Your clients span healthcare, legal, finance, and beyond. RGX is the multi-tenant AI API infrastructure that powers any product you build — PHI compliance, legal output validation, and finance audit trails are activated by a config flag, not a separate vendor contract. Your brand. Any industry. Live in 48 hours.
Pass a config.industry flag and RGX activates the right compliance layer automatically. Your clients span every vertical — your platform handles all of them.
"industry": "healthcare"
"industry": "legal"
"industry": "finance"
"industry": "insurance"
"industry": "real_estate"
"industry": "field_service"
"industry": "professional_services"
"industry": "construction"
"industry": "hospitality"
"industry": "education"
"industry": "nonprofit"
"industry": "generic"
The unfair advantage of being a VAR is that you already have what software companies spend years trying to acquire: trusted relationships with real businesses that rely on you. RGX converts that trust into a software product you own — under your brand, at your price, forever.
Every AI vendor targeting your clients is selling a promise. You walk in with a live, branded platform that already has their data in it — email connected, CRM wired, inbox populated. You demo it in the same meeting where they asked the question.
The meeting that used to end with "we'll look into it" now ends with a signed contract. Because the product is real, it's yours, and it's running right in front of them.
Services revenue requires people. More clients means more headcount. You hit a ceiling and you stay there. Software revenue compounds — each new seat adds to the base without adding to the cost stack.
The economics flip. A client you've had for 4 years who adds 3 seats next year adds revenue without a single new conversation. That's a business with a completely different ceiling.
The platform becomes load-bearing fast. Within 90 days of go-live, it's handling intake, scheduling, client communication, and follow-up. Turning it off would break their daily operations.
Clients who buy software from their VAR stay 3–4× longer than clients buying services alone. The platform turns a vendor relationship into an operational dependency — and that's the deepest moat you can build.
RGX isn't a point solution. It's the entire backend infrastructure for a B2B AI product — 16 endpoint groups covering every layer from client onboarding to billing. You don't stitch together six vendors. You make API calls to one platform that handles all of it.
You already have the clients and the trust. RGX is the AI software backend that lets you monetise both — under your brand, at your price, with no backend build required.
"We've had 35 clients for an average of four years. M365 resale is $4–$8/seat above cost and Microsoft keeps cutting it. Hardware is one-time margin. Managed services burns two techs per 50 clients. Then at a renewal meeting, a client mentioned they'd signed with a smaller competitor for an AI platform at $450/seat. We weren't even in the conversation."
MSP revenue is structurally thin — every line is either one-time, vendor-margin-dependent, or headcount-constrained. You can't grow without hiring. And every year, SaaS vendors go direct to your accounts. The gap between what you charge and what a software company charges for the same client is $300–$500/seat — and you have no way to close it without your own product.
"Every AI vendor we introduce to a client ends up cutting us out. We did the relationship work, vouched for the product, handled the rollout — and three months later they're renewing direct. We get nothing. I needed a product I actually owned and could put in front of them myself."
Technology resellers are in the worst position in the AI wave — close enough to the client to get the initial trust, but not close enough to capture the recurring software margin. Every vendor you introduce is a relationship risk. RGX gives you a product that's yours: your brand, your contract, your renewal, your upsell.
"Every client I sit down with asks some version of the same question: 'What should we be doing with AI?' I used to say 'let's explore options.' Now I open my laptop and show them their own AI platform — built on their data, under my firm's name — and I close the deal in the same meeting."
IT consultancies and field service operators have something most software companies spend years trying to build: trusted relationships with real businesses that rely on them for technology decisions. RGX converts that trust into a software product you can demo, close, and invoice for recurring revenue — without a development team.
Every business in your current client base would pay for a working AI platform. Here's what's standing between you and that revenue.
"My clients keep asking about AI.
I have nothing to show them."
The law firm managing partner, the insurance broker, the HVAC owner — every one of them has asked you some version of this. You say "we're looking into it." Then a competitor walks in, opens a laptop, and demos an AI platform with their own name on it. The client signs that week. You don't get a call back.
"M365 reselling pays me $4 a seat.
I should be charging $400."
Microsoft 365 resale: $3–$8/seat above cost — and Microsoft keeps cutting reseller margins. Hardware: one-time margin, then nothing. Managed helpdesk: $75–$150/seat with 2 techs per 50 clients. Every revenue line you have is thin, one-time, or burns headcount. You physically cannot grow without hiring.
"I know what I'd build.
I just can't afford to build it."
