Referral tracking and re-engagement

What automation does

AI automation from Grow2.ai handles two related tasks: tracking the status of every referral and systematically re-engaging contacts that have dropped out of active work. A referral is any incoming direction: a client from a partner, a patient from a physician, a lead from word of mouth. Once a referral enters the system, automation tracks every step of its progression, records gaps, and triggers touch sequences when movement stops.

Re-engagement is the second loop. Automation periodically scans the contact base, identifies those who have not interacted with the company beyond a set threshold, and launches a re-engagement campaign: a personalized message through the preferred channel, an offer to book or schedule a call, automatic transfer to the active funnel upon response.

What the automation loop includes

1. Receiving a referral from a source (partner, website form, call, EHR, CRM).
2. Assigning a status, owner, and SLA to each stage.
3. Monitoring: if a stage does not move within N days — an alert to the owner and a follow-up to the client.
4. Segmenting the base by recency of last contact.
5. Launching a re-engagement campaign based on the preferred channel (email, SMS, messenger).
6. Transferring respondents to the active funnel and notifying the manager.
7. Logging all events for reporting and compliance.

What automation does NOT do: it does not sell instead of the manager, does not negotiate price, does not make diagnoses, and does not replace a physician or consultant. The decision on the content of each touch is made by the operator using templates — the AI agent executes and escalates.

Typical configuration options

Solo (1–5 people). Minimal configuration: one incoming referral channel, one funnel, basic monitoring on two rules — "no movement for 7 days" and "inactive for 60 days". Re-engagement in one wave, one template. Calendar integration for automatic call booking. Setup takes a few days, including a base export and first run. Suited for solo consultants, single clinics, coaches — those who close deals themselves and lose leads in the chaos of inbound.

SMB (6–30 people). Referral segmentation by source and service type, rule-based owner assignment, SLA per stage. Re-engagement in 2–3 waves with escalation to a live manager. Integration with CRM or EHR, calendar, communication channels. Reporting on conversion at each stage and ROI of re-engagement campaigns. Setup — approximately one working week. Typical case: a 3-physician clinic with a coordinator, a consulting team with account managers.

Enterprise (30+ people). Multi-tenancy across branches or business units, complex routing rules, simultaneous integration with multiple systems (CRM + ERP + billing + EHR), compliance mode (HIPAA, GDPR) with audit and encryption. Re-engagement campaigns segmented by lifetime value, purchase history, and preferred channel. Setup — 2–4 weeks including approvals. A dedicated manager from Grow2.ai leads the project to stable operation and does a handoff to the internal team.

How it works

AI automation is split into three layers: data collection, rule logic, actions. Each layer is configured for a specific company — no code, in config.

Step 1. Connecting sources

Automation connects to the systems where referrals and clients already live: CRM, EHR (for clinics), calendar, communication channels (email, SMS, messengers, telephony). Grow2.ai uses standard APIs and webhooks — if you have a vertical SaaS with an open interface, integration is standard. If the system is closed, an intermediate connector is added.

Step 2. Rule configuration

The config defines funnel stages (for example: 'referral received → initial contact → booking → visit → post-visit'), SLA for each stage in days, responsible roles, communication templates by channel, and re-engagement segments. The AI agent does not invent these rules — it executes what has been defined. Rule complexity ranges from a few conditions to dozens.

Step 3. Monitoring and alerting

Every N minutes the system scans all active referrals and checks their status against SLA. Violations are classified by priority:

1. Critical — the referral is losing value (for example, insurance expires in 3 days and the visit is not booked). Alert to the responsible manager in Slack or by email + follow-up to the client.
2. High priority — the stage has not moved for longer than the SLA. Follow-up to the client by template, copy to the manager.
3. Scheduled — reminder one day before the meeting, confirmation request.

Step 4. Re-engagement loop

In parallel with active funnel monitoring, the re-engagement process for inactive contacts runs. Inactivity criteria are a parameter (30, 60, 90 days without contact). The selection is segmented, and each segment is assigned a message sequence. A client response is a trigger for transition to the active funnel and manager notification.

Step 5. Reporting

Automation records all events: sends, response receipts, link clicks, visit bookings, cancellations, returns. From this a dashboard is assembled: stage-by-stage conversion, fallthrough rate (uncompleted), ROI of re-engagement campaigns, average time-to-contact.

Alternative approaches

The manual approach breaks down at scale: with just dozens of active referrals per week, a person loses control and gaps begin to appear. No-code tools like Zapier or a workflow engine cover standard scenarios, but cannot handle segmentation, non-standard logic, and reporting — every new scenario requires additional setup, and maintenance becomes burdensome. Grow2.ai AI automation is different in that it covers the entire loop at once: rules, monitoring, communication, escalation, reporting — in one system, without stitching together 5–7 different SaaS.

Security and compliance

For clinics, HIPAA (USA), GDPR (EU), and local medical regulations are critical. Grow2.ai configures automation with these requirements in mind:

1. Personal data does not leave the CRM or EHR perimeter without explicit authorization.
2. Patient communication goes through approved channels with encryption in transit.
3. Event logs are stored in an auditable format with operator attribution (human or AI agent).
4. Access to configuration is role-based.

For consulting, NDA and control of client information leakage are important — automation operates within the client's infrastructure or in an isolated environment, and data is not transferred to third parties without authorization.

Prerequisites for implementation

AI automation is an add-on to your system, not a replacement. For the project to launch within a week and deliver results, basic conditions on the client's side are required.

Technical

1. CRM or EHR with API / webhook (or at least scheduled CSV export).
2. A calendar with the ability to programmatically create and reschedule appointments.
3. A client communication channel accessible via API: email service, SMS gateway, messenger provider, or telephony.
4. Admin rights for the client's contact person — to grant Grow2.ai access without waiting on the IT department.

Organizational

1. Documented stages of the referral or lead funnel. If no stages exist — we define them in the first session, but this adds a few days to the setup.
2. A designated point of responsibility — a manager or coordinator who receives alerts and handles escalations. Automation works not "instead of people" but "with people".
3. Client data processing consent — Grow2.ai provides consent templates.

Data

1. A contact database with up-to-date emails and phone numbers. Database quality is critical: outdated contacts → low response rate → wasted SMS budget.
2. At least 6 months of interaction history — for segmenting inactive contacts and calibrating rules.

Potential pitfalls

• Overly aggressive re-engagement. Frequent touches to inactive contacts cause complaints and unsubscribes. A conservative setup during the first 30 days reduces the risk.
• Templates without personalization. A generic "Hello, we haven't been in touch for a while" yields a low response rate. Templates are adapted to the segment and channel, not taken from defaults.
• No designated owner for alerts. If alerts go to a shared chat with no designated person — they get ignored. SLA is assigned to a specific role.
• Setup without reliance on real data. Rules set arbitrarily produce either a flood of false positives or misses. Before launch, we run the last 3 months of history through the rules and review what triggers.
• Ignoring compliance requirements in the healthcare industry. SMS containing medical details is a violation. Communication via unencrypted channels is a violation. Separating channels by data sensitivity is mandatory.