Newsletter Signup

At Gorgias we recently switched our flask & celery apps from Google Cloud VMs provisioned with Fabric to using docker with kubernetes (k8s). This is a post about our experience doing this.

‍

Note: I'm assuming that you're somewhat familiar with Docker.

‍

Docker structure

The killer feature of Docker for us is that it allows us to make layered binary images of our app. What this means is that you can start with a minimal base image, then make a python image on top of that, then an app image on top of the python one, etc..

‍

Here's the hierarchy of our docker images:

• gorgias/base - we're using phusion/baseimage as a starting base image.
• gorgias/pgbouncer
• gorgias/rabbitmq
• gorgias/nginx - extends gorgias/base and installs NGINX
• gorgias/python - Installs pip, python3.5 - yes, using it in production.
• gorgias/app - This installs all the system dependencies: libpq, libxml, etc.. and then does pip install -r requirements.txt
• gorgias/web - this sets up uWSGI and runs our flask app
• gorgias/worker - Celery worker

‍

Piece of advice: If you used to run your app using supervisord before I would advise to avoid the temptation to do the same with docker, just let your container crash and let k8s handle it.

‍

Now we can run the above images using: docker-compose, docker-swarm, k8s, Mesos, etc...

‍

We chose Kubernetes too

There is an excellent post about the differences between container deployments which also settles for k8s.

‍

I'll also just assume that you already did your homework and you plan to use k8s. But just to put more data out there:

‍

Main reason: We are using Google Cloud already and it provides a ready to use Kubernetes cluster on their cloud.

‍

This is huge as we don't have to manage the k8s cluster and can focus on deploying our apps to production instead.

‍

Let's begin by making a list of what we need to run our app in production:

• Database (Postgres)
• Message queue (RabbitMQ)
• App servers (uWSGI running Flask)
• Web servers (NGINX proxies uWSGI and serves static files)
• Workers (celery)

‍

Why Kubernetes again?

We ran the above in a normal VM environment, why would we need k8s? To understand this, let's dig a bit into what k8s offers:

• A pod is a group of containers (docker, rtk, lxc...) that runs on a Node. It's a group because sometimes you want to run a few containers next to each other. For example we are running uWSGI and NGINX on the same pod (on the same VM and they share the same ip, ports, etc..).
• A Node is a machine (VM or metal) that runs a k8s daemon (minion) that runs the Pods.
• The nodes are managed by the k8s master (which in our case is managed by the container engine from Google).
• Replication Controller or for short rc tells k8s how many pods of a certain type to run. Note that you don't tell k8s where to run them, it's master's job to schedule them. They are also used to do rolling updates, and autoscaling. Pure awesome.
• Services take the exposed ports of your Pods and publishes them (usually to the Public). Now what's cool about a service that it can load-balance the connections to your pods, so you don't need to manage your HAProxy or NGINX. It uses labels to figure out what pods to include in it's pool.
• Labels: The CSS selectors of k8s - use them everywhere!
• ‍

There are more concepts like volumes, claims, secrets, but let's not worry about them for now.

‍

Postgres

We're using Postgres as our main storage and we are not running it using Kubernetes.

‍

Now we are running postgres in k8s (1 hot standby + pghoard), you can ignore the rest of this paragaph.

The reason here is that we wanted to run Postgres using provisioned SSD + high memory instances. We could have created a cluster just for postgres with these types of machines, but it seemed like an overkill.

The philosophy of k8s is that you should design your cluster with the thought that pods/nodes of your cluster are just gonna die randomly. I haven't figured our how to setup Postgres with this constraint in mind. So we're just running it replicated with a hot-standby and doing backups with wall-e for now. If you want to try it with k8s there is a guide here. And make sure you tell us about it.

‍

RabbitMQ

RabbitMQ (used as message broker for Celery) is running on k8s as it's easier (than Postgres) to make a cluster. Not gonna dive into the details. It's using a replication controller to run 3 pods containing rabbitmq instances. This guide helped: https://www.rabbitmq.com/clustering.html

‍

uWSGI & NGINX

As I mentioned before, we're using a replication controller to run 3 pods, each containing uWSGI & NGINX containers duo: gorgias/web & gorgias/nginx. Here's our replication controller web-rc.yaml config:

‍

apiVersion: v1
kind: ReplicationController
metadata:
 name: web
spec:
 replicas: 3 # how many copies of the template below we need to run
 selector:
   app: web
 template:
   metadata:
     labels:
       app: web
   spec:
     containers:
     - name: web
       image: gcr.io/your-project/web:latest # the image that you pushed to Google Container Registry using gcloud docker push
       ports: # these are the exposed ports of your Pods that are later used by the k8s Service
         - containerPort: 3033
           name: "uwsgi"
         - containerPort: 9099
           name: "stats"
     - name: nginx
       image: gcr.io/your-project/nginx:latest
       ports:
         - containerPort: 8000
           name: "http"
         - containerPort: 4430
           name: "https"
       volumeMounts: # this holds our SSL keys to be used with nginx. I haven't found a way to use the http load balancer of google with k8s.  
         - name: "secrets"
           mountPath: "/path/to/secrets"
           readOnly: true
     volumes:
       - name: "secrets"
         secret:
           secretName: "ssl-secret"
And now the web-service.yaml:apiVersion: v1
kind: Service
metadata:
 name: web
spec:
 ports:
 - port: 80
   targetPort: 8000
   name: "http"
   protocol: TCP
 - port: 443
   targetPort: 4430
   name: "https"
   protocol: TCP
 selector:
   app: web
 type: LoadBalancer

‍

That type: LoadBalancer at the end is super important because it tells k8s to request a public IP and route the network to the Pods with the selector=app:web.
If you're doing a rolling-update or just restarting your pods, you don't have to change the service. It will look for pods matching those labels.