You've thought through the product. You know your clients need it. But multi-tenant architecture, API integrations for a dozen tools, encrypted credential storage, AI routing, compliance logging, billing — that's $400k–$800k in engineering and 12 months before you ship v1. By the time you launch, a competitor already closed your accounts.
"I can't take on more clients
without hiring 3 more people."
Every new account is a custom project. New credentials to track, a slightly different setup from the last one, documentation that lives in one tech's head, and a relationship that falls apart if that tech leaves. Revenue and headcount move together — you're stuck at the same ceiling year after year.
"I lose pitches to companies
that walk in with their own software."
They open a laptop. Branded platform, live client dashboard, AI reading messages in real time — their name on it. You bring a slide deck about response time SLAs. The prospect signs with them. Not because their IT is better. Because they look like a software company and you look like a service company.
"I've lost long-term clients
to SaaS platforms I couldn't match."
You had a 4-year relationship with that brokerage. Then they called to say they were going all-in on HubSpot. You didn't lose because your IT was bad. You lost because you didn't have a software product — and they found one. Now a SaaS company has your client and the next 5 years of their budget.
The math on your existing client base
No new client acquisition. No new staff. Just a software product layer on the relationships you already own.
"The moment I realized I could launch a software product on top of clients I've had for 5 years — without touching the underlying relationship — that was the whole pitch for me."
You're not replacing your managed services. You're adding a product layer on top — one that earns software margins on the same clients you already serve. The relationship is yours. The revenue has been sitting there.
See It Work Free →Every business you work with has the same mess. And they'd pay well for someone to fix it — that someone is you.
Your clients have this problem right now — and none of them know who's going to fix it yet. That's the opening. RGX is the backend. You build the front-end, put your name on it, and own the relationship.
Zapier, Make, and n8n are automation tools — you use them yourself to connect your own apps. RGX is something completely different.
You're the end user. You connect your own Gmail, your own Slack, and build automations that save yourself time.
No "clients" concept. There's no architecture for managing dozens of separate businesses under your account.
Background workflows. A Zap fires when something happens. Your clients still open their own apps — nothing changes for them.
Can't be resold as software. You can't put your logo on Zapier and charge a law firm $600/seat for it.
No client isolation. There's no built-in separation between Client A's data and Client B's. That's not what it was designed for.
Your clients are the users. You connect each client's tools — their Gmail, their CRM — under a product with your name on it.
Built for multi-tenant. You manage dozens of clients from one dashboard. Each one isolated, none of them touching each other.
You're selling software. Your clients log into your product — a real UI with your branding — powered by RGX running underneath.
White-label by design. Set whatever price per seat you want. Clients never know RGX exists. You keep the margin.
Full client isolation. Client A cannot see Client B's data. Each workspace is completely separate, from day one.
The short version: Zapier automates your tasks. RGX is the engine behind a software product you build, brand, and sell to your clients — one where they pay you monthly and you keep most of it.
Every industry below has 4–5 problems that cost real money every single day. The platform you build on RGX solves all of them simultaneously. When it becomes load-bearing — when turning it off breaks their operations — they don't cancel. Ever.
Click any card to see the exact workflows your platform runs for that client type.
"I have 12 attorneys. Every one of them reads their own client emails, drafts their own responses, updates their own CRM notes, and books their own follow-ups. That's $500/hr partner time going into work a paralegal shouldn't even be doing."
A potential client emails after hours. Nobody sees it until morning. By then they've called two other firms. Lost case — $15,000 in fees gone because nobody was watching the inbox.
Incoming inquiry hits email → AI reads, extracts matter type and urgency, creates CRM contact, schedules consultation via calendar, sends personalised acknowledgment within 60 seconds. Attorney gets a brief in the morning, not a cold lead.
"Where is my case?" calls cost 20 minutes each. Attorneys get 6–10 a week. That's 3+ hours of billable time spent answering questions that are in the CRM.
Client emails asking for status → AI pulls CRM case notes and recent activity → drafts personalised status update → attorney reviews and sends in 30 seconds. Every client gets a response. Zero calls.
120-page deposition transcript arrives. Attorney needs the key facts before tomorrow's hearing. Reading the whole thing takes 4 hours they don't have.
Document uploaded to knowledge base → AI generates structured summary: key admissions, contradictions, flagged passages, open questions. Attorney reads a 2-page brief and digs only into what matters. 4 hours → 20 minutes.
