Agentic GTM Hackathon 2026: Winners and Best AI Projects
What happens when 60 teams get one day to rethink how companies acquire, qualify, and expand customer accounts with AI agents? The Agentic GTM Hackathon brought builders together at Station F in Paris on July 9, 2026, to answer that question.
Across three tracks and five challenges, teams built systems for detecting buying signals, reviving closed-lost deals, preparing sales calls, enriching qualified contacts, and turning scattered CRM data into concrete actions. Gravity took first place overall, followed by Zorro and Net-Work. Other projects explored everything from personalized physical outreach to autonomous CRM maintenance.
These are the winners, standout submissions, and agentic GTM workflows that emerged from the event.
Agentic GTM Hackathon 2026 at a Glance
- Event: Agentic GTM Hackathon
- Date: July 9, 2026
- Location: Station F, Paris
- Submissions: 60
- Tracks: Acquisition, Expansion, and Free-for-all
- Overall winner: Gravity by Team Low Cortisol
- Second place: Zorro by Team Tomorro
- Third place: Net-Work
- Main use cases: Buying-signal detection, pipeline reactivation, pre-call research, contact enrichment, CRM maintenance, and personalized outreach
- Recurring workflow: Sillage detects a signal, Claude evaluates it, FullEnrich resolves contact details, and HubSpot receives the resulting data or action
What Was the Agentic GTM Hackathon 2026?
The Agentic GTM Hackathon took place at Station F in Paris on July 9, 2026. FullEnrich and Sillage helped run the one-day event, which brought together 60 teams to build practical AI systems for go-to-market work.
Participants competed across three tracks:
✅ Acquisition: 37 submissions focused on finding and converting new opportunities.
✅ Expansion: Seven submissions explored ways to grow existing accounts.
✅ Free-for-all: 16 submissions tackled broader GTM problems without a fixed use case.
Projects could also enter five challenges: Most Creative GTM, Best Use of Gamma, Best Use of Gradium, Most Viral on LinkedIn, and Most Viral on X.
Claude powered the reasoning layer behind most submissions, but teams used it in different ways. A common setup started with Sillage detecting a buying signal. Claude then assessed the opportunity and decided what should happen next. FullEnrich identified the right contact once the account qualified, while HubSpot stored the resulting data and actions.
The strongest entries went beyond generating another sales recommendation. They decided which accounts deserved attention, selected the next action, prepared the necessary content, and left room for human approval when a decision carried more risk.
Agentic GTM Hackathon 2026 Winners
Gravity, Zorro, and Net-Work took the top three places. Each project addressed a different weakness in the sales process, from cold outreach and disconnected GTM data to last-minute meeting research.
| Place | Project | Team | Main use case |
|---|---|---|---|
| 1st | Gravity | Low Cortisol | Building familiarity before sales outreach |
| 2nd | Zorro | Tomorro | Turning disconnected GTM signals into actions |
| 3rd | Net-Work | Net-Work | Preparing sales teams before customer meetings |
1. Gravity: Building Demand Before Outreach
Gravity won the hackathon with a different take on outbound sales. Instead of improving the message sent to a cold prospect, the system helps a company become familiar to that buyer before direct outreach begins.
Sillage identifies relevant signals, including job changes, hiring activity, promotions, and engagement with competitors. Gravity then builds a buyer taste profile covering the person’s interests, preferred tone, content habits, and trust triggers.
Claude-managed agents use this profile to prepare a warm-up plan. They determine what the seller should publish, where to leave comments, who to connect with, and when direct contact makes sense. Gamma converts the plan into content, while FullEnrich enriches the prospect only after the account becomes warm enough to approach.
Gravity entered all five event challenges: Most Viral on LinkedIn, Most Viral on X, Most Creative GTM, Best Use of Gamma, and Best Use of Gradium.
2. Zorro: Turning GTM Data Into Sales Actions
Zorro finished second with an intelligence layer designed to connect the tools already sitting around a CRM.
Most GTM teams collect data from prospecting platforms, enrichment providers, signal tools, and business intelligence systems. Each source may be useful on its own, but the information often stays isolated. Sales representatives receive more context without getting a clear answer on what to do next, while RevOps teams spend time maintaining workflows between separate tools.