‍

Celery

Also a replication controller that runs 4 pods containing a single container: gorgias/worker, but doesn't need a service as it only consumes stuff. Here's our worker-rc.yaml:

‍

apiVersion: v1
kind: ReplicationController
metadata:
 name: worker
spec:
 replicas: 2
 selector:
   app: worker
 template:
   metadata:
     labels:
       app: worker
   spec:
     containers:
     - name: worker
       image: gcr.io/your-project/worker:latest

‍

‍

Some tips

• Installing some python deps take a long time, for stuff like numpy, scipy, etc.. try to install them in your namespace/app container using pip and then do another pip install in the container that extends it, ex: namespace/web, this way you don't have to rebuild all the deps every time you update one package or just update your app.
• Spend some time playing with gcloud and kubectl. This will be the fastest way to learn of google cloud and k8s.
• Base image choice is important. I tried phusion/baseimage and ubuntu/core. Settled for phusion/baseimage because it seems to handle the init part better than ubuntu core. They still feel too heavy. phusion/baseimage is 188MB.

‍

‍

Conclusion

With Kubernetes, docker finally started to make sense to me. It's great because it provides great tools out of the box for doing web app deployment. Replication controllers, Services (with LoadBalancer included), Persistent Volumes, internal DNS. It should have all you need to make a resilient web app fast.

At Gorgias we're building a next generation helpdesk that allows responding 2x faster to common customer requests and having a fast and reliable infrastructure is crucial to achieve our goals.

If you're interested in working with this kind of stuff (especially to improve it): we're hiring!

We've released a new version of the Chrome Extension, with sharing features and a new navigation bar. We hope you'll love it!

‍

Share templates inside the extension

Before, the only way to share templates with your teammates was to login on Gorgias.io.

If you're on the startup plan, when you create a template, you can choose who has access to it: either only you, specific people, or your entire team.

‍

‍

The account management section is now available in the extension, under settings.

‍

New navigation

Tags are now available on the left. It's easier to manage hundreds of templates with them.
You can also navigate through your private & shared templates. Shared templates include templates shared with specific people or with everyone.

‍

‍

We hope you'll enjoy this new version of our Chrome Extension. As usual, your feedback & questions are welcome!

Today, we’re thrilled to announce that we’ve raised a $1.5 million Seed round led by Charles River Ventures and Amplify Partners, to help build our new helpdesk.

‍

We’re incredibly grateful to early users, customers, mentors we’ve met both at and Techstars.

‍

We started the journey with Alex at the beginning of 2015 with our Chrome extension, which helps write email faster using templates. We’ve been pleased all along with customers telling us about how helpful it was, especially for customer support.

‍

While building the extension, we’ve realized that a big inefficiency in support lies in the lack of integration between the helpdesk, the payment system, CRM and other tools support is using. As a result, agents need to do a lot of repetitive work to respond to customer requests, especially when the company is big.

‍

That’s why we’ve decided to build a new kind of helpdesk to enable customer support agents to respond 2x faster to customers. You can find out more and sign up for our private beta here.

‍

When a company has a lot of customers, support becomes repetitive. We want to provide support teams with tools to automate the way they treat simple repetitive requests. This way, they have more time for complex customer issues.

‍

We'll now focus on this helpdesk and on growing the team, oh, and if you'd like to join, we're hiring! We're super excited about this new helpdesk product. If you’re using the extension, don’t worry.

‍

Romain & Alex

We've been busy, but not deaf!

Last few months we got lots of feedback about our extension and found to our delight that most people are satisfied, but still a few recurrent issues came up:

• The HTML/WYSIWYG editor sucks.
• No support for Outlook.com.

‍

We listened and now we're presenting:

• A brand new editor
• Support for outlook.com
• More on the Rich-Text editor

‍

WYSIWYG editors for the web are notoriously buggy and are just difficult to develop.

‍

I have yet to see one that is bug free. There are few venerable editors that do a good job like TinyMCE, FKEditor or CKEditor.. but they are big and all have edge cases that break the intended formatting and add a lot of garbage html.

‍

There are newer good quality editors in town such as Redactor. The one that got my attention and finally landed in Gorgias is this wonderful editor called which is super lightweight, uses modern content-editable (no i-frames) and 'just works' most of the time. That's not to say it's perfect, but it's good enough and I'm satisfied with it's direction in terms of development.

‍

Enjoy it and as always send us bug-reports or feedback on: support@gorgias.com

TL;DR:

• Manual QA is time-consuming—Auto QA does the heavy lifting. It frees up team leads by automatically reviewing conversations with accuracy and consistency, so they can focus on improving support.
• Auto QA scores 100% of private text conversations, whether handled by a human or AI Agent. It evaluates support quality based on Resolution Completeness, Communication, and Language Proficiency.
• Auto QA supports multiple languages but provides feedback in English. It can assess tickets in any language supported by OpenAI’s GPT-4, ensuring global teams can benefit from automated QA.‍
• Start with individual meetings before a team-wide rollout of Auto QA. One-on-one conversations help address specific agent concerns and ensure a smooth transition.