The CRM has notes from 6 months ago. Every attorney keeps their own version in email. When a client calls someone new, they start from zero. Malpractice risk is real.
Every email sent and received, every call logged, every document uploaded — automatically written to CRM with timestamp, attorney name, and matter tag. The record builds itself. Audit trail is always current.
Intake goes back to missed overnight emails. Status calls return. The CRM goes stale. Document review takes all night again. Attorneys don't cancel software that saves them 3 billable hours a day.
"January through April I have 200 clients all emailing me at once. I spend 2 hours every morning just triaging. Half of it is 'did you get my documents?' or 'where's my refund?' I went to school for accounting, not inbox management."
Each client owes a different set of documents. Chasing them manually — one email at a time — takes a staff member 3 weeks every year. Returns can't start until docs arrive. Deadlines slip.
Knowledge base holds each client's document checklist. Platform reads their incoming emails, marks off received items, sends personalised follow-up for what's still missing — automatically, every 5 days — until the file is complete. Staff touches nothing.
During busy season, 40% of inbound emails are "where is my return?" Answering each one manually burns 10 minutes. Multiply by 200 clients over 12 weeks. That's weeks of staff time on questions that have answers in the system.
Client emails asking for status → AI reads inquiry, pulls their CRM record, checks current return stage, drafts a personalised reply with current status and expected completion date. CPA approves in 15 seconds. All 200 clients feel personally attended to.
Tax planning meetings require pulling last year's return, CRM notes, this year's documents, and any outstanding questions — from 4 different places. Takes 45 minutes per client. CPAs skip it and wing it. Clients notice.
Night before the meeting: AI reads email history, pulls CRM notes, scans uploaded documents from knowledge base, generates a one-page brief — open items, last year's highlights, what the client asked about, what was promised. 45 minutes → 0. CPA walks in prepared.
Every client has a different deadline. April 15 is the bulk but extensions, quarterly estimates, and business filings scatter across the calendar. Missing one costs penalties and a client.
Platform reads the calendar, checks each client's deadline, sends personalised reminder email + SMS at 30 days, 14 days, and 3 days out — with exactly what's still needed and how to submit it. Zero manual tracking. Zero missed deadlines.
Document chasing is manual again. 40% of the inbox is status questions again. Prep goes back to 45 minutes per meeting. Deadlines live in someone's head. You cannot remove the thing that got a CPA through tax season.
"I spend every Monday morning reading last week's emails, updating notes, and writing performance summaries. That's 3 hours of my time that should be billable. And when the market drops, I get 40 emails in 2 hours and I can't answer all of them personally."
80 clients. Each one has different holdings, different risk tolerance, different life events. Sending meaningful personalised updates to all of them manually is impossible. So most get nothing. They feel ignored. They leave.
Every Friday: AI reads each client's email history, pulls CRM notes, scans uploaded statements from knowledge base, writes a personalised weekly digest for each client with portfolio highlights and a relevant note. Advisor reviews batch, sends with one action. 80 clients, 20 minutes total.
Market drops 4%. 40 anxious clients email within 90 minutes. Each needs a calm, personal response that acknowledges their specific situation without constituting advice. Responding to all 40 takes all day. Some wait 2 days. They start looking elsewhere.
Client emails arrive → AI reads each one, pulls their risk profile and recent conversation history from CRM, drafts a personalised, calm response referencing their specific situation — framed as informational. Advisor reviews 40 drafts in 30 minutes, sends. Every client hears back within the hour. Full audit trail logged per regulatory requirement.
Each quarterly review requires pulling account history, CRM notes, uploaded statements, and recent communications — from different places. 45 minutes per client × 20 quarterly reviews = 15 hours of prep every quarter. It either doesn't happen or it's rushed.
48 hours before each review: AI pulls CRM history, reads recent email threads, scans uploaded account statements, generates a one-page brief — portfolio summary, open client concerns, what was promised last quarter, suggested talking points. Advisor walks in with everything. 45 minutes of prep → zero.
Clients whose advisors don't proactively reach out leave at 3× the rate of clients who receive regular unprompted contact. Birthdays, account anniversaries, life events — advisors know they should be reaching out. They don't have time.
Platform monitors CRM for trigger events — birthday, anniversary, major market move, life event noted in conversation. Automatically drafts and sends personalised outreach via email or SMS. Every client touched at every relevant moment. Attrition drops. Referrals go up.
