Stop clicking. Start shipping.
MaxAI is the AI operator for NetSapiens. Onboard customers in minutes, not days. Handle support tickets without dropping the call. Run bulk edits that would take a tech all afternoon — in a single prompt, with a full audit trail.
No credit card for the trial. Bring your own NetSapiens credentials and start operating within 15 minutes.
Your techs are clicking.
A new 50-seat customer is six-to-eight hours of portal work. A caller-ID change for a territory move is an afternoon. An auto-attendant with five menu options and a professional greeting is another thirty-five minutes. Every one of these is a rote, portal-shaped task that a senior VoIP engineer has to babysit instead of doing something that actually grows the business.
50 subscribers, 40 devices, 2 AAs, 2 queues, domain config, caller IDs, e911, greetings, testing.
Territory moves, re-branded CIDs, refreshed e911 after the annual audit. One click per user, 50+ users.
Subscriber + dialplan + 5 control rules + menu rules + answer rule + greeting record + upload + test.
One operator, all your tools.
MaxAI isn't a wrapper around NetSapiens. It understands the primitives deeply — dialplans, answer rules, AAs, queues, timeframes, voicemail, recording, DIDs — and wires in carrier APIs, domain-copy, and TTS alongside.
- ▸Subscribers, devices, AAs, queues, timeframes, answer rules, dialrules
- ▸Read, write, confirm-to-commit — not guess-and-hope
- ▸Every write gated behind a YELLOW or RED confirmation modal
- ▸Export a working customer, save as a template
- ▸Deploy to the next customer with one click — subscribers, AAs, audio, dialplans all copied
- ▸Edit templates in chat OR the visual editor (inline table with save/delete)
- ▸OpenAI TTS for AA greetings and voicemail greetings
- ▸Per-domain voice preference (Onyx, Nova, Shimmer, etc.)
- ▸Whisper auto-transcribes existing greetings so you can see what’s really there
- ▸Bandwidth, Commio built-in — search, order, release, port-out
- ▸DID inventory across every customer in one query
- ▸DID reroutes confirmed atomically with pre-state snapshot
- ▸Every tool call logged with user, tenant, risk level, pre-state snapshot
- ▸Sensitive disclosures (voicemail PIN) require operator-captured reason
- ▸Rolled-back operations tracked separately
- ▸Tokens stream as the model produces them
- ▸Tool-call progress strip shows what’s running in real time
- ▸Verified-state result cards appear inline when writes commit
Things you can't do in the portal.
The NetSapiens UI was built for click-by-click config. These examples are one prompt in MaxAI and compounding hours saved every month.
Turn a template + a CSV into a live customer.
- ▸Template carries the proven config (AAs, queues, dialplans, timeframes, audio).
- ▸MaxAI diffs the template roster against your CSV and patches every subscriber in one atomic transaction.
- ▸CIDs formatted, e911 set, AA greeting written and synthesized — all confirmed once, all audited.
“Deploy the "restaurant-standard" template to a new domain for Verdant Bistro. Use the CSV I'm pasting for the user list — replace every subscriber in the template with these names and extensions. Generate caller-IDs from their respective direct numbers. Deploy a professional AA greeting matching their business hours schedule.”
Build a full auto-attendant in one prompt.
- ▸One tool call builds: AA subscriber row, dialplan, 5 control dialrules, 4 menu option dialrules, answer rule.
- ▸OpenAI TTS generates the greeting audio in your saved domain voice.
- ▸Atomic upload — audio + stored script updated together so the NS UI shows the correct transcript.
“Create AA x691 for Acme Legal. Options: 1 for Intake (220 Sarah), 2 for Billing (221 Marcus), 3 for Directory, 4 for After-hours voicemail. Professional greeting mentioning all four options. Deploy with the onyx voice and wire invalid-input + digit-timeout to the same voicemail.”
Handle the support call without dropping it.
- ▸Voicemail auto-transcribed on read — no “want me to transcribe it?” loop.
- ▸Block lands on the correct answer rule (NS stores these in a non-obvious place; MaxAI gets the wire format right the first time).
- ▸PIN reveal is RED-gated: operator types a reason in the modal, audit log records everything except the PIN value itself.