Customer satisfaction scores (CSAT) have long been the go-to metric for measuring support quality, with 53% of customer experience leads relying on them. However, CSAT only tells you part of the story. 

When customers rate their experience 3 out of 5, what does it really mean? Did they rate the agent’s actions or the company’s policies? Was an agent helpful or inefficient? Did they take unnecessary steps to get to the answer?

Quality assurance checks can fill these gaps, but manual QA is a heavy lift. Team leads often struggle to review more than a small sample of conversations, leaving many issues unchecked.

Auto QA redefines quality assurance for today’s support teams. It transforms QA from a manual task into an automated feedback engine that helps your team deliver excellent support, every single time.

Let's dive into how Auto QA works, how accurate its scoring is, and how you can add it to your support workflow to start improving customer conversations today.

What is Auto QA?

Gorgias Auto QA upgrades the customer service QA process by automatically evaluating 100% of private text conversations, whether handled by a human or AI Agent. 

Each message is scored on metrics like Resolution Completeness, Brand Voice, and Accuracy, helping teams fix and address areas of improvement.

With an automated QA process, brands can:

• Save time: Automated quality checks help team leads focus on the most critical tickets.
• Ensure consistency: Both human agents and AI agents are evaluated with a unified, comprehensive quality score.
• Boost performance: Agents can receive targeted coaching to provide more consistent customer experiences.
• Meet customer expectations: Customers benefit from higher-quality support with quicker resolutions and accurate responses.

How Auto QA works 

Let's explore a real-life scenario: A customer reaches out about a product issue, seeking troubleshooting help. Here’s how the interaction unfolds:

Customer: "Hi, my device broke, and I bought it less than a month ago. -Kelly"

Support Agent: "Hi Kelly, please send us a photo or a video so we can determine the issue with your device. -Michael"

The ticket is eventually closed, but the customer doesn't leave a CSAT score.

In this case, Auto QA would provide the following insights:

• Communication Score: 3/5. Reason: The agent's wording could benefit from more empathy.
• Resolution Score: "Complete". Reason: The agent effectively addressed the customer's concerns.

How accurate is Auto QA’s scoring?

Auto QA uses a comprehensive scoring system that evaluates conversations on communication proficiency and knowledge accuracy.

To ensure accuracy, Auto QA only scores interactions with at least 250 characters and messages from both agents and customers. It's also smart enough to filter out automated responses, spam, and bot messages.

Auto QA automatically scores three main aspects:

1. Resolution Completeness: Did the agent solve everything the customer asked about? This area is scored with a "Complete" or "Incomplete.” For instance, it correctly marks a ticket as "Complete" when a customer resolves their issue or when there's no clear question to address.
2. Communication Quality: How well did the agent listen and show empathy? Uses a 1-5 scale, looking at how well your agents acknowledged a customer’s concerns and communicated the solution.
3. Language Proficiency: Did the agent communicate properly? Uses a 1-5 scale to check spelling, grammar, and syntax.

For deeper feedback, certain criteria require manual scoring from team leads:

• Accuracy: How accurate was the information provided by the agent?
• Efficiency: How quickly did the agent handle the ticket? How well did they minimize the number of follow-ups?
• Internal Compliance: How closely did the agent follow your team’s internal processes and brand guidelines?
• Brand Voice: How well did the agent use brand vocabulary, greetings, sign-offs, and tone of voice?

How to integrate Auto QA into your workflow

Whether you're just starting with quality checks or transitioning from manual QA, Auto QA can seamlessly fit into your existing processes. Here's how to get started.

1. Set your standards

What does “good” look like for your team? Review Auto QA's scoring system and decide which metrics matter most for your brand, from Resolution Completeness to Brand Voice. This will help you set realistic targets for your team to work toward.

Tip: Start by prioritizing a couple of areas. This could look like prioritizing a 5/5 Resolution Completeness score while deprioritizing Brand Voice. As your team gets comfortable with Auto QA, you can ramp up to improving Brand Voice.

2. Agree on a scoring system

Since some criteria—Accuracy, Efficiency, Internal Compliance, and Brand Voice—require manual scoring, it’s best to agree on how your team will use the scoring scale.

For example, each score from 1 to 5 receives a distinct piece of feedback. Here’s what that would look for the Efficiency criteria:

• 1/5 stars: Excessive back-and-forth that could have been avoided
• 2/5 stars: Resolution took longer than necessary due to poor process
• 3/5 stars: Average handling time with some unnecessary steps
• 4/5 stars: Quick resolution with minimal back-and-forth
• 5/5 stars: One-touch resolution

3. Prepare your agents

Start rolling out Auto QA through individual meetings with agents rather than overwhelming your team with a general training session. One-on-one conversations allow you to better address each agent's specific questions and concerns. Make sure to cover the following:

• Explain that Auto QA is meant to help make conversations consistent, not police agents
• Explain the scoring criteria and what each score means
• Highlight which criteria agents should prioritize

If regular one-on-one meetings aren't part of your routine, consider introducing Auto QA during your weekly team meetings or through a dedicated training session. Just remember to leave plenty of time for questions and walk through multiple examples to ensure everyone is comfortable with the system.

