Freshservice AI limitations: What you need to know in 2026

Freshservice markets itself as an "AI-powered ITSM platform" that delivers enterprise capability without complexity. The reality is a bit more nuanced. While Freshservice's Freddy AI suite offers genuine automation benefits, several significant limitations exist that Freshworks doesn't highlight in its marketing materials.

Let's break down what Freddy AI actually delivers, where it falls short, and what you should consider before committing.

What is Freddy AI?

Freddy AI is Freshservice's artificial intelligence layer, built on top of Microsoft's Azure OpenAI infrastructure. It's positioned as a three-part system:

• Freddy AI Agent: Autonomous conversational AI that handles employee requests through Slack, Microsoft Teams, email, and the self-service portal
• Freddy AI Copilot: Assistance tools for IT agents, including reply suggestions, ticket summarization, and writing assistance
• Freddy AI Insights: Analytics and proactive insights with root cause analysis for IT leaders

Freshworks claims Freddy AI delivers "356% ROI in under 6 months" and "66% ticket deflection with AI-powered self service." These numbers come from their own Benchmark Report 2025 and a commissioned Forrester TEI study. But here's the catch: those results require the full Freddy AI suite, which brings us to the first major limitation.

The Enterprise paywall: AI features aren't for everyone

Here's something Freshservice doesn't make obvious on its pricing page: Freddy AI Agent is Enterprise-only.

If you're on Starter ($19/agent/month), Growth ($49/agent/month), or Pro ($99/agent/month), you don't get the autonomous AI Agent at all. Here's what's actually available:

Source: Freshservice Pricing

The Pro plan only includes Freddy AI Copilot, which drafts replies and summarizes tickets but doesn't autonomously resolve anything. To get the autonomous AI Agent that Freshservice showcases in its marketing (the one that deflects tickets and resolves issues without human intervention), you must upgrade to Enterprise.

And here's the kicker: Enterprise pricing isn't published. You have to "contact us" for a quote. This lack of transparency makes it difficult to budget or compare costs. Based on typical enterprise software pricing patterns, you're likely looking at significantly more than the $99/agent Pro plan, plus potential overage charges for AI usage.

There's also no standalone AI purchase option. You can't buy Freddy AI Agent as an add-on to your existing plan. It's Enterprise or nothing.

Session limits that could stall your operations

Even if you do upgrade to Enterprise, you're not getting unlimited AI usage. Each Enterprise license includes 1,200 AI Agent sessions per year.

A "session" is defined as any interaction a unique user has with Freddy AI Agent within a 24-hour period. So if one employee chats with the AI three times in one day, that's one session. If they come back the next day, that's a second session.

Let's do some math. If you have 100 employees and each interacts with the AI Agent just once per week on average, you're looking at 5,200 sessions per year. With 1,200 sessions per license, you'd need roughly 5 Enterprise licenses just to cover basic usage. And that's assuming relatively light usage.

What happens when you hit your limit? According to Freshservice's documentation, Freddy AI Agent continues operating to prevent service disruptions, but "additional sessions will be charged." The rates aren't publicly disclosed. You have to contact your Customer Success Manager to find out what those overages cost, and to increase your limits.

There's no rollover of unused sessions either. If you have a slow month, those sessions don't carry over. They reset at the start of each billing cycle.

Source: Freddy AI Agent Pricing FAQs

API rate limiting

Beyond AI sessions, Freshservice also imposes a 1,000 API calls per hour limit across your entire account. This affects integrations, custom workflows, and any automation that relies on Freshservice's API.

When you hit this limit, Freshservice returns an error with a "retry-after" header telling you how long to wait. For teams running heavy automation or integrations with multiple tools, this can create bottlenecks. Your workflows might pause mid-process, waiting for the rate limit to reset.

Source: Freshservice API Documentation

Customization constraints users actually complain about

Freshservice's Gartner Peer Insights reviews reveal real-world frustration with customization limitations. The platform has a 4.4/5 rating overall, but the critical reviews highlight specific pain points.

Everything is locked behind a paywall and even then features are fairly rigid and custom functionality etc is not really possible

Gartner Peer Insights

Other users report:

• "Portal setup can feel somewhat cumbersome, with limited customisation before diving into CSS"
• "Form logic is pretty basic... lack of flexible dynamic fields or conditional behaviour"
• "If you need to build highly specific, complex reports, you may hit a wall"
• "Higher complexity to the product than initially promised"

Source: Gartner Peer Insights - Freshservice Reviews

The pattern here is clear: Freshservice works well out of the box for standard ITSM workflows, but teams with specific customization needs often find themselves working around the platform rather than with it.