“What does extension 221's voicemail actually say? Block the number the caller just identified as spam. Reveal 221's voicemail PIN — the customer forgot it and I've verified via callback.”
Update e911 emergency caller ID for every real user across a domain.
- ▸Scope disambiguation happens in-chat — “all extensions” means real users, not system accounts (MaxAI asks if unclear).
- ▸One batch transaction updates 50+ subscribers AND the domain default.
- ▸Result card shows the full extension list so you can sanity-check before you walk away.
“Update the emergency caller ID for every real user (not AAs or queues) at htel.23281.service to 6168280099. Also set the domain-level default to match in case a new user gets added.”
The math at one customer per week.
This is conservative. Most operators underestimate how much time disappears into the portal because it's spread across every tech, every ticket, every week.
That's 15+ forty-hour weeks of tech time, back on your calendar.
Or: pay for MaxAI more than eleven times over with a single tech's reclaimed hours — and that's assuming only ONE tech uses it.
Assumptions: 50-seat customers with 2 AAs / 2 queues / caller-ID + e911 setup; 10 support tickets/day/tech at 5 min saved each; one bulk audit per month; 40 new AAs / year across your base. Scale with your actual numbers — the gap widens for larger customers.
Pricing that pays back week one.
Three tiers. No per-seat gotchas. Bring your own Anthropic key if you'd rather pay usage at cost.
Save 15% on Starter or Pro by paying annually.
- NS read + basic MACD tools
- Chat agent with full audit log
- 500k monthly token allowance
- Community support
- Everything in Starter
- 10M monthly tokens included
- Carrier tools (Bandwidth, Commio)
- Domain-copy templates + deploy
- AI-generated greetings (TTS + Whisper)
- SkySwitch reporting integration
- Email support
- Everything in Pro
- Customer billing module
- BYOT — bring your own telecom carrier
- Support-desk integrations (Zendesk, Freshdesk, HubSpot)
- Self-hosting — run it in your own infra
- BYO LLM + storage API keys (Anthropic, OpenAI, S3, R2)
- Dual-approval change requests + SLA support
Credit-pack pricing subject to change — trending down as model costs drop.
Show-special promo pricing available with a coupon code at signup.
Frequently asked questions.
Does MaxAI work with our NetSapiens deployment?
If you can get an API token out of your NS portal, MaxAI works. We've tested on recent NS versions; older deployments may have fewer API surfaces but the core (subscribers, devices, AAs, queues, audio, recording) is widely supported.
How does MaxAI handle permissions?
Every tool has a risk classification: GREEN (read) executes immediately, YELLOW (write) requires one-click operator confirmation, RED (destructive) requires typed phrase + pre-state snapshot. VIEWER-role users are restricted to GREEN tools. Nothing about this can be bypassed by the agent.
Can we bring our own Anthropic key?
Yes, on Enterprise. Put a key in your tenant settings and MaxAI routes calls through it — you pay usage directly to Anthropic, we skip our credit metering. BYOK calls still record a usage-exempt ledger entry so your audit trail is complete.
What's the audit trail like?
Every tool call, confirmation, and LLM request produces an AuditLog entry scoped to your tenant. Entries include user identity, risk level, input params, output (or error), and — for YELLOW/RED writes — a pre-state snapshot. Sensitive fields like voicemail PINs are redacted from logs but visible in the UI result card.
Can we test on a sandbox before touching production?
Yes. Point MaxAI at a non-production NS tenant using a separate credential. When you're ready, add the production credential as a new entry — MaxAI routes each call to the right instance based on the domain realm.
What about multi-NS-instance operators?
MaxAI natively routes per-domain. Add one credential per NS instance; any operation targeting a specific domain automatically picks the matching instance. Cross-instance audits (e.g. 'list every domain across every server') are first-class.
How long to roll out?
~15 minutes from signup to first tool call. Add NS creds → start chatting. Templates and TTS voice preferences can wait until you've gotten comfortable with the basics.
Your next customer is easier than the last one.
Stop rebuilding the same auto-attendant. Stop clicking through fifty subscribers. Stop apologizing for provisioning delays.