Monday prep is back to 3 hours. Market volatility response is back to all-day manual work. 60 of 80 clients go back to getting nothing personalised. The advisor who couldn't respond fast enough during the last correction already knows what that costs.
"We get 150 inbound leads a month from the website. My agents are on the phone working renewals. By the time someone circles back to the new lead, it's been 6 hours. We've already lost half of them. The competitor down the street responds in 4 minutes."
Insurance leads have a 5-minute window. Respond in under 5 minutes: 9× more likely to convert. Respond in 30 minutes: 21× less likely. Every hour agents spend on existing clients, new leads go cold and sign with someone else.
Lead submits web form → webhook fires → AI reads, qualifies coverage type and urgency, logs to CRM, sends personalised response within 60 seconds with a quote request link and calendar link to book a call. Agent picks up a warm, pre-qualified lead. Not a cold one from 6 hours ago.
A policy renews and the client never hears from their agent beforehand. They shop around, find a better rate, and leave. The agency finds out when the cancellation notice arrives. Every lapsed renewal is 100% revenue loss on a relationship that took years to build.
90 days before each renewal date: AI sends personalised renewal outreach email — references their policy, their history with the agency, asks if anything has changed. 60 days out: follow-up with updated quote. 30 days out: SMS reminder. Agent only gets involved if the client responds with a question. Retention goes up without adding headcount.
Client calls to report a claim. Staff takes handwritten notes. Someone types them into the system later — sometimes. The client gets no acknowledgment for hours. They're already anxious. The experience sets the tone for everything that follows.
Client emails or texts to report a claim → AI reads, extracts incident details, date, parties involved, urgency level → creates claim record in CRM → sends personalised acknowledgment with claim number and next steps within 2 minutes. Full transcript logged for compliance. Adjuster receives a structured brief, not a phone message.
A state audit requires documentation of every client communication for the past 3 years. Staff spends 2 weeks pulling emails, call logs, and notes from 4 different systems. Half of it is missing. Fines are real.
Every interaction — email, SMS, call log, document upload — automatically timestamped and logged to the audit trail with client ID, agent ID, and policy number. Full audit report exports in minutes. Three years of documentation, structured, searchable, complete.
Lead response goes back to hours. Renewals go unworked until it's too late. Claims intake is handwritten notes again. The audit trail is back to 4 disconnected systems. Revenue leaks at every single one of those points simultaneously.
"My front desk staff spends the first 4 hours every morning on the phone — scheduling, rescheduling, confirming, chasing patients who missed their appointment. They're doing the work of a booking system and a call center at the same time. I pay them $22/hr to answer the same 6 questions on repeat."
Every appointment scheduled is a phone call. Every reschedule is another. A 200-patient practice gets 40–60 scheduling interactions a day. That's 2 full staff members doing nothing but answering the same questions.
Patient texts or emails to request an appointment → AI reads request type (new patient, follow-up, urgent), checks calendar availability, books the slot, logs to patient CRM record, sends confirmation with date/time/location and prep instructions from knowledge base. Zero staff touch for routine bookings.
Industry average no-show rate: 18–23%. Each no-show is a slot that can't be filled on 2 hours notice. For a practice seeing 40 patients a day at $200 average, that's $1,600+ lost every single day to patients who simply forgot.
48 hours before appointment: personalised SMS reminder with appointment details and one-tap confirm/cancel. 2 hours before: final reminder. If cancelled same-day: AI sends SMS to waitlisted patients offering the slot immediately. No-show rate drops 50–60%. Slots fill automatically.
Patients don't get sick between 9 and 5. Emails and texts arrive evenings and weekends. They wait until morning for a response — by which time they've gone to urgent care, called a competitor, or simply gotten more anxious. First response time is a patient satisfaction metric.
Patient contacts practice after hours → AI reads inquiry, pulls from practice knowledge base (hours, services, common symptoms, intake process), responds with relevant information immediately, books appointment if needed, flags urgent cases for on-call staff. Patients get answers at 10pm. PHI scrubbed from every LLM interaction before processing.
New patient intake requires collecting insurance information, medical history, and consent forms — before the appointment. Staff calls patients to collect it. Most patients don't have insurance cards handy. It takes 3 attempts per new patient.
Appointment booked → AI immediately sends personalised intake email with secure form links from knowledge base. 48 hours before appointment: reminder if forms incomplete. Information flows directly to CRM. No phone tag. Patient arrives with everything done. Staff spends intake time on actual clinical prep.