4. Establish a review schedule

To solidify QA checks, create a simple routine for reviewing Auto QA insights with the Auto QA Report (navigate to Statistics > Auto QA). 

• Weekly: Do a quick check of automated scores.
• Monthly: Analyze trends and patterns across conversations. 
• Quarterly: Review and adjust quality benchmarks.

5. Act on insights

Once you’ve collected a substantial amount of Auto QA data, there are a few follow-up actions you can take to continue having high-quality conversations:

• Set the example by sharing high-scoring conversations in your team meetings.
• Coach agents individually by reviewing their tickets together. Celebrate high-scoring conversations and provide targeted feedback on areas for improvement. This immediate, personalized approach helps agents grow faster than general training sessions.
• Increase product and policy knowledge by refining internal guidelines on brand voice, escalation processes, and more.

Remember, Auto QA works alongside your existing processes—it doesn't replace them. Start small, focus on the metrics that matter most to your team, and scale up as you get comfortable with Auto QA.

Brands are excited about the power of Auto QA

We invited leading ecommerce brands to beta test Auto QA, and their feedback highlights how it's transforming quality assurance across support teams of all sizes.

amika's support team values the complete visibility beyond CSAT: "Auto QA dramatically widens the volume of tickets we can review," they share. "A 5-point scale only tells you so much, and relying on consumers providing feedback limits what you're able to learn from."

Peachybbies' CX team enjoys real-time improvement: "Being able to give real-time feedback is pivotal, especially during peak times," their team explains. "Auto QA catches pretty much everything I'd want a human QA agent to catch."

OSEA Malibu's managers discovered operational insights: "It helps managers understand when a macro or process is leading to incomplete conversations versus when an agent made a mistake," their support lead shares.

Bring quality into every conversation with Auto QA

By prioritizing QA, your team can identify potential problems early, reduce errors, and improve overall performance, leading to a smoother, more reliable experience for customers––and your CX team. 

In the long run, brands focusing on QA can gain a competitive edge. Book a demo now to see what Auto QA can do for you.

{{lead-magnet-1}}

There are tons of CX metrics you could be tracking. But where you spend your time is crucial as a customer experience leader. 

According to recent data, these are the top five CX metrics for you to prioritize and improve on in 2025.

{{lead-magnet-1}}

Why CX metrics are essential for success 

Not tracking CX metrics is like putting a loaf of bread in the oven but leaving baking time to chance. Without a set timer, you could end up with an underbaked bowl of dough or a burnt mess. Unless you have a sixth sense, it’s going to be really challenging to end up with something good. 

In the same vein, metrics provide clear parameters for success. Meet or exceed them and your team is doing well; fall short and you’ll be better equipped to identify pain points and solve them. 

Here are a few additional reasons why setting customer support metrics is key to success.

• Measure success and ROI. By tracking KPIs like resolution time, first response time, and CSAT, you can gauge the health of your customer support program and potentially justify investments in CX initiatives in the future.

• Identify customer and team pain points. Metrics help uncover areas where customers or your team is struggling. For example, high resolution times or low CSAT scores signal friction in the experience that you can address. 

• Create accountability within your team. When everyone on your team understands what success looks like, it aligns efforts and keeps everyone focused on shared goals.

• Prioritize resources. Metrics guide CX leaders on where to allocate resources—for example, leveraging AI and automation to tackle repetitive tickets when ticket volume adds up or resolution times are getting high.
• Get proactive. Metrics reveal trends in customer behavior which can help you predict customer needs and make proactive adjustments in your CX strategy. By monitoring customer sentiment and acting on feedback, CX leaders can create more personalized and positive experiences.

Tip 💡: AI and automation can be valuable sidekicks as you look to optimize and improve on metrics. That’s especially true for busy periods: in 2024, 70% of CX leaders relied on AI and automation during peak seasons.

Resolution time should be your main focus for 2025

Customers are done with being patient. One study found that two thirds of respondents valued speed to reply just as much as product price. 

A recent survey we ran found the same thing. 

In our 2024 customer expectations survey, we asked CX leads and agents which metric they used to track success. Here’s what they said:

• Resolution Time (71%)
• First Response Time (59%)
• CSAT (53%)
• Revenue or Sales Impact (41%)
• Ticket Volume (41%)

Resolution time is going to be a key differentiator for your team this year. It should be your primary focus when it comes to optimizing different facets of your customer service strategy. 

Top 5 CX metrics for 2025 & how to improve them with AI 

1) Resolution time 

Resolution time is the average time it takes to resolve a customer request from start to finish.

How do you calculate resolution time?

To calculate resolution time, you’ll take the total resolution time within a set period and divide it by the total number of customer interactions your team tackled within that same time frame.

Average resolution time = Total resolution time in a defined period / Total number of customer interactions resolved in that period

How to use AI & automation to improve it

According to a 2023 study from Statista, 70% of support leaders noted that the customer support metrics that AI had the greatest positive effect on was resolution time.

You can use automation tools to send Macros to answer common questions, or leverage AI to interact as an agent via email or chat. The instant nature of these tools means that customers won’t have to wait in a queue for your team to get to them.

For example, Wildride implemented Gorgias's AI Agent to manage an influx of 1,000 tickets per week. After AI Agent took over 33% of email inquiries, the team saw a 24% decrease in resolution time. That allowed the team to focus on more complex issues, streamline their support process, and make their customers happier. 

