
SOC 2 is an independent auditor’s report on how a company protects customer data. It is not AI-specific. It applies to AI tools the same way it applies to any software: the auditor tests the controls around the system, not whether the AI gives good answers. For an owner buying AI voice agents, AI follow-up or a CRM setup, the report is one signal among several. The bigger questions are who handles your lead data, who is accountable when something goes wrong, and whether the calls and texts the AI sends stay inside the rules.
This guide explains SOC 2 in plain terms, what it does and does not say about AI, and how to judge an AI lead generation setup as a business owner. It is written for personal injury firms, home services contractors, insurance agencies and local service businesses, not for enterprise IT teams.
TL;DR
- SOC 2 is a voluntary report written by an independent CPA firm, built on five criteria from the AICPA: security, availability, processing integrity, confidentiality and privacy.
- It is not a law and not a certificate. Customers ask for it because they want outside assurance on data handling.
- It does not grade AI quality. It covers the controls around a system, and a company can choose to add AI-related controls to its scope.
- For AI that calls and texts leads, consent and telemarketing rules matter as much as any audit report.
- A done-for-you partner that runs your ads, AI agents and CRM in one place leaves fewer handoffs where lead data can leak.
Table of Contents
- What SOC 2 actually is
- Does SOC 2 cover AI?
- Is SOC 2 legally required?
- The rule that matters more for AI follow-up: TCPA
- How to judge an AI lead system without being an IT expert
- A worked example: five vendors versus one
- Where this matters by industry
- How MediaBloom delivers this done-for-you
- Ready to put AI on your leads?
- Our perspective
- FAQ
- Sources
What SOC 2 actually is
SOC stands for System and Organization Controls. The framework comes from the American Institute of Certified Public Accountants (AICPA), the body that sets standards for CPAs in the United States. The AICPA describes SOC 2 reports as reports on a service organization’s controls relevant to security, availability, processing integrity, confidentiality or privacy. Those five areas are called the Trust Services Criteria.
In plain words:
- Security: is the system protected against unauthorized access?
- Availability: is it up and usable as promised?
- Processing integrity: does it process data completely and accurately?
- Confidentiality: is sensitive business information restricted to the right people?
- Privacy: is personal information collected, used and kept as the company says it is?
A company picks which criteria to include. Security is the baseline. Many reports stop there, others add availability or confidentiality.
Type 1 and Type 2
There are two versions. A Type 1 report looks at whether controls are designed properly at a single point in time. A Type 2 report tests whether those controls worked over a period, commonly several months. Type 2 carries more weight because it shows the controls operated day after day, not just on paper.
A report, not a certificate
People say “SOC 2 certified,” but a SOC 2 is an attestation report issued by a licensed CPA firm. There is no pass or fail badge. The report lists what the auditor tested and any exceptions found. That is why two companies that both “have SOC 2” can be very different in what they covered.
Does SOC 2 cover AI?
Not by name. The Trust Services Criteria were written before generative AI became common in business software, and they do not contain a separate AI chapter. What they do cover still applies when AI is part of the system:
- Security: who can access the AI platform, its prompts, call recordings and the lead data it touches.
- Availability: whether a voice agent that answers your phones is up when your customers call.
- Confidentiality and privacy: how caller details, claim information or home addresses are stored and shared with AI model providers.
- Processing integrity: whether lead data moves correctly between the AI, the CRM and your calendar.
A vendor can ask its auditor to add AI-specific controls to the scope, such as how models are approved, how training data is handled and how outputs are monitored. That is a choice, not a requirement. So a SOC 2 report tells you how a vendor protects data. It does not tell you whether the AI books appointments well, qualifies leads accurately or sounds natural on the phone.
What the report will not tell you
- Whether the AI’s answers are accurate or on-brand.
- Whether the vendor’s calling and texting practices follow telemarketing consent rules.
- Whether your ads, AI follow-up and CRM work together.
- Whether anyone on the vendor’s team will tune the system after launch.
For a business owner, those four points drive results. The audit addresses the data side only.
Is SOC 2 legally required?
No. No federal law says a software company must hold a SOC 2 report. It exists because customers, especially larger ones, ask for it in vendor reviews and contracts. If you run a roofing company, a PI firm or an insurance agency, no regulator will ask whether your marketing vendor has one. Your own clients and partners are more likely to ask how you handle their information.
The rules that do apply to you sit elsewhere, and for AI lead generation the most direct one is the telemarketing law.
The rule that matters more for AI follow-up: TCPA
The Telephone Consumer Protection Act (TCPA) governs autocalled and prerecorded calls and texts to consumers. In February 2024 the FCC ruled that calls using AI-generated voices count as artificial voices under the TCPA, which means the consent requirements for prerecorded and artificial voice calls apply to them.
What that means for you as a business owner, in practical terms:
- An AI voice agent that calls leads is treated like any other automated outreach. Consent and opt-out handling matter.
- Speed-to-lead follow-up that texts or calls a person who just filled out your form depends on what that form told them they were agreeing to.
- Cold outbound and database reactivation campaigns need extra care about whose numbers are on the list.
This is a general overview, not legal advice. For your specific campaigns and consent wording, talk to your own attorney. Law firms should also take questions about advertising and intake rules to their state bar or ethics counsel.
How to judge an AI lead system without being an IT expert
A SOC 2 report is one input. Owners who buy AI for lead generation get better outcomes by looking at the whole chain a lead travels. Use this list.
1. Map where lead data goes
A typical lead passes through an ad platform, a landing page, a form tool, a CRM, a texting tool, an AI voice or chat tool and a calendar. Every handoff is a place where data is copied, a connection can break or a lead can sit untouched. Count the tools and the logins. Fewer is easier to secure and easier to fix.
2. Know who is accountable
If the ad vendor blames the CRM, and the CRM blames the AI tool, nobody owns the missed lead. Pick a setup where one team answers for the result from ad click to booked appointment.
3. Check consent at the source
Look at the form or ad that captures the lead. Does it collect consent in a way your attorney is comfortable with? The AI follows the permission the lead gave. It does not create it.
4. Look at what happens after launch
AI agents need scripts adjusted, objections added and calendars synced. A platform you configure once and leave alone drifts. A managed service keeps tuning.
5. Treat audit reports as one signal
If your own clients or insurers require a vendor security report, factor it in. Otherwise weigh it alongside the points above.
A worked example: five vendors versus one
Picture a home services contractor running Facebook ads. A lead submits a form at 7:40 p.m. Here is the chain with separate vendors: the ad agency’s form feeds a CRM, a separate automation tool triggers a text, a different AI tool calls, and a booking tool puts the appointment on a calendar. Five tools, five sets of permissions, five places for a connection to break silently.
Now picture the same lead in one managed system: ads, AI follow-up and CRM built and run by the same team. The form, the pipeline stage, the call, the text and the calendar slot are all configured to work together, and one team watches whether the lead got booked. Nothing here changes the telemarketing rules. It reduces the number of seams where leads and data get lost.
That is the reasoning behind how we build. You can see the same logic applied to industries in our pages on AI speed-to-lead and intake for law firms and AI customer service for insurance agencies.

