Inspiration
We noticed sales teams don't lose deals because of bad leads, they lose them in the middle, through small execution failures like missed follow-ups and poor prioritisation. We initially explored intent signals, but users told us straight up: intent data isn't trustworthy, and they didn't want more noise. After speaking with reps, managers, and founders, the insight was clear, the real problem is execution.
What it does
Delta is an AI sales execution coach. It connects CRM, email, meetings, and notes into one live deal view, learns from won and lost deals to spot patterns and risks, flags what's slipping, recommends next steps, scores execution, and closes the loop by drafting responses and updating the CRM automatically.
How we built it
We coded the project together with the help of AI coding agents like Codex.
Challenges we ran into
Our initial direction focused on intent signals, but users pushed back hard, so we pivoted. Beyond that, we're navigating trust in AI recommendations, messy fragmented data across tools, and keeping our scope tight to avoid feature creep.
Accomplishments that we're proud of
We validated the problem with real users and gathered around 1,700 online views and discussions. Instead of forcing our initial idea, we listened, pivoted, and let user feedback directly shape the product.
What we learned
Teams don't need more data or more leads — they need a system that acts during execution. Users want high-confidence, actionable insight on real deals, not more noise.
What's next for Delta
We're starting in real estate, finance, consultative B2B, and service-heavy sales where deals are high-value and workflows are messy. From there, we'll expand into recruiting, account management, and customer success through founder-led outreach, design partners, and a land-and-expand model
Built With
- crm
- fastify
- llm
- next.js
- supabase
- typescript
Log in or sign up for Devpost to join the conversation.