How to scale customer support on Telegram and Discord: Insights from Privy and Alchemy

Many web3 and crypto companies now offer support over channels like Discord and Telegram. But as your customer base grows, Discord communities, Telegram groups, and Slack channels create fragmented support experiences that are difficult to manage at scale.

In our recent webinar, Chad Doan (Ops at Privy) and Bastien Moyroud (Head of Solutions Engineering at Alchemy) shared how they've built scalable support operations that meet customers where they are. Both companies support thousands of developers and have evolved their support strategies as they've grown.

Key takeaways from the webinar

1. Web3 customers prefer different communication channels than traditional B2B buyers

One of the biggest differences in web3 support is channel preference. While enterprise customers are typically comfortable with Slack or email, many web3 teams prefer Telegram on principle.

As Chad explained: "There's a contingent of web3 customers that just don't trust Slack. They just don't trust these legacy products, if you will, that are more catered towards enterprise. I think when Telegram was launched, it had this motto of decentralization and transparency, that I think something like Slack just inherently doesn't have."

Cost also plays a role. Chad noted: "Telegram can be free for companies to use, and with Slack that might not necessarily always be the case. We even have, today, people saying: I don't want to sign up for Slack Connect."

This means web3 support teams need to be flexible and meet customers on their preferred platforms — whether that's Telegram, Discord, or Slack.

2. Developer communities drive product feedback and customer relationships

Both Privy and Alchemy started with large developer communities before implementing formal support systems. Privy's developer Slack community now has over 8,000 members, while Alchemy built an active Discord community for educational content and support.

These communities served multiple purposes, beyond just answering customer questions. Bastien explained Alchemy's approach to VIP Telegram channels: "We'd build a relationship with these customers. They would feel like they were almost talking to their friends and have VIP treatment. We'd also get feedback on the product, and we'd be able to solve their issues really quickly."

The personal touch mattered too. Bastien shared: "Our entire team was there. Our cofounders were there. Our sales team was there. Engineers were there."

As companies scale, maintaining this community feel while professionalizing support operations becomes a key challenge.

3. Consolidation becomes critical as you scale beyond 30 support channels

Both speakers agreed that around 30 active support channels, you hit the inflection point where manual processes break down.

Bastien specifically called out this threshold: "I'd say around the 30 mark. At some point, it just gets too hectic."

Chad offered a more qualitative indicator: "As soon as ownership is sort of unclear, it might be time that you think about implementing a support system. At the end of the day, if you're looking through your messages and you're like: Wait, I don't remember who's working on this! I think it's probably a sign that something is broken there."

Before implementing Pylon, both teams struggled with fragmented workflows. Chad told us how Privy manually managed Telegram support before Pylon: "We had people screenshotting what they saw on Telegram and sharing an image of that into our internal Slack, and saying: Hey, can someone jump in and respond?"

4. Threading and assignment features prevent issues from falling through the cracks

One of the most valuable features both teams highlighted was the ability to assign tickets and track ownership across channels.

Chad explained: "We know we all live busy lives. We're inundated with Slack notifications, and things can sort of fall through the cracks that way. I think the thing that Pylon does well to solve that is, you can actually assign tickets to people. They can look back at it. They can escalate things internally, create internal threads."

Bastien also called out internal threading as the most underrated feature: "The amount of times that we had to be like: Oh, where did you post about this? Or, where did you ask Andrew about this? Or, have you asked this? Now, we have all this in the same UI. It seems silly, but it's been such an improvement in having customer conversations. I have my internal threads, I have internal threads with the team, and also my internal debugging thread."

5. AI features help teams maintain quality at scale without replacing the human touch

Both companies are leveraging AI to accelerate their support teams, not replace them. The focus is on speed and consistency rather than full automation.

Chad shared how Privy uses AI for product feedback loops: "We recently launched hCaptcha, which is a new offering to mitigate bot activity on Privy apps. And we literally went to Pylon and said: Hey, in the past ‘X’ amount of time, who's asked about this feature?"