Front desk is back to 4 hours of phone calls. No-show rate is back to 20%. After-hours inquiries go unanswered until morning. New patient intake is phone tag again. The practice was running on 2 fewer front desk staff because of this platform. That's immediately visible in payroll.
"We get 200+ leads a month. My agents are showing homes, in negotiations, driving between appointments. 60% of our leads never hear back within 24 hours. I watch deals go to competitors who respond before my agents even see the notification."
Studies are unambiguous: lead conversion drops 80% after the first hour. A brokerage getting 200 leads a month and responding in 4+ hours average is converting at maybe 2%. The same leads, responded to in under 5 minutes, convert at 10–15%.
Lead hits website → webhook fires → AI qualifies (buyer/seller, price range, timeline, area), scores urgency, logs to CRM, assigns to right agent based on territory, sends personalised response within 60 seconds with relevant listings from knowledge base and a calendar link to book a showing. Agent gets a warm lead with full context, not a cold form submission.
Scheduling showings is 15 back-and-forth messages per appointment. Agents spend 90 minutes a day on scheduling logistics alone. That's time not spent closing deals. Buyers get frustrated waiting and move on.
Buyer requests showing → AI checks agent calendar, offers 3 available slots via SMS, buyer confirms with one reply, appointment created in calendar, confirmation sent to both parties, reminder sent 2 hours before. 15 messages → 2. Agent involvement: zero until they show up.
Buyers and sellers are anxious throughout a transaction. They email asking for updates at every stage. Agents respond when they can — which is often hours later. Anxiety becomes frustration. Referrals evaporate.
At every transaction milestone — offer submitted, accepted, inspection scheduled, appraisal ordered, clear to close, closing date set — AI sends personalised status email to both parties. Clients feel informed at every step without the agent lifting a finger. Satisfaction scores and referral rates increase measurably.
Past clients are the best source of referrals. Most agents lose touch within 6 months of closing. The database goes cold. A competitor who sends a market update email every month gets the referral call instead.
Platform monitors CRM for past clients. Monthly: personalised market update email with their neighbourhood stats from knowledge base. Annually: home anniversary email. Any time market conditions shift significantly: proactive outreach. Every past client hears from the agent every month. Referrals increase 40%.
Lead response goes back to hours. Showing scheduling is back to 15 messages per appointment. Transaction updates stop. The past client database goes silent. A brokerage that converted 12% of leads is back to converting 2%. That math is immediately felt in closed deals.
"My dispatcher manually reads every service request, finds an available tech, calls them, creates a work order, and texts the customer a confirmation. Each booking takes her 45 minutes. We lose jobs because customers call 3 companies and go with whoever responds first. That's usually not us."
Customer needs AC fixed today. They text 3 HVAC companies. The one that responds in 4 minutes with availability and a confirmation gets the booking. The ones that respond in 2 hours get voicemail.
Customer texts in → AI reads request type and urgency, checks technician availability in CRM, creates work order, notifies the right tech via Slack with job details, sends customer a confirmation with tech name and arrival window via SMS — in under 90 seconds. 45-minute dispatch process → 90 seconds. Zero dispatcher involvement for routine bookings.
Tech shows up at a job with no history. Last visit was 8 months ago. Customer mentions the previous tech told them something. Current tech has no idea what was said, what was fixed, or what to look for. Customer loses confidence. Job takes longer.
When job is assigned, AI pulls complete customer history from CRM — previous visits, reported issues, parts replaced, what was promised — and sends it to the tech via Slack before they arrive. Tech walks in knowing the full picture. Jobs are faster. Customers feel recognised. Reviews improve.
Tech drives 40 minutes to a job. Customer forgot and isn't home. That's a wasted hour, a missed slot, and a frustrated tech. It happens 15–20% of the time with no reminder system.
Evening before job: personalised SMS reminder with tech name, arrival window, and what to have ready. Morning of job: 2-hour reminder. If customer can't make it, they reply to reschedule — AI handles the rebooking and notifies the tech immediately. No-shows drop 70%.
After the job, nobody follows up. No invoice reminder. No review request. No maintenance follow-up in 6 months. Revenue from recurring maintenance contracts is the highest-margin business in field service. Almost nobody captures it.
Job marked complete → AI sends invoice via email with payment link. 48 hours later: review request SMS. 6 months later: personalised maintenance check-in email based on what was serviced. Recurring revenue that was invisible becomes a structured pipeline.