Where this matters by industry
Personal injury and MVA law firms. Intake involves sensitive details about injuries and accidents, and speed decides who signs the case. The priority is a fast, consistent first response and a clean handoff into your case management flow. Our pay-per-signed-case model means firms pay only for signed retainers, not leads.
Home services contractors. Roofing and HVAC leads go cold in minutes, and callers want an appointment, not a conversation. The risk here is missed calls and slow follow-up, not data science. Booked appointments are the measure.
Insurance agencies. Policyholder questions and quote requests arrive around the clock. AI handles tier-one requests and routes the rest, and your CRM should record every interaction.
Local service businesses running paid ads. If you pay for every click, a lead that never gets a call-back is wasted spend. Follow-up speed and CRM hygiene protect the budget.

How MediaBloom delivers this done-for-you
MediaBloom is a Tampa, Florida digital marketing and AI agency. We have worked with 300+ businesses since 2022 and hold a 4.8-star Google rating. We are not a software box you configure alone. We build the ads, the AI agents and the CRM, and our team books the appointments.
That covers four areas:
- AI systems: voice agents, AI receptionists, speed-to-lead and database reactivation that call, text and book appointments 24/7. See AI voice agents for inbound and outbound calls.
- CRM optimization: GoHighLevel setup, pipelines, automations and follow-up so no lead slips through. See our CRM automation and GoHighLevel setup service.
- Advertising: Facebook, Instagram, Google and ChatGPT ads, with our setters booking qualified appointments. See ads management.
- Pay-per-signed-retainer case acquisition for personal injury and MVA law firms.
Because one partner covers lead generation, AI follow-up and CRM, nothing falls through the cracks between vendors. If you want to see how the pieces fit, our guide to setting up AI lead capture walks through the setup.
Ready to put AI on your leads?
If your leads are slipping between ads, phones and a CRM, the fix is a single system someone runs for you. MediaBloom builds it and manages it, with campaigns live in days. Start with AI Voice Automation and tell us which of the four areas you need first.

Pricing is set after a conversation about your lead volume and tools, since the scope differs for a PI firm, a roofer and an insurance agency.
Our perspective
We hear “is it SOC 2?” more often as AI tools spread. It is a fair question about data handling. It is only one question.
From where we sit, the question that decides whether AI pays off for a PI firm, a roofer or an agency is simpler: when a lead comes in, does someone call it back within minutes, qualify it and book it? Most leaks we think about are operational. A form that does not trigger a text. A pipeline stage nobody updates. A voice agent that was never connected to the calendar.
That is why we build the whole chain ourselves: the ads, the AI agents, the CRM and the appointment setting. When one team owns every step, there are fewer vendors holding your data, fewer logins to manage and one place to look when a lead is missed.
For law firms, we go a step further. You pay for signed retainers, not leads, so our incentive is the same as yours: cases that sign.
MediaBloom
FAQ
Does SOC 2 cover AI?
SOC 2 does not have a dedicated AI section, but its five criteria apply to any system, including one that uses AI. A company can ask its auditor to include AI-related controls in scope. The report speaks to data protection, not to how accurate or effective the AI is.
Is SOC 2 legally required?
No. It is a voluntary report issued by an independent CPA firm under AICPA standards. Companies obtain it because customers and partners ask for outside assurance. Laws that apply to AI calling and texting, such as the TCPA, are separate from SOC 2.
What is the difference between SOC 2 Type 1 and Type 2?
Type 1 reviews whether controls are designed properly at one moment. Type 2 tests whether the controls worked over a period of time. Type 2 gives a fuller picture because it shows ongoing operation.
Do AI voice agents have to follow telemarketing rules?
Yes. The FCC ruled in 2024 that AI-generated voices are artificial voices under the TCPA, so consent rules for artificial voice calls apply. Talk to your attorney about the consent wording on your forms and the lists you call.
Sources
- AICPA & CIMA: SOC 2 reports and the Trust Services Criteria
- FCC: AI-generated voices in robocalls under the TCPA