2) First Response Time (FRT)

First response time is the length of time it takes for a customer service team to send the initial reply to a customer inquiry.

How do you calculate first response time? 

To calculate average first response time, take the total amount of time it took for your team to respond to initial customer requests and divide by the total number of tickets within a set time frame. 

How to use AI & automation to improve it

Your team is busy––when they’re not tackling repetitive questions, they’re helping customers with complicated or high-effort requests. All of that work is going to bog down your FRT, especially during more buzzy periods like sales, new releases, or over the holidays. 

By using AI to jump in to handle those more routine requests, you can significantly reduce your FRT and give your team time back to tackle more heavy-lift needs. 

For example, AI Agent helped Glamnetic achieve a 91% improvement in first response time during Black Friday Cyber Monday (BFCM) 2024. They got FRT down from their pre-AI Agent time of eight minutes to 40 seconds. 

Here’s what that looked like in practice: 

3) Customer Satisfaction Score (CSAT) 

CSAT scores show how satisfied customers are with a product, service, or interaction, typically gathered through surveys.

How is CSAT calculated? 

CSAT is calculated via a five-point rating scale survey sent to customers after a support interaction, where one is the worst experience and five is the best. While it can be calculated in different ways, at Gorgias the average of all survey responses is your CSAT score.

How to use AI & automation to improve it

When customers reach out for support, they’re expecting a fast response––regardless if they have an issue or are contemplating their next purchase. 

That’s why using automation or AI tools to provide that lightning quick response, even if it directs shoppers to a self-service resource, can be extremely effective in raising CSAT scores. These responses could be sent by an AI agent that responds like a human agent would or an automated Macro built to fire off pre-crafted templates to common questions. 

In luxury golf brand VESSEL’s case, customers felt that the AI responses were helpful and seemed on-par with the level of support they’d expect from a human agent. 

“Our customers expect almost immediate responses, and so being able to automate that, even if it's not necessarily the exact answer that they're looking for, but being able to send over information to give them the reassurance that we're looking into it or trying to find an answer, whatever it may be, that's been a huge help to our team,” says Lauren Reams, the Customer Experience Manager at VESSEL. 

4) Revenue or sales impact 

The direct or indirect effect of customer service or business activities on generating sales or revenue.

How do you calculate it?

There are different ways to calculate revenue generated and the sales impact of customer support, and quantifying the indirect impact can be difficult. But generally, the formula looks like this: 

ROI = [ (Money earned - Money spent) / Money spent ] x 100

Resource: How to measure & improve customer service ROI

How to use AI & automation to improve it

Leveraging AI and automation can provide significant cost savings because it acts as an additional agent who can tackle repetitive questions, translating to money saved on the time it would take for human agents to manually answer those questions. 

The results are tangible: by automating 48% of inquiries, Dr. Bronner's saved $5,248 in the first month, and $100K in the first year. 

Jonas Paul Eyewear saw revenue influenced by AI Agent as well: the team tracked $600 of sales revenue directly to the tool after it effectively answered pre-sales support questions from shoppers. 

5) Ticket volume 

Ticket volume is the total number of customer service inquiries that a team receives over a specific period of time.

How do you calculate it?

The customer support tool you use will be able to calculate ticket volume for you, as it’s the total number of tickets that have come in within a set amount of time. If you don’t use a CX platform yet and are still using something like Gmail or Excel, you’ll perform this count manually.

How to use AI & automation to improve it

Set rules to trigger automated responses to common questions, or ask an AI agent to completely take them off your team’s plate. 

Arcade Belts, for example, saw a 50% reduction in ticket volume by using Gorgias’s AI Agent. 

How to get buy in to improve your CX program

Tracking CX metrics is valuable for more than just gauging your program's effectiveness. The more you improve upon your CX metrics, the more you can leverage them to prove your support function’s value within your company.  

1. Tie CX to revenue. Show how improvements in customer satisfaction or repeat purchase rates directly impact revenue growth. 
2. Show industry benchmarks. Compare your team’s stats to competitors or industry averages to demonstrate how well your support strategy is working.  
3. Demonstrate your team’s impact on sales and retention. Use the metrics you’ve collected to show support’s impact on converting customers asking pre-sales questions and getting repeat customers. 
4. Ask to expand your team’s budget. Pitch acquiring additional buy in and resources by presenting revenue generated, costs saved through tools like AI and automation, and happy customers created. 

How to use metrics to evaluate AI performanceIf you want to transform customer experience for the long term, the AI tools you use should never be “set it and forget it” solutions. Just as you do with your human agents, you can use metrics to evaluate your AI agent to make sure it’s performing well. If you use Gorgias, you’ll find these metrics under the AI Agent dashboard. 

To review AI Agent’s performance: 

‍It’s also easy to retrain your AI's performance by adjusting settings like Guidance, refining the internal documents it draws from, setting up brand voice, or creating a Handover topic list to escalate certain types of tickets to human agents.

Start tracking top CX metrics 

Whether you’re new to being a CX leader or you’re a seasoned pro, tracking and improving on your CX metrics will help your team stand out among the rest. A key way to improve them is to leverage AI and Automation tools, and Gorgias is here to help you do it.

