Challenges
Together AI's previous support platform wasn't built for technical B2B products, causing lost tickets, fragmented information, and limited visibility into customer health across the entire lifecycle.
When Sarung Tripathi joined Together AI as VP of Customer Experience, he discovered the team was using a platform originally designed for B2C chatbot interactions to manage enterprise technical support. As the customer base grew and technical complexity increased, critical operational blockers emerged.
Tickets were getting lost at scale, and the platform lacked essential integrations for technical products. There was no way to pull observability data from tools like Grafana to inform support decisions or credit strategies based on customer uptime metrics.
The platform operated as a simple ticket queue without the necessary, comprehensive customer intelligence needed to manage emerging AI-native companies and complex enterprise relationships proactively. Knowledge management was scattered across different tools, and information silos between support and customer success meant scheduling 30-60 minute meetings to transfer context manually.
"[Our previous support platform] simply wasn't built for a technical product. The robustness of feature set is not there for B2B, especially for technical products that have deep observability integration challenges."
Sarung Tripathi, VP of Customer Experience
Solutions
Pylon delivered the unified, AI-powered command center that Together AI needed to manage enterprise customer relationships at scale.
Together AI evaluated multiple platforms, but Pylon emerged as the clear choice based on three factors: Slack-native architecture for their communication-heavy workflows, rapid feature delivery that matched Together AI's own velocity, and a product vision centered on AI-powered customer intelligence.
Unified multi-channel operations
With 80% of support happening in Slack, 15% in Discord, and 5% across other channels, Together AI needed a platform that could consolidate communication without forcing engineers to context-switch between tools.
Support engineers now work from a single Kanban view in Pylon, handling tickets regardless of origin channel. Solution architects communicating with customers during evaluations have all interactions automatically tracked, eliminating manual reporting while giving leadership complete visibility into deal progress.
"If you walked into our office right now, you would see support engineers just have that kanban view of Pylon on their thing, and they're handling tickets. It doesn't matter where it's coming from, it's in Pylon, and they're answering it."
Sarung Tripathi, VP of Customer Experience
AI-Powered Customer Intelligence
Pylon's AI capabilities transformed how Together AI understands and responds to customer needs. Rather than replacing human judgment, the AI accelerates the team's ability to surface insights and take action on what matters most.
Account-level AI summaries provide instant context on customer health, recent issues, and interaction history. The team customizes prompts for different use cases, tailoring summaries for executive reviews, support triage, and cross-team handoffs so each stakeholder gets exactly the context they need without scheduling meetings or digging through tickets. This eliminated 10 hours per week (per agent) previously spent on manual context gathering across the team.
Monthly executive reviews now run entirely from AI-generated reports that pull customer health scores, ARR data, recent ticket activity, and status updates automatically. For at-risk accounts, AI extracts specific technical issues that require exec-level attention, eliminating 10 hours per month of manual data aggregation and reducing executive review prep time by 90%.
The philosophy resonates with Together AI's needs: AI should accelerate human expertise in solving complex technical problems, not automate away the judgment required to manage enterprise relationships.
"The AI summaries let us move from identification to action immediately. Instead of scheduling meetings to understand what's happening with an account, everyone has the context they need to start solving problems right away."
Sarung Tripathi, VP of Customer Experience
Proactive customer success through Account Intelligence
Account Intelligence fundamentally changed how Together AI manages customer relationships by shifting from reactive problem-solving to proactive optimization. The platform continuously aggregates signals across tickets, Slack conversations, observability data from Grafana, and customer interaction patterns to automatically surface opportunities for improving customer outcomes.
For Together AI's complex technical customers, this visibility captures the full customer journey beyond support tickets. The platform identifies deployment optimization opportunities, unresolved technical blockers, or engagement patterns that signal when customers need additional support. These insights enable the team to intervene early and drive measurable improvements.
"Based on the data Pylon was showing us in our executive review, we identified things we could fix in this customer's deployment. Within 24 hours, we put together a tiger team to deliver an optimized API endpoint, ultimately delivering them between 2-3x performance improvement."
Sarung Tripathi, VP of Customer Experience
Account Intelligence transforms customer operations from reactive support to proactive relationship management, enabling Together AI to continuously optimize and grow their most valuable customer relationships.
Results
Together AI achieved measurable improvements in retention and revenue while transforming how support and customer success teams operate.
The migration to Pylon delivered immediate impact across the business. Within months, Together AI saw meaningful improvements in both gross and net revenue retention as Account Intelligence enabled the team to identify and resolve churn risks before they escalated. The platform's unified view of customer health and proactive alerting fundamentally changed how the team protects and grows their most valuable accounts.
- Notable improvement in GRR and NRR through proactive churn identification and rapid response
- 24-hour average response time from identifying at-risk accounts to deploying technical solutions
Operationally, the transformation was equally significant. Cross-team handoffs that previously required long meetings now happen instantly through shared context in Pylon. Support and customer success teams redirected hours previously spent on manual data gathering and context switching to solving actual customer problems. The unified platform eliminated friction across the entire customer operations workflow.
- 90% reduction in handoff time from 30-60 minute context-sharing meetings to instant access
- 10 hours saved per week across support and customer success teams, redirected to solving customer problems
- 5x faster response time with unified view eliminating channel switching
- 100% of customer interactions automatically tracked across Slack, Discord, and other channels with zero manual logging
The impact extends beyond efficiency metrics. Support engineers now spend their time debugging customer AI infrastructure challenges like time-to-first-token optimization and inference latency rather than navigating fragmented systems. Solution architects focus on technical evaluations and closing deals instead of updating multiple platforms. Customer success leaders run data-driven executive reviews without manual reporting.
"Every minute they have to spend figuring out some SaaS platform is a minute they're not spending debugging why customers' time-to-first-token or inference latency on their AI system is too high. That's where our customers need them spending time. Because of what you guys offer, that is where they can allocate their time."
Sarung Tripathi, VP of Customer Experience
Pylon delivered on Together AI's vision for a customer command center where any information about any customer is immediately accessible to anyone who needs it. The platform serves as both the tactical tool for daily support operations and the strategic intelligence layer for executive decision-making. For a company serving customers with the industry's most complex AI infrastructure challenges, this proves that the right platform transforms customer operations from reactive ticket management to proactive, data-driven relationship management at enterprise scale.