NOTE: Demo visuals use either blurred real data or synthetic placeholders to protect customer privacy.
Replacing the Broken Relay
The current support flow sends each customer question through multiple people before it reaches a correct answer, introducing delay and inconsistency. Existing AI assistants offer speed, but they often invent or misstate facts. Andy’s update replaces this “broken telephone” model with grounded, traceable automation that behaves more like a trusted peer.
The old friction points are clear:
- Information gets outdated or fragmented across documents.
- Tone and phrasing vary between teams.
- AI responses often sound confident without proof.
Together these frictions slow deals and dilute trust, making every verified answer a small victory.
Deploying a Careful Assistant
Built on the Nora framework, the bot checks its work like an engineer—reading product code, verifying across multiple repositories, and abstaining when the evidence is weak. Human evaluations still shape its benchmark, but deployment in Slack now brings the system into daily GTM use.
Its current capabilities include:
- Scanning product code and live sources to ground responses.
- Cross-checking across multiple systems before replying.
- Abstaining when data conflicts or confidence is low.
- Logging SME feedback to refine performance.
- Expanding coverage to new product and roadmap domains.
In addition, a roadmap bot and Gong transcript integration are being tested to unify related question workflows. Both extend the core vision: one assistant that learns directly from product truth and customer context.
Driving Accuracy and Reach
Early usage shows the bot answering about 30 percent of GTM Help questions and reducing Slack dependency across sales and support. Teams see fewer engineering escalations and faster handoffs. Its verified approach builds momentum where uncertainty once stalled progress.
The outcomes speak clearly:
- Faster responses during sales and troubleshooting,
- Higher factual confidence across all product topics,
- Consistent tone aligned with customer messaging,
- A clearer path to proactive, direct-to-customer assistance.
Next, the team is reviewing tone and positioning for external readiness. Once validated, reps will be able to copy bot answers straight to customers.
Signals from the Field
Observers noticed how the bot’s precision inspired confidence. Watching an AI verify code like an engineer and decline when uncertain set a new standard for responsible automation.
This marks an important cultural turn. The GTM organization is learning to trust AI not because it answers everything, but because it knows when to stop.
