Gong
Each of these is a real decision Gong is working through, written so you can make the call yourself — incomplete information, someone who pushes back, no answer key.
Gong is an enterprise B2B SaaS provider of revenue-intelligence and sales-execution software that captures and analyzes customer interactions and CRM data. It serves sales reps, managers, revenue operations, enablement, and senior revenue leaders, focusing on deal risk, forecast predictability, coaching, and reduced administrative work. Gong's platform processes millions of customer interactions every day. Gong runs a data pipeline that ingests and captures customers' interactions from external communication systems such as emails, calendars, phone and web conferencing calls, and texts. Gong runs a multi-cloud infrastructure on AWS and Azure. Gong's production environment powers hundreds of Java microservices and ML model-serving workloads. Gong builds backend services and APIs using Java, Spring Boot, and cloud technologies. Gong builds an internal agentic framework that leverages and integrates tools such as LangChain, LangSmith, and ADK.
Your first one is free. No card.
11 live right now.
The rest
Build a value programme for sales managers so sponsors see what changed
Pilot customers can see deal signals but cannot yet tie them to better forecast inspection. You must weigh a simple programme customers will adopt against evidence executives will trust.
Commit the renewal of automated plays once privacy sees human review
The customer has six weeks to decide whether automated plays stay in its forecast workflow. You have to weigh measurable adoption gains against evidence that leaders can defend to privacy and finance.
Rebuild trust in deal boards while forecast reviews continue
Sales managers are abandoning deal boards when opportunity fields disagree with Salesforce. You must weigh a fast workflow reset against the controls that keep CRM data dependable.
Weigh AI briefs against fixes to Kafka correlation
Missing interactions leave sales managers reviewing active deals with incomplete evidence. You must weigh faster investigation briefs against the proof engineers need before acting on live capture.
Flag deals with thin evidence without mislabeling late-ingested ones
Managers cannot tell whether a quiet deal lacks buyer activity or lacks captured evidence. You must weigh a useful early signal against the chance that delayed ingestion mislabels a deal.
Correct duplicate emails before forecast reviews dispute the evidence
Duplicate email activities are appearing in active deal histories before the next forecast review. You must weigh fast correction against preserving the evidence needed to prove every change is safe.
Recover call identities while meetings keep flowing
Completed customer meetings are arriving without attaching to the right deal activity. You must weigh fast recovery against duplicate records and incorrect account history.
Surface stalled deals without mistaking missing activity for health
Managers open many deal records before finding a stalled opportunity, and activity capture is uneven across accounts. You have to weigh broad ranking coverage against evidence standards that keep coaching priorities credible.
Stop account mixing in resumed sessions before the rollout commits
Five resumed sessions may have crossed account boundaries after failover, though the traces are partly inconclusive. You have to weigh a broad context shutdown against preserving forecast usefulness while every suspected exposure is verified.
Fix the adapter for CRM data while the Java contract stays changed
Successful grounded CRM tool calls fell to 91.8% after a response-contract change. You have to weigh a narrow adapter repair against a broader Java compatibility path that consumes the same release window.