Introducing Context-Aware Follow-Up Suggestions

Today we're releasing Context-Aware Follow-Up Suggestions — a feature that tracks the flow of your conversation and proactively suggests what to say or ask next, before there's an awkward silence while you think of it yourself.

What Changed

Until now, Neothi was reactive. You'd hear a question, and the overlay would surface information to help you answer it. That's useful, but it only covers half the conversation. The other half — the moments where you need to drive the discussion — was on you entirely.

With context-aware follow-ups, the overlay now suggests follow-up questions you could ask, talking points that build on what was just discussed, and natural transitions to steer the conversation toward your strengths.

Here's what that looks like in practice. You're in a behavioral interview. The interviewer asks about a time you handled a disagreement with a teammate. You give your answer. There's a beat of silence. Previously, you'd either wait for the next question or scramble to think of something to add. Now, the overlay might suggest: "Mention the process change that came out of that disagreement — ties back to your leadership narrative" or "Ask: 'Would you like me to walk through how we changed our review process after that?'"

The difference is subtle but real. Instead of the interviewer controlling the entire flow, you have natural entry points to highlight things that matter.

How It Works

The feature maintains a sliding context window of the last several minutes of conversation. It does three things simultaneously:

Topic tracking. It identifies what's currently being discussed and detects when the topic is shifting — the transition between "tell me about your background" and "let's talk about a technical challenge" is a moment where a well-placed comment can redirect the conversation.

Gap detection. It cross-references what's been said against your profile — resume, past projects, key metrics you've uploaded — and identifies relevant points you haven't mentioned yet. If you talked about leading a migration but didn't mention the performance improvement numbers, it'll nudge you.

Suggestion ranking. Not every possible follow-up is worth surfacing. The system ranks suggestions by relevance to the current topic, how naturally they'd fit into the conversation flow, and how strongly they connect to your prepared talking points. Low-confidence suggestions are suppressed entirely.

Suggestions appear in a dedicated section of the overlay, visually separate from answer assistance. A quick glance tells you the difference between "here's help with the current question" and "here's what you could bring up next."

Early Results

We beta-tested with 200 users over six weeks. The numbers:

• 34% of users asked at least one suggested follow-up question per interview
• Conversation flow self-ratings improved from 3.6/5 to 4.2/5 with the feature enabled
• Users reported feeling more in control of interview pacing — less "waiting for the next question" and more "steering the discussion"
• No increase in detection risk — suggestions are paced to match natural conversation rhythm

The most common feedback: "It's like having a coach whispering 'now mention the metrics' at exactly the right moment."

Availability

Context-aware follow-ups are enabled by default for all Pro and Enterprise users starting today. Starter plan users can enable the feature in Settings → Overlay → Follow-Up Suggestions. Works across all interview types: technical, behavioral, and case study.