Dispatch is back to 45 minutes per job. Techs show up blind. No-shows are back to 15%. Post-job follow-up stops. This company went from booking 70% of inquiries to booking 95%. That difference — in a 25-tech operation — is hundreds of thousands of dollars annually.
"I have 3 active projects and 20 subcontractors. I spend 2 hours every morning just texting and emailing to find out where things stand. If I don't chase them, they go silent. I have no idea what's actually happening on site until I drive out there."
Each sub uses a different communication method. Some text, some email, some use WhatsApp. The PM manually aggregates status updates from all of them every day. Something always falls through the cracks.
Each morning: AI sends personalised daily check-in to every sub via their preferred channel (SMS or email), asks for status on their active tasks. Responses are read, parsed, and logged to the project CRM. PM sees a single structured summary by 8am — who's on track, who's behind, what needs attention. No manual aggregation.
Project owner emails "what's the status?" every other day. PM writes a status report from scratch each time. Each report takes 30 minutes to compile. On 3 active projects, that's 90 minutes a day on reporting instead of managing.
Client emails for status → AI pulls current project milestones from CRM, reads recent sub updates, generates a structured project summary with percent complete, recent completions, next milestones, and open issues. PM reviews and sends in 2 minutes. Client gets a professional report. PM didn't write a word.
RFIs (requests for information) arrive via email and get buried. Change orders are requested verbally and never documented until there's a dispute. Untracked changes are how projects end in lawsuits.
Every incoming email with RFI, change order, or scope question flagged by AI, logged to CRM with date/sender/project tag, and assigned a tracking number. Automated acknowledgment sent within minutes. 7 days with no response: automatic follow-up. Complete paper trail on every item. Disputes become conversations instead of lawsuits.
Subs work off different versions of plans. The latest revision is in someone's email. Three subs are working off a drawing that was superseded 6 weeks ago. Rework is expensive. Arguments are expensive. Both are avoidable.
All drawings, specs, and permits uploaded to knowledge base with version control. When a sub asks "which version of the electrical plan is current?" — AI answers instantly from knowledge base. When plans are updated, AI sends notification to all relevant subs automatically. Everyone works from the same document. Always.
Sub status updates are back to 2 hours of morning texts. Client reports are back to 30-minute manual writes. RFIs are buried in email again. Someone's working off the wrong plan version. Construction margin is already thin — this platform is the difference between a profitable project and a dispute.
"My recruiters spend 3 hours a day on email — sending job descriptions, following up on applications, coordinating interviews, checking references. They're not recruiting. They're doing administrative work that a system should handle."
A recruiter manages 40 active candidates. Each one needs regular touchpoints — status updates, interview prep, document requests, offers. Doing it manually means candidates fall through the cracks and accept other offers first.
At every pipeline stage change in CRM: AI sends personalised email to candidate with next steps, what to expect, and what to prepare. Interview confirmed: prep email sent automatically with job description, company background from knowledge base, and logistics. Offer made: congratulations + next steps sent immediately. Candidate always knows where they stand.
A client sends a job order. Recruiter confirms. Two weeks pass. Client emails "any progress?" Recruiter writes a status update from scratch. The relationship feels fragile because communication is reactive, not proactive.
Job order received → logged to CRM → AI sends weekly update to client with pipeline status: candidates sourced, screened, in process, rejected with reason. Client sees structured progress without asking. Recruiter looks proactive without lifting a finger beyond the actual recruiting work.
Every placement requires references, background check consent, and onboarding documents. Chasing these from candidates manually adds 3–5 days to every placement cycle. Delays cost placements when clients have urgent needs.
Candidate moves to offer stage → AI immediately sends personalised email with all required document links and deadlines. 48 hours with no action: automated follow-up via SMS. Compliance checklist tracked in CRM. Nothing falls through. Placement cycle compresses by 3–5 days on average.
Candidate communication is back to manual. Client updates are back to reactive emails. Compliance chase adds days to every placement. Recruiters are back to spending half their day on process instead of placements. A firm doing 30 placements a month compresses to 18.
"I have 12 short-term rental properties. Guests text me at all hours — check-in instructions, WiFi passwords, where to park, what restaurants are nearby. I'm answering the same 8 questions 40 times a week. And if I don't respond fast, they leave a bad review about 'unresponsive host.'"
80% of guest questions are identical: check-in time, check-out time, parking, WiFi, local recommendations, early check-in availability. Answering them manually is purely reactive labour that scales with property count.