Zorro brings these signals together and acts on behalf of the sales representative. Rather than adding another dashboard or producing a list of recommendations, it converts available data into concrete actions.
The project targets a common problem in modern sales stacks: companies gather large amounts of account intelligence, but much of it never influences an actual sales decision.
3. Net-Work: Delivering Better Pre-Call Intelligence
Net-Work took third place with a pre-call research agent for revenue teams. Before a scheduled meeting, Claude gathers information from three live data sources and creates a one-page brief.
The brief appears in the tools the sales team already uses, including the web app, Slack, or the calendar invitation. It helps the representative understand who controls the budget, who recently changed roles, and which existing relationships could provide a warmer path into the account.
Its main insight concerns customer referrals. A useful introduction may already exist through a current client, but the connection is often split between LinkedIn, CRM records, and colleagues’ personal knowledge. Net-Work surfaces that path before the meeting, when the sales representative can still use it.
Other Standout Agentic GTM Projects
The remaining submissions approached GTM automation from several directions. Some revisited opportunities already sitting in the CRM, while others looked for new ways to reach buyers who had ignored every digital channel.
Re:lay: Reopening Closed-Lost Deals
Re:lay monitors closed-lost opportunities for new buying signals. When something changes at the account, a Claude agent reviews the original loss, checks whether the situation has evolved, and decides whether the deal deserves another attempt.
If the opportunity qualifies, the system identifies the right person, writes a re-engagement play, and pauses for human approval. Sillage detects the signal, FullEnrich supplies verified contact details, and HubSpot handles the two-way CRM sync.
Gamma also creates a personalized “why now” deck for each play. Gradium adds a voice interface that lets users run or approve a play out loud.
Revibe: Prioritizing a Dormant Pipeline
Revibe also targets inactive and closed-lost opportunities, but places more emphasis on prioritization and budget control.
The system imports dormant accounts from HubSpot, then combines Sillage signals with external context from Exa and Linkup. It scores each account across three dimensions: Readiness, Willingness, and Priority.
Only accounts that pass the required score trigger paid contact enrichment through FullEnrich. Claude then prepares a personalized, costed, and owner-assigned action plan that fits the available budget. The approved plan can be pushed back to HubSpot in one click.
Knock Knock: Taking Outbound Sales Offline
Knock Knock steps in after emails, calls, LinkedIn messages, and ads have failed. It turns a new buying signal into a personalized physical gift sent to the relevant decision-maker.
Sillage identifies the reason to contact the account, and FullEnrich finds the person behind the decision. Claude researches public information to confirm a genuine personal interest before recommending the gift.
The package includes a QR code that calls the sales representative directly. Instead of sending another digital follow-up, Knock Knock tries to turn a final outbound attempt into an inbound conversation.
Ouija: Bringing Lost Opportunities Back Through Voice
Ouija also watches closed-lost accounts for signs that an opportunity may be active again. It scores each account to determine whether the deal is worth revisiting, then briefs the sales representative through a live voice séance.
The project competed in the Expansion track and gave the familiar pipeline reactivation workflow a more theatrical interface.
Sentinel: Cleaning Up Overdue CRM Tasks
Sentinel works overnight on the backlog of overdue tasks stored in a CRM. It reviews the unfinished work, scans the related accounts for relevant signals, and reschedules tasks based on what now deserves attention.
The system does not require the representative to reorganize the entire backlog manually. It prepares the updated schedule so the user can review and validate the proposed changes.
FOMO AI: Choosing the Right Internal Expert
FOMO AI decides who inside a company should contact a prospect at a specific moment. Depending on the situation, that person could come from sales, product, compliance, or the founding team.
The system drafts the email in the selected expert’s writing style and sends it to that person through Slack for approval. The message only goes out under their name after they confirm it.
This approach treats the sender as part of the sales strategy. Instead of assigning every opportunity to the same representative, FOMO AI chooses the person whose role and knowledge best fit the conversation.
How Agentic GTM Workflows Turn Signals Into Actions
Many submissions followed a similar foundation. They detected a commercial signal, evaluated whether it mattered, enriched the account, and sent the result back to the CRM. The main difference came from the action chosen after qualification.
1. Detect a Buying Signal
→ Sillage monitored events that could make an account worth revisiting. These included a job change, promotion, hiring activity, competitor engagement, or another change affecting the timing of an opportunity.