‍Get started with AI Agent →

{{lead-magnet-2}}

TL;DR:

• AI Agent on Chat automates up to 50% of chat conversations. It ensures customers get fast, context-aware answers, product recommendations, and seamless handovers to human agents when needed.
• AI Agent goes beyond automated tools like Flows and article recommendations. Unlike pre-set FAQ flows or suggested Help Center articles, AI Agent can handle complex inquiries like modifying orders and providing personalized product recommendations.
• Setting up AI Agent on Chat is quick. Brands can activate AI Agent with a few clicks, improving efficiency during peak seasons and reducing the need for follow-ups.‍
• Updating AI Agent’s knowledge and behavior ensures the best customer experience. Businesses should refine their Help Center, set Guidance instructions, personalize AI Agent’s tone, and test responses before going live.

It’s clear that shoppers want answers fast—chat accounts for 20% of all customer support tickets.

The appeal is obvious: Chat is an easy-to-access customer service channel for quick questions and a convenient and subtle way to cross-sell complementary products.

But without the right chat tool, brands risk losing these valuable opportunities.

Introducing AI Agent on Chat, a conversational AI assistant that can automate up to 50% of chat conversations. This new feature upgrades chat by combining agent knowledge with superhuman efficiency and response times.

Now, customers can guarantee personalized interactions at any point of the shopping journey—whether they’re looking for a quick answer or a tailored recommendation.

With AI powering every interaction, one-to-one conversations become a seamless part of every customer experience.

Why Chat is better with AI Agent

Before AI Agent, customers reaching out through chat outside business hours had two options: following pre-set Flows (automated FAQ conversations) or browsing through suggested Help Center articles. 

These features are great for quick answers to basic questions, but AI Agent takes support to the next level by handling more complex needs like modifying orders or offering personalized product recommendations.

With AI Agent in Chat, customers enjoy dynamic, real-time conversations available on multiple channels. AI Agent generates personalized responses that match exactly what customers ask for, automating 50% of chat interactions so agents get time back to upsell, create stronger relationships, and craft better experiences.

Related: How to optimize your Help Center for AI Agent

The key features of AI Agent on Chat 

Upgrade your chat support from a basic Q&A tool into an intelligent assistant that handles customer inquiries 24/7. Here's how AI Agent makes that possible:

Real-time conversations

AI Agent responds within 15 seconds or less, offering fast responses that result in frictionless conversations. Unlike traditional chatbots, AI Agent also adapts to your brand’s unique tone of voice to enhance the customer experience and assure shoppers their questions will be taken care of. 

24/7 availability

Today’s shoppers expect instant responses regardless of time zone or business hours. AI Agent on Chat means customers get the help they need, when they need it. This availability leads to higher customer satisfaction and fewer abandoned carts.

Instant product recommendations

AI Agent understands context and customer intent. Whether a shopper needs help finding the right product size or changes their mind and wants to compare features, AI Agent customizes its recommendations for each person.

Intelligent handovers

Some conversations, like technical issues or complaints, need a human touch. AI Agent recognizes these situations and smoothly transfers them to the right agent. 

Using Handover topics, you can choose which types of inquiries should go straight to human agents. Then, if AI Agent lacks the confidence to provide an answer or can’t locate relevant knowledge in its database, it automatically escalates the conversation.

Read more: Handover rules

Why enable AI Agent in Chat now?

Based on Hiver’s 2024 study, 62% of customers prefer live chat to other support channels. With AI Agent in Chat, agents can cut down average response times while customers get the answers they need in one conversation with zero wait times or follow-ups.

Easy setup

AI Agent on Chat is ready to use in a few clicks. Simply connect your Shopify store and Chat widget to AI Agent, and you’re ready to resolve questions asked by visitors and loyal customers faster than you ever have.

Capture the growing demand for live support

Chat is often a customer’s first touchpoint with your brand, whether they’ve just discovered your brand or are on their third order. Meet customer expectations by being available with AI Agent on Chat. The faster you can ease their concerns, the faster they can head to checkout.

Maximize team efficiency

AI Agent makes scaling support effortless, especially during peak seasons like Black Friday. While it handles repetitive support tickets like order status and shipping questions, your team can focus on high-priority tasks like requests from VIP customers.

Eliminate the need for follow-ups

Drawing from knowledge sources like your Help Center and policy pages means AI Agent can often resolve inquiries within one conversation. No more unnecessary back-and-forths. Quick resolutions = happier and more loyal customers.

How to activate AI Agent on Chat

Ready to get started? Here’s how to activate AI Agent on Chat:

1. Click Automate in the top left menu.
2. Select your store from the sidebar, then click on AI Agent.
3. In the Settings tab, under Chat Settings, select one or more Chat from the dropdown menu.
4. Toggle Enable AI Agent on Chat on.
5. Select Save Changes at the bottom of the page.

Already use AI Agent for email? No need to set up Guidance and Handover topics all over again—AI Agent will behave the same way in Chat.

Best practices for setting up AI Agent on Chat

Get the most out of AI Agent on Chat by following these best practices. 

1. Prepare and optimize your knowledge base

The Help Center is AI Agent’s brain. This customer knowledge database is the key to AI Agent’s accurate and on-brand responses. To ensure your AI Agent is as trained as your human agents, include important topics in your Help Center like shipping, returns, cancellations, and account management.

No articles yet? No problem! Gorgias has 20+ article templates for you to use and modify. Or, even better, check out the AI Library for AI-generated articles based on your customer tickets.