Guest texts or emails with any common question → AI reads inquiry, pulls relevant answer from property knowledge base (check-in instructions, parking details, WiFi, local guide, house rules), responds immediately in personalised conversational tone. 80% of guest messages handled with zero staff involvement. Response time: under 30 seconds.
Guest arrives. They didn't read the check-in email. They text: "Where do I get the key?" "What's the door code?" "Where do I park?" Each of these is a separate message requiring a response. Multiplied by 12 properties, check-in day is chaos.
Day of arrival: AI sends personalised check-in briefing via SMS with door code, parking instructions, WiFi details, and check-in steps — timed to 2 hours before their expected arrival. If they ask anything anyway, AI answers from knowledge base instantly. Check-in calls drop 90%.
Happy guests leave. Nobody asks them to review. 30% of guests who had a great experience would leave a 5-star review if asked within 24 hours of checkout. Most properties ask nobody and get reviews from 5% of guests — mostly the unhappy ones.
2 hours after checkout: personalised thank-you SMS with a direct review link. 48 hours later: follow-up email if no review yet. Review rate jumps from 5% to 30–40%. Rating average increases. Ranking improves. Bookings go up. The platform paid for itself in higher occupancy.
Guest questions go back to manual responses at all hours. Check-in chaos returns. Review rate drops back to 5%. Property ranking falls. The operator was managing 12 properties with 1 person because of this platform. That becomes impossible immediately.
"We get 80 enrollment inquiries a month for our tutoring centre. My admissions coordinator follows up on maybe 40 of them — the ones she gets to between everything else she's doing. The other 40 go cold. Every one of those is a student we didn't enroll."
Parent emails asking about programs. Nobody follows up for 3 days. Parent enrolled their child somewhere else on day 2. Every unfollowed inquiry is a lost enrollment. At $2,000/student/year, 40 lost inquiries per month is a $80,000/month revenue gap.
Inquiry arrives via email or form → AI reads, identifies grade level and subject need, pulls relevant program info from knowledge base, responds within 60 seconds with personalised program recommendations and a calendar link to book a consultation. 3 days later if no booking: follow-up SMS. Every inquiry followed up. Every time.
Parents want progress updates. Teachers and coordinators spend 30–45 minutes a day writing individual update emails. The updates are generic because there's no time for personalisation. Parents feel disconnected. Retention suffers.
Weekly: AI pulls each student's CRM record, session notes, and progress data from knowledge base, generates personalised progress email for each parent — specific to their child's sessions, what was covered, what improved, what to practice at home. Teacher reviews in batch, sends. Parents feel genuinely informed. Retention rate increases significantly.
Parent needs to reschedule. They email. Staff responds when they see it. Back and forth takes 4 messages and half a day. Meanwhile the slot sits open, another student who wanted that time wasn't offered it, and revenue is lost.
Parent emails to reschedule → AI reads request, checks calendar, offers 3 available slots via reply, parent confirms in one message, calendar updated, both parties confirmed. Slot change complete in under 10 minutes. If original slot opens: AI immediately texts the waitlist. No revenue gap.
Half of inquiries go unfollowed. Parent updates go back to generic weekly emails. Rescheduling is 4 messages and half a day. This school was enrolling 60% of inquiries instead of 30%. The platform difference is directly measurable in enrolled students.
"We have 1,200 donors. When someone gives, I want to thank them personally and tell them exactly what their gift is doing. We send one bulk email a month. I know it's not enough. But we have 3 staff and no time. The donors who care most are getting the most generic treatment."
A donor gives $500. They receive a bulk thank-you email 3 days later that doesn't mention the amount, doesn't reference their history, and reads like it was written for 10,000 people. Because it was. Donors who feel unappreciated reduce or stop giving within 12 months at a rate of 40%.
Donation received → AI immediately reads donor CRM record (giving history, programs they've supported, last interaction), generates a genuinely personalised thank-you email — references their specific gift, their history with the org, and what this gift specifically enables. Sent within 5 minutes of the donation. Every donor, every gift, same treatment as a major donor gets manually today.
A donor who gave last year didn't give this year. Nobody noticed until the year-end report. By then they've been absent 11 months. Reactivation rate after 12 months of silence is under 10%. Catch them at month 6 and it's 40%.
At 6 months since last gift: AI generates personalised re-engagement email — references their past giving, shares a specific impact story from knowledge base relevant to their previous donation area, includes a soft giving prompt. At 9 months: follow-up with a different impact angle. Caught early, re-engaged before they're gone.