Bastien emphasized how Alchemy maintains the human element in support: "One of the differentiated value adds that Alchemy has is the human touch, the ability to talk to our engineers directly. That doesn't go away." They're using AI to automatically kick off investigations when tickets come in, saving their team time on information gathering.

Bastien also highlighted simpler AI improvements: "We tag tickets, like what blockchain it was on, what request method, that kind of stuff. In our own system, before, we relied on the team to tag them. It was lot of: Hey guys, please don't forget to tag them correctly. Now AI basically does it for, I'd say, 90% of tickets."

Questions from the audience

How does Pylon handle threading on Telegram when the platform doesn't natively support it?

This was a major concern for attendees, since Telegram's lack of threading is a well-known limitation.

Chad shared his experience: "It doesn't introduce native threading within Telegram, obviously. But when issues come through to Pylon, I would say that Pylon AI or the underlying logic does a pretty good job of separating issues based on the content. So our Telegram threads are — it's just a dump, right? You have technical questions, you have marketing questions, you have sales questions. And I'd say that Pylon does a really good job of separating those into separate issues."

We use AI to bundle related messages into topic-based issues, giving teams a structured view even when Telegram itself doesn't support threading.

Why did you choose Slack over Discord for your developer communities?

For Privy, Discord’s voice-based platform just didn’t scale well for customer support.

Chad explained: "I think the decision was based around Discord being more of a voice-based community, with an emphasis on real-time communication. Like 10-person voice calls, associated normally with gamers. To us, it didn't really scale well. I mentioned that our developer Slack community now has 8,000 people in it that just does not translate to live voice calls."

As for Alchemy, they found that Discord messages were easy to get lost in — which became a problem when they were trying to meet fast response time standards for B2B support.

According to Bastien: "We eventually said: Hey, let’s just keep this for community and get the community excited to share new tutorials, then help where we can. But when you truly do support, we're beholden to response times. We wanna get back to teams as quickly as possible. It's very easy from our perspective to get lost in the sauce on Discord."

What makes Slack easier to use than Telegram for support teams?

Bastien identified three key advantages of Slack:

1. Threads: “You can't thread in Telegram. Well, you kind of can, but it's a nightmare."
2. Cross-team collaboration: “Not everyone in our team has Telegram. I'm thinking specifically on the engineering side: If we have an issue and we're debugging it, we're having to go back and forth. I'd rather just say: Hey, let's add the engineering team there so they have visibility."
3. Integrations, specifically how easy they are to manage: “The ability to even build with the API is unmatched... Either external tools we can hook in or build ourselves. It's just a blessing to work with Slack’s API relative to the other ones."

How do you use AI to improve your product feedback loops?

Chad explained how he uses Pylon’s Ask AI to understand which customers asked for features, and who to reach out to once Privy launches something new: "The AI features have been super helpful in our general product feedback loop. I'm thinking specifically of the Ask AI feature, which is kind of like talking to an LLM. You can just ask it any question, except the benefit is that it can read all of your Slack messages all at once."

He elaborated on how Privy uses Ask AI for feature launches: "Whenever Privy launches a new feature, we want to make sure that we're reaching out to every customer that's asked about this in the past. The Pylon Ask AI feature is a really good way for us to do that."

Plus, Chad highlighted why this improves the customer experience: "I think if you're a customer, to hear back when in the past you've expressed interest for a certain feature — it’s a really good customer experience. To really know that your voice is heard, and your asks are being tracked and ticketed, and things are getting followed up on."

Build support systems that scale with your company

Privy and Alchemy’s stories show that scaling support in web3 requires flexibility, consolidation, and implementing AI for the right workflows.

The key is finding a balance between the personal, community-driven feel that web3 customers want, and implementing the processes and tools to support hundreds or thousands of accounts.

Want to learn more about scaling customer support across Telegram, Discord, and Slack? Connect with our team to see how web3 and crypto companies are doing it with Pylon.