AI is a world-class co-pilot, but a dangerous auto-pilot. While Large Language Models (LLMs) can synthesize tax research in seconds, they are also designed to be agreeable, which means they will confidently cite a tax code that doesn’t exist if it fits the pattern of the conversation.

In this episode of Firm Ground, Jatin Narang (Founder of Verito) tackles the operational side of AI compliance. We move past the technical settings and dive into the Human-in-the-Loop framework: the non-negotiable process every tax firm needs to ensure AI efficiency never comes at the cost of professional accuracy.

In this episode, we cover:

• The Hallucination Trap: Understanding why AI makes things up and why confidence in a model's tone is its most dangerous trait.
• The 3-Step Framework: A tactical workflow for moving from an AI-generated draft to a final, compliant client deliverable.
• Algorithmic Bias: How to spot when a model’s training data might be skewing your tax planning or advice.
• The Transparency Standard: Best practices for disclosing AI usage to clients without undermining your firm’s value.
• Staff Accountability: How to train your team to maintain professional skepticism in an automated environment.

Key Takeaway: Accuracy is the product you sell. Efficiency is just how you produce it. This episode provides the roadmap for balancing both without violating your professional standards.

🔒 Need help getting compliant?
→ VeritShield WISP (audit-ready in 5 business days): verito.com/wisp 
→ VeritSpace Private Cloud Hosting: verito.com/hosting 
→ VeritGuard Managed IT + Compliance: verito.com/managed-it 

Firm Ground is a series from Verito Technologies for tax preparers and accountants navigating technology, compliance, and security. New episodes cover the tools, regulations, and risks that matter to your practice, explained in plain English. 

Subscribe: Be sure to follow Firm Ground for more insights on the intersection of tax, technology, and security.

If your tax firm is using free AI accounts, you aren't the customer, your client data is the product. Most practitioners assume that a simple delete chat clears their liability, but the reality of model training is much more complex.

In this episode of Firm Ground, Jatin Narang (Founder of Verito) moves past the theory and into the tactics. He breaks down the exact Opt-Out architecture required to keep your firm compliant with IRS and FTC standards while still leveraging the power of LLMs.

In this episode, we cover:

• The Opt-Out Architecture: Why standard data controls often fail and how to use the Privacy Portal to ensure your data stays private.
• Consumer vs. Enterprise: The massive security gap between the $20/month Plus accounts and true Enterprise-grade AI.
• The API Advantage: Why connecting via API is inherently safer than using the standard chat interface for firm-wide deployment.
• The Training Kill-Switch: How to verify, not just hope, that your data is being excluded from future model training.
• Compliance Check: How these settings map back to your Written Information Security Plan (WISP).

Key Takeaway:
Don't wait for a data breach to audit your AI settings. This episode provides the 2026 roadmap for securing your digital practice.

🔒 Need help getting compliant?
→ VeritShield WISP (audit-ready in 5 business days): verito.com/wisp 
→ VeritSpace Private Cloud Hosting: verito.com/hosting 
→ VeritGuard Managed IT + Compliance: verito.com/managed-it 

Firm Ground is a series from Verito Technologies for tax preparers and accountants navigating technology, compliance, and security. New episodes cover the tools, regulations, and risks that matter to your practice, explained in plain English. 

Subscribe: Be sure to follow Firm Ground for more insights on the intersection of tax, technology, and security.

Click here to watch a video of this episode.

In the inaugural episode of Firm Ground, Jatin Narang (Founder of Verito) dives into the hidden risks of the AI tools we use every day. From the new QuickBooks-Claude integration to the training trap hidden in consumer AI accounts, Jatin breaks down the compliance lines tax and accounting firms cannot afford to cross.

In this episode, we discuss:

• The Training Trap: Why ChatGPT and Claude train on your conversations by default and the specific steps to opt out.
• The Pentagon Controversy & Claude: How consumer accounts quietly shifted data policies and what it means for your privacy.
• QuickBooks AI (MCP): The risks of direct data integration and why it’s more dangerous than simple copy-pasting.
• The Regulatory Reality: Why IRS Publication 4557, the FTC Safeguards Rule, and GLBA apply to AI tools, even if they don’t mention AI by name.
• The 5-Step Guardrail Checklist: A tactical framework every firm should implement this week to mitigate Shadow AI risks.

Key Statistics Mentioned:

• $4.63M: The average cost of a Shadow AI data breach (IBM 2025).
• 295%: The spike in ChatGPT uninstalls following the Pentagon data deal (Feb 2026).
• 77%: The percentage of employees currently using AI tools without official company oversight.

Sources referenced in this episode: 

📄 IBM 2025 Cost of a Data Breach Report — ibm.com/reports/data-breach
📄 LayerX 2025 Enterprise AI Report — go.layerxsecurity.com/the-layerx-enterprise-ai-saas-data-security-report-2025
📄 Anthropic Privacy Center (MCP + Training) — privacy.claude.com/en/articles/10023580-is-my-data-used-for-model-training
📄 Anthropic Consumer Terms Update — anthropic.com/news/updates-to-our-consumer-terms 📄 Intuit-Anthropic Partnership — investors.intuit.com/news-events/press-releases/detail/1305 📄 IRS Publication 4557 — irs.gov/pub/irs-pdf/p4557.pdf
📄 FTC Safeguards Rule — ftc.gov/legal-library/browse/rules/safeguards-rule
📄 Samsung ChatGPT Data Leak — cybersecuritydive.com (April 2023)
📄 ChatGPT 295% Uninstall Spike — techcrunch.com/2026/03/02/chatgpt-uninstalls-surged-by-295-after-dod-deal

🔒 Need help getting compliant?
→ VeritShield WISP (audit-ready in 5 business days): verito.com/wisp 
→ VeritSpace Private Cloud Hosting: verito.com/hosting 
→ VeritGuard Managed IT + Compliance: verito.com/managed-it 

Firm Ground is a series from Verito Technologies for tax preparers and accountants navigating technology, compliance, and security. New episodes cover the tools, regulations, and risks that matter to your practice, explained in plain English. 

Subscribe: Be sure to follow Firm Ground for more insights on the intersection of tax, technology, and security.

In this episode of Beyond Downtime: The Verito Advantage, we break down how tax and accounting firms can outsource work without creating security or compliance risks.

This conversation comes from a recent NATP session and is more relevant now than ever, with firms navigating IRS requirements, FTC Safeguards updates, and rising pressure to keep client data protected.

You will learn what a secure outsourcing process actually looks like, how to evaluate third-party vendors, what red flags to avoid, and how to protect your firm’s reputation when you rely on outside help.

We also cover the practical steps firms take to stay compliant, build a stronger security posture, and keep operations moving during busy season.

Want IT that protects your data and supports your firm every day?
Start here → Verito.com

In our debut episode of Beyond Downtime: The Verito Advantage, we take you back to Thomson Reuters SYNERGY 2024, where industry leaders gathered to discuss the future of tax and accounting technology.

Recorded live at last year’s event, this conversation is packed with more relevant insights than ever, from preparing your firm for the FTC Safeguards Rule to what a WISP really needs to include, to the practical tech upgrades firms are making to minimise risk and maximise performance.

We also share how firms reduce IT stress and elevate their client experience with smarter, more secure cloud infrastructure.

Want to protect your firm with IT that just works, securely?
Start here → Verito.com

eAFT6YA5vV1L5RGE3HIo