Coordinating 50 volunteers for an event requires 3 staff members sending individual emails, tracking RSVPs in a spreadsheet, sending reminders, and following up on no-shows. It takes more staff time than the event itself.
Volunteer opportunity posted → AI sends personalised invitation to relevant CRM contacts based on their interest tags. RSVP tracked automatically. 48 hours before event: reminder with logistics from knowledge base. Day of: final confirmation SMS. Post-event: personalised thank-you with impact summary. Entire coordination flow runs automatically. Staff focuses on the event, not the logistics.
Grant funders require periodic impact reports. Staff writes them from scratch, pulling data from spreadsheets, photos from a shared drive, and anecdotes from memory. Each report takes 2 days. Most small nonprofits underreport and underperform on grants because the reporting overhead is crushing.
Staff uploads program data, photos, and outcomes to knowledge base throughout the period. When report is due: AI pulls all relevant data, generates a structured draft report with metrics, narrative, and supporting evidence organised by grant requirement. Staff edits and personalises. 2-day write → 2-hour review. More grants applied for. More grants won.
Donor acknowledgment goes back to bulk emails. Lapsed donors go unnoticed for 11 months. Volunteer coordination is back to 3 staff and a spreadsheet. Grant reports take 2 days each. A 3-person nonprofit was operating with the communication capacity of a 12-person org. That's immediately gone.
Every workflow above runs on email, SMS, CRM, calendar, calling, Slack, Teams, and knowledge base endpoints — all under your brand, through one API key.
No approval forms. No waiting. Try it free, go live when you're ready.
Hit the sandbox at /demo — no approval, no credit card. You get a live API key instantly and 500 test requests to explore the full platform. Takes 2 minutes.
When you're ready to go live, sign up for a production node at /demo. Your API key arrives by email instantly — format: rgx_live_…. Store it securely. First invoice on the 1st of next month.
Each client you onboard becomes an active seat on your node. One POST /clients creates their workspace. A few more calls connect their Gmail, CRM, Slack, and SMS. See the Deployment Guide for the full step-by-step — most ship their first client in under 48 hours.
You charge your clients whatever your market supports. RGX bills you $2k/mo base + $30 per active seat. Everything above that is yours. Check your usage stats and estimated invoice anytime at /api/v1/usage.
Want to run your own AI model instead of RGX's? Set passthrough: true — RGX logs and routes everything for compliance without processing it. Your model, our infrastructure.
Everything is provisioned under your brand, your domain, your API key naming. Your clients never know RGX exists. Here's what a live integration looks like:
All channels feed the same backend. Your clients get one unified workspace — not five tools taped together.
One flat model. No variable tiers. No surprise fees. The more clients you add, the better your unit economics get.
Your base monthly fee for a production VAR node. Covers infrastructure, API access, provisioning, and usage tracking. Flat — never increases as you add clients.
One seat = one active client. $30/month per client. You charge your clients whatever your market supports — $200, $500, $1,000+. The difference is entirely yours.
Every API call is logged to your node. Message volume, processing usage, active seats, and estimated invoice — all visible anytime at /api/v1/usage.
Your margins improve as you scale. Your cost is $2,000 + ($30 × clients). You set your own per-client pricing. 20 clients at $500/mo = $10,000 revenue against $2,600 cost — $7,400 net.
Illustrative only. Your actual pricing, client volume, and retention determine your earnings.
Full documentation at /docs. Authenticate with your Master API Key on every request via the X-Api-Key header.
Confirm your API key is valid and your node is active. Run this before going to production — takes 200ms to return a green light.
Send any channel payload — WhatsApp, SMS, email, calls — and RGX logs it with a timestamp. Optionally trigger AI processing on the last message. All events tagged to your node.
Send any text input, receive a structured AI response. Set passthrough: true to log without processing — for teams running their own AI model.
Your node's usage stats — requests, active seats, seat charges, and estimated invoice. Current month and last 6 months.
RGX is for operators who already have client relationships and want to launch a recurring software revenue line — not individuals exploring a vague interest in AI.
RGX is an operator platform. If you manage client relationships and want to own a software product, you're in the right place. The sandbox is free — validate before you commit.
Start with the free sandbox — no commitment.
Get a live API key instantly and 500 free requests to build and test. Sign up for a production node in minutes when you're ready to go live.
Get Started Free →Try the sandbox free — live API key in 2 minutes. Sign up for a production node when you're ready. First invoice on the 1st.
Free sandbox · No approval required · Production node in minutes