2. Decide Whether the Signal Matters
→ Claude acted as the reasoning layer. It reviewed the signal alongside the available account context and determined whether the sales team should respond.
This step prevented every detected event from automatically becoming another task. Re:lay checked whether the reason behind a lost deal had changed, while Revibe scored accounts before spending money on enrichment.
3. Identify the Right Contact
→ FullEnrich resolved contact information once an opportunity met the project’s qualification criteria. Several teams delayed this paid step until they had enough evidence that the account deserved attention.
4. Select and Prepare the Next Action
→ Each project took the workflow in a different direction. Gravity created a warm-up plan, Re:lay prepared a re-engagement play, Knock Knock recommended a physical gift, and FOMO AI chose an internal expert to contact the prospect.
5. Keep a Human in Control
→ Some systems paused before taking an external action. Re:lay required human approval before sending, while FOMO AI asked the selected expert to approve the drafted message in Slack.
A similar workflow can be tested during a focused corporate hackathon. Iterate helps companies bring technical teams together to build and evaluate working AI concepts around real business problems.
What the Best Agentic GTM Projects Had in Common
The strongest submissions did more than generate emails or summarize CRM records. They addressed a specific decision inside an existing sales process.
They Started With Timing
Gravity, Re:lay, Revibe, and Knock Knock all relied on a change at the account before recommending outreach. The signal gave the sales team a reason to act now instead of launching another generic sequence.
They Qualified Before Enriching
Contact enrichment did not always happen at the beginning of the workflow. Gravity waited until the buyer had warmed up, while Revibe only enriched accounts that passed its score threshold. This kept paid data usage focused on opportunities that had already shown potential.
They Worked With Existing GTM Tools
Several teams built around systems that revenue teams already use. HubSpot stored and synchronized CRM data, Slack handled approvals or delivery, and calendar invitations carried pre-call information.
The agent connected these tools and moved information between them rather than forcing the user to work from a separate interface.
They Produced a Defined Next Step
Each project ended with a clear output. That could be a warm introduction, an approved re-engagement play, a physical package, a reorganized task list, or a one-page meeting brief.
This made the systems easier to evaluate. Teams could judge whether the output helped someone complete a real GTM task instead of measuring a broad promise of better productivity.
They Preserved Human Approval Where It Mattered
Re:lay and FOMO AI did not send external messages as soon as the agent produced them. They placed a review step between the recommendation and the final action.
The event showed that agentic automation does not have to remove people from the process. It can research, prioritize, and prepare the work while leaving the final customer-facing decision to a human.
Build Your Own AI Hackathon With Iterate!
The Agentic GTM Hackathon gave teams one day to turn familiar sales problems into working AI projects. The submissions focused on concrete workflows, from preparing calls and reviving lost deals to prioritizing accounts and approving outreach.
Does your company have an AI use case that deserves more than another planning meeting? Run a corporate hackathon with Iterate and give your teams a focused environment to build, test, and present practical solutions.
Conclusion
The Agentic GTM Hackathon 2026 showed what sales agents look like when they are built around a precise task. Gravity prepared buyers before outreach, Zorro turned disconnected signals into actions, and Net-Work surfaced useful context before meetings.
Across the wider field, teams explored pipeline reactivation, account prioritization, CRM maintenance, internal expert selection, and physical outreach. The best projects shared a simple trait: they used AI to move a real GTM workflow forward.
Frequently Asked Questions
When did the Agentic GTM Hackathon 2026 take place?
The event took place at Station F in Paris on July 9, 2026.
Who won the Agentic GTM Hackathon?
Gravity, built by Team Low Cortisol, finished first overall. Zorro took second place, followed by Net-Work in third.
How many projects entered the hackathon?
The event received 60 submissions across three tracks: 37 in Acquisition, seven in Expansion, and 16 in Free-for-all.
What did teams build during the event?
Teams built AI systems for tasks such as detecting buying signals, preparing sales calls, reactivating closed-lost deals, enriching contacts, organizing CRM tasks, and planning personalized outreach.
Which technologies appeared across the submissions?
Claude served as the reasoning layer in most projects. Teams also used tools including Sillage for buying signals, FullEnrich for contact enrichment, HubSpot for CRM synchronization, Gamma for content creation, and Gradium for voice interfaces.