REBOOT THE WAY
WORK GETS DONE.

Most clients see measurable ROI by week 8, full payback in 4–6 months. Baseline is locked week 1, final ROI report lands week 12. Refund terms live on the pricing page.

Two safety nets. (1) Every agent runs in shadow mode for 2–4 weeks before going live — your human reviewer decides when it’s ready, measured against the baseline you define in week 1. (2) If we miss the ROI target, the deployment fee is refunded per the schedule on the pricing page. The same schedule appears verbatim in every SOW.

Your agents, your data, your IP. Every agent configuration, workflow, prompt, and integration we build for you is yours — fully documented and exportable. Your data never leaves your tenancy and is never used to train any public model. Forge (the platform) is licensed — it’s the orchestration, monitoring, evals, and safety layer your agents run on. If you cancel, you keep every agent definition and every bit of data; you just lose the runtime. We sign DPAs, NDAs, and source-escrow agreements on request.

Yes. Forge ships with native connectors for SAP S/4 & Business One, Microsoft Dynamics 365 (F&O and BC), NetSuite, QuickBooks, Sage Intacct, Salesforce, HubSpot, Slack, Teams, Gmail, Outlook, Snowflake, Databricks, and 60+ others. If you have something custom, we build the connector during week 5–6 of deployment — included.

Two things: (1) headcount band (it’s a proxy for transaction volume) and (2) the number of agents and integrations you need. Deployment is one-time and fixed. Monthly is recurring and covers the platform, agent inference, monitoring, and ongoing tuning. Extra agents are $800/mo, extra integrations $400/mo. No usage fees, no surprise bills.

No — it gives them their week back. Our clients use the freed-up hours to push into projects they never had time for: customer onboarding, supplier negotiations, data quality, new product launches. We’ve never had a client let staff go after deployment. The economic case is doing more with the same team, not the same with fewer.

You can — and many of our clients tried first. The hard part isn’t the LLM, it’s the orchestration: integrations, error handling, observability, retries, evals, role-based access, audit logs, and the change management to get a finance team to trust an agent posting to the GL. That’s what Forge is. A 2-engineer in-house build typically takes 9–12 months and stalls; we’re live in 90 days with the safety guarantees.

They automate within their own platform. We automate across your entire stack. Microsoft Copilot for Business Central is great — inside Business Central. SAP Joule is great — inside SAP. NetSuite Text Enhance is great — inside NetSuite. None of them can draft an invoice in Business Central, reconcile it against a vendor statement in QuickBooks, update the opportunity in HubSpot, and notify the buyer in Slack in a single agent run. That cross-system orchestration across the five-to-ten tools your business actually runs on is the entire point of Forge. Add to that the managed delivery model — we install it, integrate it, tune it, and run it every month — and you get outcomes in 90 days instead of a license you have to figure out yourself. We’re additive to Copilot, not a replacement for it.

Yes. We publish a redacted sample SOW showing scope, milestones, baseline methodology, ROI targets, and the refund schedule verbatim. Request the sample SOW pack → and we’ll send it back within an hour.

The team that built Forge. Every deployment is run by a dedicated delivery pod of 2–3 agent engineers plus an integration specialist, owned end-to-end by the same people who write the code. No offshore handoffs, no sales-to-CSM swap. That’s why we’re capped at 10 founding clients in 2026 — it’s the ceiling on what the team can deliver without cutting corners. Meet them in the team section →