent
It is surprisingly complicated to get to the true customer intent. While the customer is talking the agent is thinking about which apps and data to pull-up. Maybe a combination of real-time speech analytics and automation can be utilized?
01:30 to 04:30
Resolution
The bulk of the call is spent trying to resolve the issue … it involves endless searching in a knowledge base. Our agents have to look at one article after another. If we can surface the right information, we can shave off a ton of time here.
04:30 to 06:00
Disposition
It’s hard to believe, but our agents are taking 1:30 to disposition the calls. So much note taking that has to happen and fields to be entered. We can definitely automate some of that.

How to Identify Use Cases for Robotic Desktop Automation In Your Contact Center

A very good methodology is outlined in Forresters “Use RPA To Deliver Better Customer Service Experiences” research brief where a heat map approach is also advocated. In their approach, they suggest using a heat map based on the agent actions (instead of the call stages as defined above). For example, finding where a lot of Cut & Paste actions are occurring, or where an Agent is going to Excel to do a calculation. For each of these actions, the benefits of RPA can be analyzed. In the same report, they also advocate for analyzing calls for both their similarities and differences. Where calls have similarities of actions and processes, those are strong candidates for automation.
Speech Analytics have an additional highly valuable, yet often overlooked, use case. Gartner have proposed an excellent use-case in their research note titled “Use Speech Analytics to Optimize Contact Center Costs With Self-Service, Process Improvement and Deeper Engagement”. Given the constant pressure to cut costs, Gartner advocate for using speech analytics to discover which types of inquiries could be automated or avoided, before addressing the improvements. In many respects, it is similar to the heat-map approaches discussed above, but using technology to provide additional insight.
Armed with this information, surgically go in and solve a specific problem. This is often a judgement call when you look at the intersection of complexity to implement vs value gained. While the above example is purposely a little simplistic, the key point here is that you don’t have to implement everything in one go. In fact, we recommend that you don’t! Your agent assistance vendor should have a sound methodology and strong professional services to guide you through a proper analysis phase. If they don’t, just run away.
real-time conversational ai

How to Identify Use Cases for Assisting Agents
Using Speech Analytics

Prioritize Capabilities

It is important that you map out the capabilities you need. First, you can prioritize and roll out the most needed capabilities first. Second, knowing the full set of capabilities you need will help with vendor selection. Ideally, most of your capabilities will be provided by one or two vendors, instead of relying on a patchwork of various technologies, and trying to integrate them all.

How to Get Started with Agent Assist?

Deliver the real-time assistance your agents need. Get up and running in weeks.
Agent Assist
Call Center Scripting Edition
Guide agents especially on the most dynamic conversations with customers
Agent Assist
Call Center RPA Edition
Guide agents and automate mundane tasks to help agents focus on the customer experience
Agent Assist
Call Center AI Edition
Guide agents and automate tasks based on real-time alerts from conversational AI and RPA

Design & Build

Now that you’ve identified what use case you want to implement first, the design and build can start. The good news is that you have a clearly defined use case and a relatively small implementation effort. We recommend that most environments look for use cases that can be implemented from start to finish (live in production) in 3 months or less. The exception to this rule is when you are looking to implement a full unified agent desktop, as these projects generally have a longer implementation time.
In our example above, you may decide to implement the Identification and Verification use case first … Estimates are a 6 week implementation and the net result is saving 18 seconds from each and every call….sounds like a no-brainer.

Test & Launch

Agent adoption is a key, yet often neglected, aspect of implementing agent-assistance solutions. That is what we recommend deploying frequently in bite-sized pieces so that agents can quickly appreciate the benefit being delivered, without massive disruption. Moreover, agents, upon seeing early results, will often become your key product managers with great suggestions for improvements.
The beauty of choosing well defined use cases to implement is that you can typically launch them into production without requiring a corporate act-of-congress.

Continuous Improvements

No matter the perfection of your design or the purity of your intent, you will always find room for improvement once your code is in production. And this is the beauty of continuous improvement … make changes, push them to production, and monitor for additional improvements. It is important that, at the outset, expectations are set within the organization that these projects are not “once and done”, and that they will continually be improved. The best results we see are from organizations starting with a specific use case and then continually refining the solution based on “real world” performance.
Discovery
Crawl
Walk
Run
Goal
Assess processes and journeys to determine fit for solution archetypes
Design and implement a customer facing pilot for select use cases​
Incorporate best practices and lessons learned from crawl phase and prep to scale the program​
Design and implement additional use cases per roadmap​
Duration
2-4 weeks​
8-12 weeks​
8-12 weeks​
12+ weeks​
Consulting
Review use cases across journeys and touchpoints​

Evaluate fit from a CX, EX and business impact standpoint​
Design use cases with client team​

Measure solution effectiveness
Review adoption analytics in the context of the crawl phase

Engage in consulting workshops with client to agree on go-forward approach
Design use cases with client ​

Implement uses cases

Use agile processes & measure effectiveness
Implementation
Identify use cases for crawl phase

Technical assessments​
Design use cases with client team​

Implement use cases

Launch Pilot​
Train Client and Partner Teams on best practices to scale the program​

Incorporate lessons learned​
Enable client to own / manage platform​
Deliverables
Use case Design artefacts​

Business Case Assessment​
Exec dashboards reviewing solution effectiveness​
Recommendations to scale the program​

Blueprint & Roadmap from business and tech perspectives​
Exec dashboards reviewing solution effectiveness​
Success Criteria
Use cases for Pilot and Pilot success criteria​
Achievement of pilot success criteria​
Agreement to scale the program​
Achievement of success criteria​

Voice of the Agent

While many organizations have implemented Voice of the Customer solutions, the agent is quite often neglected. Understanding the Voice of the Agent is critical to improving your agent assisted solutions and for maximizing adoption.
In Gartner’s guide “Effortless CX – boosting agent experience and productivity” , they advocate for turning your VoC programs inwards. Those focus groups, surveys, and observations can be applied to your agents in order to gain a better understanding of their day-to-day world. It is also imperative that you provide channels for your agents voices to be heard and acted upon.

Capability Maturity Model for Agent Assist

While the above is meant to illustrate a concept, your vendor or systems integrator should have a formal methodology for implementing agent assistance solutions. These methodologies will ensure rigor in the process and drive successful outcomes. Outside of the vendors, Gartner and Forrester also offer pragmatic approaches to deploying assisted service technologies.
By way of example, in their excellent “How To Build A Modern Agent Desktop And Transform Customer Service Experiences” document, Forrester lays out a four phase approach to adding capabilities to your desktop.

Continue to the Next Page

Buying Agent Assist