Data privacy and residency concerns

Freshservice's approach to AI training raises several privacy considerations that aren't immediately obvious.

By default, your data is used to train Freddy AI's "collaborative models." These are models trained on de-identified data from multiple Freshservice customers to improve shared AI performance. While Freshworks states they never use your data to train third-party models (like OpenAI's), they do use it for their own collaborative improvements.

Want to opt out? You can't do it through the admin panel. You have to email support@freshworks.com and request an opt-out. This isn't a self-service toggle; it's a manual process that requires contacting support.

There's also a data residency issue for non-US/EU customers. If you're hosted outside these regions, your data may be transferred to Microsoft's Azure OpenAI service in the US for processing. Freshworks notes that "as Azure OpenAI becomes available in your respective region, Freshworks will localize the service," but there's no timeline for this.

Source: Freddy AI Data and Safety FAQs

How eesel AI addresses these Freshservice AI limitations

If you're evaluating AI for IT service management, it's worth looking at how alternatives approach these same challenges. At eesel AI, we've taken a different approach to AI implementation that addresses many of the limitations Freshservice users encounter.

Transparent pricing without Enterprise gates

Our AI Agent is available on our Business plan at $639/month (annual billing), not hidden behind an Enterprise paywall. You know exactly what you're paying upfront. No "contact us for pricing." No surprises.

We also price by AI interactions, not per agent. This means your costs scale with actual usage, not headcount. Our Business plan includes 3,000 interactions per month, and you can see exactly how you're tracking against that limit in real-time.

No arbitrary session limits

Instead of counting "sessions" (a 24-hour window per user), we count individual AI interactions. This is a more granular and fair approach. A user who asks three quick questions counts as three interactions, not one session that might have covered dozens of exchanges.

If you do hit your limit, we don't charge surprise overages. You'll get a notification, and the AI pauses until you upgrade or your cycle resets. No hidden fees.

Plain-English customization

Rather than forcing you into rigid configuration options or CSS workarounds, eesel AI lets you customize behavior using plain English prompts. Want the AI to handle refund requests differently for VIP customers? Just write: "For VIP customers, CC the account manager on all refund-related responses."

No coding. No complex workflow builders. Just describe what you want in natural language.

Full data control

Your data is never used to train general AI models. This is contractually guaranteed with our subprocessors, including OpenAI. We also offer EU data residency on our Business plan, so your data never leaves the region if that's a requirement for your organization.

Opting out of features is self-service through your dashboard. No support tickets required.

Progressive rollout with confidence

One of our core principles is that AI should start with guidance and level up to autonomy based on performance, not based on which plan you bought. You can run simulations on thousands of past tickets before going live, measure quality, identify gaps, and gradually expand the AI's scope as it proves itself.

This "teammate" model means you verify the AI understands your business before giving it more responsibility. Mature deployments achieve up to 81% autonomous resolution, with a typical payback period under 2 months.

Choosing the right AI for your IT service desk

Freshservice's Freddy AI has genuine strengths. The platform is well-established, the integration with Microsoft Teams and Slack is polished, and for teams that fit its workflow patterns, it works well. The 4.4/5 Gartner rating from 193 reviews suggests most users are satisfied.

But the limitations are real:

• Freddy AI Agent is Enterprise-only with opaque pricing
• 1,200 session limit per license per year with undisclosed overage rates
• 1,000 API calls per hour can bottleneck integrations
• Customization requires workarounds for non-standard needs
• Data opt-out requires manual support contact

If you're a mid-sized IT team looking for autonomous AI capabilities, the jump from Pro ($99/agent) to Enterprise (unknown, but likely significantly higher) just to access AI Agent features is a substantial barrier.

For teams that want AI autonomy without the Enterprise gate, transparent pricing, and the ability to customize through natural language rather than code, eesel AI offers a different approach. Our Business plan includes full AI Agent capabilities, unlimited bots, and the ability to test and roll out gradually with confidence.

The right choice depends on your specific needs, budget transparency requirements, and how much customization flexibility your workflows demand.