2. Set restrictions with Guidance

AI tools perform best when you set limitations. A Guidance is the main way to control AI Agent’s behavior. It is a set of written instructions that outline how AI Agent should interact with customers, handle certain requests, and more.

We recommend publishing a Guidance on the top five questions you receive from customers.

Tip: AI Agent prioritizes Guidance above Help Center articles. Unlike Help Center articles, the content in your Guidance will not be customer-facing.

3. Personalize AI Agent's voice

The beauty of AI Agent is its ability to speak like one of your agents. Select from Friendly, Professional, or Sophisticated presets—or create a custom tone that aligns with your brand.

Need help finding your brand voice? Here are seven brand voice examples.

4. Test AI Agent’s responses before going live

Use test scenarios to see how AI Agent responds to common customer questions, such as order status, shipping questions, and return policies. To cover all your bases, test AI Agent as both a new and returning customer to make sure it delivers accurate responses no matter the customer's need.

5. Improve AI Agent’s behavior

AI Agent becomes smarter as it learns from you. Like a human agent, give your AI Agent feedback on its responses, from how it speaks, which topics it escalates, and what actions it takes in certain scenarios. 

There are multiple ways to give AI Agent feedback on a ticket:

• Mark AI Agent’s message or any of the resources it used as correct or incorrect.
• Suggest that AI Agent use a different resource if a better or more correct piece of knowledge exists.
• Report an issue to the Gorgias Product team.

Coming soon: Actions on Chat

Soon, AI Agent will be able to perform actions like accessing Shopify order details and executing third-party app actions, such as updating shipping addresses and order cancellations, directly in Chat.

Excited to deliver an elevated chat experience? Book a demo now to experience the power of AI Agent on Chat.

{{lead-magnet-1}}

Managing customer support as a Shopify store owner can feel like juggling too many tools at once.

Constantly switching tabs to look up orders, update customer information, or track returns wastes valuable time. Plus, it prevents your team from focusing on what really matters––delivering quick, personalized customer service. 

Gorgias’s Shopify integration solves this. It keeps all your Shopify data in one place, so your team spends less time toggling tabs and more time helping customers. The result? Faster responses, better service, and more revenue.

Below, we break down the eight key capabilities of this integration, each paired with practical use cases to showcase its real-world value.

{{lead-magnet-1}}

1. View Shopify data in tickets

What it does: Shopify order data is displayed directly within support tickets, allowing agents to view essential details like order status, customer information, and transaction history without leaving the helpdesk.

Use case: An agent handling a “Where’s my order?” request can instantly check tracking information and update the customer.

The fashion retailer Princess Polly improved their customer experience team’s efficiency by using Gorgias's deep integration with Shopify. Agents can view and update customer and order data directly within Gorgias, eliminating the need to switch between multiple tabs.

Taking a streamlined approach led to a 40% increase in efficiency, an 80% decrease in resolution time, and a 95% decrease in first response time. 

2. Perform Shopify Actions

What it does: Agents can update Shopify order and customer data with Shopify Actions right in Gorgias.

Key features:

• Create a new order: Add existing products or custom items, apply discounts, modify quantities, add notes and tags, and choose to charge taxes. Then set the order as Paid or Pending and email the invoice to the customer.
• Duplicate an order: Replicate an existing order and make adjustments as needed.
• Cancel/refund an order: Cancel or refund orders by setting quantities to refund, specifying shipping amounts to refund, providing reasons for cancellation, restocking items, and notifying the customer.
• Edit shipping address: Update the shipping address for an order.
• Insert product links: Add product links or product cards from tickets so customers can add the product to their cart quickly.
• Display the customer’s cart: View the exact items the customer has in their cart at the moment they reach out via Chat.

Use case: Agents can perform Shopify actions directly from Gorgias, such as adding products, applying discounts, updating quantities, or issuing refunds.

3. Embed customer-specific Shopify data in Macros

What it does: Create templated responses called Macros with dynamic Shopify variables to automatically incorporate customer-specific information. 

Key features:

• Dynamic variables: Macros can include variables that pull real-time data from Shopify, such as order status, tracking numbers, and customer details.
• Automated actions: Beyond inserting dynamic content, Macros can perform actions like tagging tickets, setting statuses, or assigning conversations to specific agents. The automation streamlines workflows and ensures consistent handling of similar inquiries.

Use case: A customer inquires about their order. With one click, the agent uses a Macro that pulls in the order status and expected delivery date, creating a faster and more personalized response.

Take Try The World, a gourmet subscription service, needed a robust Shopify integration to handle an increasing volume of customer inquiries. By switching to Gorgias, they gained the ability to unify conversations and embed Shopify data directly into Macros. Now, agents could quickly generate personalized responses that included order details, tracking links, and customer-specific information. 

Try the World’s support team’s efficiency skyrocketed, enabling them to handle 120 tickets per day, up from 80, and reduce response times to just one business day. 

4. Provide product information with Macros

What it does: Macros with embedded Shopify data let agents quickly and accurately share pre-sale information like product links, stock availability, and discount codes, helping to convert prospective customers into buyers.

Key features:

• Dynamic Shopify variables in Macros: Agents can use dynamic variables to pull real-time product information.
• Pre-built responses for common questions: Macros can include templated responses tailored for pre-sale inquiries, such as providing direct links to products or applying discount codes.

