octa squared. The infrastructure around octa.
The model is one component. The infrastructure is what compounds. octa squared combines six algorithms running in parallel: online experimentation, ensemble signal ranking, multi-model rotation, sequence policy learning, model distillation, and the macro analysis loop.
Online experimentation Ensemble ranking Model distillation Macro analysis Multi-model rotation Sequence policy The model gets retrained on a release cadence. The process gets sharper every hour.
Most AI for sales pitches imply the model keeps learning in real time. It does not. Model weights change when training runs land. What changes continuously is the orchestration around the model.
One model component
octa retrains on a release cadence: continued pre-training plus long-horizon RL on the corpus. Weights change weekly to monthly, not per campaign.
Six algorithms running in parallel
Online experimentation, ensemble signal ranking, multi-model rotation, sequence policy learning, model distillation, and the macro analysis loop.
Every campaign generates data
Every send, reply, meeting, landing page click, and deal stage transition becomes validated learning per segment.
Back into training data
Validated learnings flow back into the next octa pretraining and fine-tune pass. The process feeds the model. The model feeds the process.
Great compounding systems were never one algorithm.
The systems that defined search, research, and game-playing were infrastructures combining many loops, with outcomes feeding back into the inputs.
Modern web search
Hundreds of weak signals are ranked continuously. No single signal decides. The ranking infrastructure, not any single signal, is the moat.
Self-play game agents
Search proposes moves, self-play generates new training positions, and the network learns from outcomes. The combination is what learns.
Online experimentation platforms
Variants compete for traffic. Outcomes are measured. Winners get more allocation per segment. The platform itself is the learner.
octa squared
Six GTM algorithms run around the octa model. Each loop validates inputs for the others. Validated learnings flow back into the corpus.
Six loops running in parallel. Each one sharpens the others.
The combination is the point: online experimentation tunes variables, ensemble ranking decides priority, multi-model rotation picks output, sequence policy decides the next step, distillation drops cost, and macro analysis watches the system.
Online experimentation
graph8 runs continuous controlled experiments on campaign variables: subject lines, send times, cadence depth, channel mix, landing page layout, and voice openers.
Ensemble signal ranking
Accounts, contacts, intent signals, and sequence variants are scored by combining many weak signals into one ranked output. Outcomes re-weight the signals.
Multi-model rotation
Generator, critic, and editor roles rotate across models per task. The winning output ships. The competition log feeds training.
Sequence policy learning
Each touch is a state. From each state there are many possible next actions. octa squared updates the state-to-action policy from real outcomes.
Model distillation
Frontier models generate canonical exemplars for hard GTM tasks. A retrieval pool serves those exemplars to cheaper open source students.
Macro analysis loop
Weekly and monthly reports span capacity, customer outcomes, algorithm performance, and segment shifts so humans can review and course-correct.
Each algorithm's output is another algorithm's input.
Six loops, but the loops are connected. Validated learnings from one loop become the starting material for another.
The infrastructure feeds the next training run.
The day-to-day loop is the infrastructure. Validated learnings, winning exemplars, surviving sequence policies, and re-tuned rankings flow back into the corpus that retrains octa.
The model release cadence is weekly to monthly. The infrastructure cadence is hourly. The two cadences feed each other.
Winning exemplars
Variable locks
Sequence policies
Distilled pool
Capture. Label. Validate. Train. Ship.
The five steps that turn yesterday's campaigns into next week's model.
Capture
Every campaign outcome, winning variant, state-to-action transition, and distilled exemplar lands in the corpus.
Label
The infrastructure tags segment, variable, channel, and intent without humans in the hot path. Humans review aggregates.
Validate
Held-out replay on octa Bench. If the new exemplar would have won historical campaigns, it survives.
Train
Surviving exemplars enter the next octa continued pre-training and long-horizon RL pass.
Ship
New octa weights deploy. The six algorithms now run with a sharper component. The next loop starts.
What octa squared is doing right now.
Live across every graph8 customer org. Every loop runs on its own cadence. The combination is what compounds.
6 algorithms
Running in production today across every graph8 customer org.
Hourly
The cadence where experiment results, rankings, and routing decisions update.
Weekly
The cadence where macro analysis re-tunes the infrastructure.
You pay for what graph8 executes. Never for seats.
Execution credits send, call, and run agents. Contact credits reveal and enrich the people you target. Three ways in.
A free entry point to the full platform. No credit card.
- 1,000 contact reveals to start
- 500 execution actions free
- Full platform access, no card
- Then $0.05 per credit
Unlimited contact data, in the tools you already use.
- Unlimited contact data, fair use
- 2,500 execution credits a month
- Priced per user, not per org
Unlimited users and data. Every app, one bill.
- Campaigns, dialer, AI voice, and AI workflows
- Meetings routed and booked for you
- 10,000 execution credits a month
- Unlimited users, no seat fees
- 7-day free trial, no credit card
Everything in the Team plan
$99 / mo, all includedThe $99 Team plan is the full platform: everything below, for the whole org. MCP Unlimited and Pay as you go are lighter entry points with usage and seat limits.
Your GTM, executed
10,000 credits included- Campaigns and sending infrastructure Mailboxes, domains, and deliverability run for you, end to end.
- 5-channel sequencer Email, phone, LinkedIn, SMS, and WhatsApp in automated sequences.
- Dialer and AI voice agents Clone any rep voice for inbound and outbound calls with live booking.
- AI inbox Replies triaged, drafted, and answered across every channel.
- Meeting routing and booking Route leads to the right rep and book the meeting automatically.
- CRM, pipeline, and quotes Deals, quotes, and AI-prioritized actions in the AE Cockpit.
- Studio content engine Brand, newsletters, landing pages, ads, and campaign content.
- AI agents and workflows Operator and Executor agents for RevOps, deals desk, and SDRs.
Data without limits
Included- MCP Unlimited for every user A $25 per user value, included for the whole org on the Team plan.
- No monthly cap on contact data Reveals, lookups, enrichments, and exports run org-wide, every seat.
- Buying signals Visitors, intent, hiring, and job-change signals on every account.
- Waterfall enrichment Third-party providers, pay on match or bring your own keys.
- Free search and browse Your whole team explores 700M+ contacts and 100M+ companies.
Org-wide by default
Always on- Unlimited users, no seat fees Invite everyone. Pricing never scales with headcount.
- Every feature unlocked No tiers gating voice, enrichment, sequencer, campaigns, or desktop.
- Built-in CDP and integrations Warehouse, 500+ connectors, two-way CRM sync included.
- Month-to-month, no contracts Cancel anytime.
- Fair-use rate limits Limits apply on all plans. They guard against scraping, not real GTM work.
Execution credits send, call, and run agents. Contact credits reveal and enrich the people you target.
Bring your motion. Watch the infrastructure spin.
See which variables lock, which model wins each task, which signal gets ranked up, which sequence policy fires, and what the next loop ships.