AI sales agents, a CRM, and automated email campaigns for Taos Global Group: Unknown Behavior

01 · The problem

Three business lines, three rulebooks, no one connecting them.

A firm that sells insurance, accounting, and investment services under one roof has an advantage most single-line competitors don't: a client who needs one thing usually needs another. In practice that advantage went unrealized. A commercial client buying a new property and a personal tax client reporting a capital gain are both walking, obvious referrals to another division, but nobody was positioned to catch the trigger in the moment and make the handoff.

Qualifying leads was its own problem, separate from cross-selling. Insurance intake, tax and accounting onboarding, and wealth management all run under different regulatory frameworks, so a single generic intake bot couldn't ask the right questions for all three without either getting one of them wrong or getting watered down into something useless for all of them.

02 · The work

Agents built per line of business, a CRM that sees across all three.

Building AI sales agents purpose-built for outbound and inbound selling, not a general-purpose support bot repurposed for three unrelated product lines. Each pillar, insurance, accounting, and investments, gets intake and qualification logic that respects its own regulatory framework, while a consolidated client profile behind the scenes is what makes cross-selling possible: the same client record informs the insurance agent, the tax agent, and the investment agent, so a trigger caught in one division can hand off to another.

The CRM is designed around that same cross-division view, plus automated email campaigns that follow up between calls on their own schedule. Because this touches regulated outbound calling in a regulated industry, compliance is built in from the start: agents disclose they're automated per California's SB 1001 bot-disclosure law, and call handling follows California's two-party consent-to-record rule under CIPA, both stricter than the federal TCPA baseline.

03 · What the agents do

Four jobs, one client record underneath all of them.

Multi-LOB cross-selling & referral traps

Trigger detection: spotting when an insurance client needs a tax professional, like a commercial client buying a new property
Warm handoffs: introducing a personal tax client to an investment advisor when they report a capital gains event
Consolidated profiling: scanning data across all three divisions to surface under-insured or under-invested clients automatically

High-volume insurance intake & quoting

Instant data gathering: collecting VINs, property addresses, and business details by chat or voice to populate carrier forms
Multi-carrier syncing: structuring risk profiles to pull comparative rates from carriers like Progressive and Travelers instantly
24/7 policy explanations: answering routine questions on deductibles, coverage limits, and exclusions any time

Accounting & tax lead qualification

Entity triage: chatting with new entrepreneurs to determine if they need basic bookkeeping or complex corporate tax structuring
Document chasing: automated, conversational reminders to upload W-2s, 1099s, and expense receipts before deadlines
Seasonal scale: absorbing the tax-season influx of onboarding and service inquiries without adding temporary staff

Wealth management funnel nurturing

Schedule optimization: qualifying high-net-worth leads by asset size and booking them directly onto an advisor's calendar
Educational nurturing: compliance-approved, personalized follow-ups on retirement or education planning
Risk-tolerance surveys: guiding prospects through interactive risk-assessment questionnaires before their first human meeting

04 · Behind the work

Notes from the engagement.

The cross-selling piece is the part that actually moves revenue, and it's the part a single-purpose chatbot can't do. It only works if the client profile is shared across divisions rather than siloed per product line, so the insurance agent noticing a commercial property purchase and the tax agent noticing a capital gains event are reading from, and writing to, the same picture of the client.

Regulatory compliance shaped the build as much as the sales logic did. Insurance, accounting, and investments each carry their own rules for what an agent can say and how a lead can be qualified, on top of California's stricter bar for automated outreach generally, so the qualification flows had to be built per line of business rather than as one generic script wearing three different hats.