OpenAI demonstrates Spark, a ChatGPT-powered workspace agent that autonomously researches, scores, and e-mails inbound sales leads while updating the CRM.
Watch a guided walkthrough of an agent that qualifies inbound leads, drafts personalized follow-ups, and keeps the CRM up to date.
TL;DR: A walkthrough of Spark, a workspace agent that automatically researches, scores, and personally e-mails every new inbound lead while keeping the CRM updated.
## From Prompt to Agent in Seconds
Matthew, the builder, opens ChatGPT and types a plain-English request:
> “Research every new lead → score them against our qualification rubric → send the first outreach e-mail → draft follow-ups → set reminders so nothing falls through the cracks.”
ChatGPT instantly drafts a step-by-step plan, lists the required tools (Gmail, web search, CRM), and surfaces any missing configuration. After Matthew confirms, the agent—nicknamed Spark—is created and ready to run.
## Live Demo: Qualifying AGI’s Mark Reed
A real-world lead arrives through the website contact form: Mark Reed from AGI Inc. asks how ChatGPT Business differs from ChatGPT Enterprise and how a sales team would use either product.
1. **Background job triggers**
Spark polls the form endpoint on a schedule; for the demo it’s started manually.
2. **Research & enrichment**
The agent searches public sources to enrich the contact and company record.
3. **Lead scoring**
It applies the SMB sales team’s qualification rules and assigns a score.
4. **First-touch e-mail**
A personalized message is composed and sent through Gmail within seconds.
5. **Queued follow-ups**
A second e-mail is pre-drafted and saved in Gmail drafts; a reminder task is scheduled for the next touchpoint.
## Key Takeaways for Sales Teams
- Natural-language prompting is enough to spin up a repeatable workflow.
- The agent can operate autonomously in the background, syncing with existing tools.
- Every step—research, scoring, copywriting, scheduling—is transparent and editable.
Spark is one example of the workspace agents teams can build on ChatGPT to speed up lead response and improve meeting-book rates.
Source: https://www.youtube.com/watch?v=HJlME6S-LJc
OpenAI built an AI-powered inbound sales assistant using its own APIs to handle surging lead volume for ChatGPT Enterprise and Business, achieving 98% response accuracy through a rep-feedback training loop. The tool personalizes responses in prospects' native languages and seamlessly hands off qualified leads to human reps.
OpenAI presents ChatGPT applications for sales teams across prospecting, discovery, meeting preparation, outreach, proposals, deal management, and objection handling.
OpenAI introduces ChatGPT agent, a unified agentic system combining web browsing, deep research, and conversational intelligence to autonomously handle complex multi-step tasks. The agent can navigate websites, analyze information, run code, and create deliverables while keeping users in control with permission requests.
Genspark launched Super Agent, a no-code AI agent platform powered by GPT-4.1 and OpenAI's Realtime API, enabling users to automate complex real-world tasks like phone calls, slide creation, and video generation. The product reached $36M ARR in 45 days after Genspark pivoted from AI search to agentic AI in April 2025.
OpenAI launches workspace agents for ChatGPT Business and Enterprise plans, allowing teams to build and share AI agents that autonomously handle workflows, integrate with tools, and operate under admin-controlled permissions and audit logs.