Use case: A customer asks if a specific product is available in their size and color. The agent can apply a Macro that automatically pulls the product's inventory details and includes a discount code, sending a response like this:

“Hi {{ticket.customer.firstname}},
Great news! The product {{ticket.customer.integrations.shopify.products[0].title}} is currently in stock in the size and color you’re looking for. You can check it out here: [Product Link]. Use the code WELCOME10 at checkout for 10% off your first order! Let me know if you have any other questions!”

How it helps:

• Eliminates manual search and typing for agents.
• Ensures accurate, real-time product information for customers.
• Improves the likelihood of converting inquiries into sales.

5. Enable self-serve order management in Chat 

What it does: Using Gorgias’s chat widget, customers can track orders or manage their purchases on their own with no agent assistance needed.

Key feature:

• Order management automation: Customers can access real-time order information, including status updates and tracking details, through the chat interface. This automation reduces the volume of live chat inquiries by up to 30%.

Use case: A customer wants to check the status of their recent purchase. By accessing the Chat widget on your website, they can enter their email and order number and receive instant updates on their order's progress, including shipping and delivery information, without waiting for an agent's response.

How it helps:

• Automates routine inquiries and frees up your support team to handle more complex issues.
• Enhances customer satisfaction thanks to immediate responses.
• Reduces the need for multiple communication channels, consolidating support interactions in one place.

6. Use Shopify variables in Rules


What it does: Rules paired with Shopify variables can automate various support tasks, such as identifying specific customer segments or tagging tickets, to boost efficiency and consistency.

Key features:

• Automated tagging: Rules can automatically tag tickets based on specific Shopify data. For instance, you can set up a Rule to tag tickets from customers with high order counts or significant total spending as "VIP."
• Prioritization of tickets: Rules can prioritize tickets that meet certain criteria, such as high-value orders or repeat customers.

Use case: A customer with a history of substantial purchases contacts support. A rule detects that the customer's total spending exceeds a predefined threshold and automatically tags the ticket as "VIP." 

This tag can then trigger other workflows, such as assigning the ticket to a senior support agent or escalating its priority.

How it helps:

• Improves customer experience by prioritizing high-value customers.
• Maintains consistent service quality.

7. Track revenue with reporting

What it does: Gorgias offers comprehensive reporting that allows you to measure how your support interactions influence sales.

Key features:

• Tickets converted: Tracks the number of support tickets that led to a sale within five days of the ticket's creation.
• Conversion rate: Calculates the percentage of created tickets that resulted in sales, helping you assess the effectiveness of your support team's interactions.
• Total sales from support: Sums the revenue generated from orders associated with converted tickets, accounting for refunds and order adjustments to provide accurate figures.

These metrics are accessible under Statistics → Support Performance → Revenue in your Gorgias dashboard. You can filter the data by integration, ticket channel, tags, or specific time periods to gain detailed insights.

Use case: By analyzing Revenue Statistics, you can identify which support channels or agents are most effective in driving sales. For example, if live chat interactions have a higher conversion rate, you might allocate more resources to that channel. 

Additionally, recognizing top-performing agents can inform training programs to elevate overall team performance.

For example, One Block Down, a Milan-based streetwear brand, struggled to manage a growing volume of customer inquiries across multiple platforms. By integrating Gorgias with Shopify, they centralized all customer interactions into a single platform, giving agents instant access to crucial information like order history and returns directly within tickets.

The setup allowed the team to measure the direct impact of their support efforts on revenue. 

The result? An impressive 1,000% increase in support-generated revenue and a 1-hour average first response time. By connecting the dots between customer service and sales performance, One Block Down demonstrated how proactive, data-driven support can directly influence the bottom line.

How it helps:

• Quantifies the revenue generated from support interactions.
• Faster team optimization with data-driven insights.
• Understanding the correlation between support interactions and sales can help refine customer service strategies.

‍

8. AI Agent integration

What it does: AI Agent automates Shopify actions like canceling orders, editing order details, and reshipping items.

Key features:

• Cancel Shopify order: AI Agent can automatically cancel unfulfilled orders upon customer request, restocking the items and issuing a full refund. A confirmation email is sent to the customer once the cancellation is complete.
• Edit order shipping address: When a customer needs to update their shipping address, AI Agent verifies if the order is unfulfilled, confirms the new address with the customer, and updates it in Shopify accordingly.
• Replace order item: AI Agent facilitates item replacements in orders by confirming the item to be removed and the new item to be added, checking stock availability, adjusting payments if necessary, and sending an updated order confirmation to the customer.
• Reship order for free: In cases where an order is lost in transit or arrives damaged, AI Agent can duplicate and resend the order at no additional charge.
• Remove order item: If a customer decides to remove an item from their order, AI Agent can handle the removal, restock the item in Shopify, process the refund for the removed item, and notify the customer of the updated order details.

Use case: A customer realizes they've entered an incorrect shipping address shortly after placing an order. They contact support, and AI Agent promptly verifies that the order is unfulfilled, confirms the correct address with the customer, updates the shipping information in Shopify, and sends a confirmation email—all without human intervention.

How it helps:

• Automating routine order management tasks reduces the workload on human agents.
• Quick and accurate responses to order modification requests lead to a better customer experience.
• Automated processes ensure consistency and accuracy in handling order changes, reducing the likelihood of human error.

{{lead-magnet-2}}