Freshsales AI in 2026: What Freddy Actually Does, and Where It Stops
Freshsales ships its own AI, and it’s called Freddy. If you searched “Freshsales AI” you almost certainly landed on Freshworks’ Freddy AI page, and the honest answer to “is Freshsales AI real” is yes — contact scoring, deal risk signals, next-best-action nudges, and generated sales emails are all in the product today. The question worth asking is a different one: does any of it solve the reason your pipeline is unreliable?
For most SMB sales teams on Freshsales, it doesn’t. Freddy is a ranking and drafting layer. The thing eating your forecast is that a rep talked to a prospect on Tuesday and nobody typed it into the CRM. Freddy scores whatever is in the record. If nothing is in the record, Freddy scores nothing.
What Freddy AI actually does inside Freshsales
Four capabilities, and it’s worth being precise about each:
- Contact scoring and customer fit. Freddy analyzes engagement signals — email opens, link clicks, call outcomes, activity history — plus historical contact properties, and produces a score that shifts as behavior changes. Freshworks documents this as on by default at the account level once you’re on a qualifying plan, with signals you can tune yourself in contact scoring settings.
- Deal insights and risk signals. Freddy reads deal activity — emails, calls, stage movement — and buckets open opportunities by health: trending, at risk, likely to close, gone cold. Inside a deal it surfaces the why (“no response in five days”) rather than just the label.
- Next-best-action. Off the back of score and fit rating, Freddy proposes the specific move to advance the contact.
- Email generation. Subject line and body for prospecting, with tone selection. Useful. Also the least differentiated thing on the list, since any general AI assistant writes a prospecting email.
Notice what all four have in common: they are advisory. Freddy tells a rep which deal is rotting. It does not go get the missing information, and it does not do the follow-up.
The plan and pricing reality
This is where a lot of “Freshsales AI” evaluations go sideways, because the AI is split across the base plans and a metered add-on.
Freshworks’ CRM pricing currently runs Growth at $9/user/month, Pro at $39, and Enterprise at $59, billed annually. The Freddy features described above — contact scoring, deal insights, the AI email writer — sit on Pro and above. Accounts created before 4 March 2024 keep some of it on Growth as a grandfathered carve-out, which is why you’ll see contradictory answers in forums: two teams on “Growth” genuinely have different feature sets depending on signup date.
Separately, Freddy AI Agent is a metered add-on billed at $49 per 100 sessions, and Freddy Copilot is a per-seat flexi add-on starting around $29/agent/month that you can buy for a subset of users rather than the whole team. So the real cost of “Freshsales with AI” for a six-rep team is not the $9 headline that got you interested in Freshsales in the first place — budget Pro seats plus whatever agent sessions you consume.
If you chose Freshsales over Salesforce and HubSpot specifically on price and simplicity, that math matters. It’s also worth sanity-checking against the wider field before you upgrade; we keep a running comparison of the best AI CRM tools and AI automation options for CRM-adjacent stacks.
The Freshsales API: genuinely open, and you can self-serve the key
Unlike a lot of vertical software, Freshsales does not gate its API behind a partner agreement. Any user can generate credentials:
Click your profile picture → Profile Settings → API Settings. Your API key is displayed there, along with your bundle alias — you need both. The direct path is https://yourdomain.freshsales.io/personal-settings/api-settings.
From there, the Freshworks CRM API is a straightforward REST surface:
- Base URL:
https://<bundle-alias>.myfreshworks.com/crm/sales/api/<resource> - Auth header:
Authorization: Token token=<your-key> - Read and write on contacts, accounts, deals, notes, tasks, appointments, and sales activities, plus marketing lists, search across entities, and CPQ products and documents.
- Rate limit: roughly 1,000 requests per hour per account before you get a 429.
That is a good API. It means anything you’d want an AI assistant to do to your pipeline — create a contact, log a call as a sales activity, move a deal stage, attach a note, book a task — is reachable programmatically, by you, today, without a sales call.
What ChatGPT and Claude can and can’t reach
Neither ChatGPT nor Claude has a native Freshsales connector. That’s the plain answer.
Freshworks does run its own MCP server, but it’s scoped to Freshservice and Freshdesk, it’s in beta behind an Early Access Program, and it’s Enterprise-plan only — see the Freshservice MCP documentation. Freshsales is not part of it. We’ve covered the same situation from the support side in our Freshdesk MCP and Freshservice MCP writeups.
What exists for Freshsales is third-party MCP wrappers — Apideck, viaSocket, Composio and similar — that take your Freshsales API key and expose CRM operations as tools to Claude or ChatGPT. They work. They also mean handing your CRM credentials to a middleman, and they inherit the limitation that matters most:
Scheduled work is no longer the dividing line — ChatGPT and Claude both run tasks on a timer without anyone watching. The line that still holds is event triggers. Nothing kicks off because something happened in one of your apps; it kicks off because the clock said so, which means anything time-sensitive waits for the next run.
For a sales team, that limitation is the whole problem. Nothing fires when a deal sits untouched for nine days. Nothing fires when a call ends. A rep still has to remember to open a chat and ask.
The actual failure mode: CRM hygiene, not CRM intelligence
Talk to a five-person sales team on Freshsales and the complaint is never “our lead scores aren’t smart enough.” It’s:
- A rep had a 40-minute discovery call and logged “good call” as the note, three days later.
- Two deals in the pipeline haven’t had an activity logged since May, and nobody noticed until the pipeline review.
- The follow-up email a prospect was promised on Thursday went out the following Wednesday, or not at all.
- Half the deal stages are wrong, so the forecast is fiction — and Freddy’s forecast is built on that same fiction.
Freddy AI is downstream of every one of these. Better scoring on stale data produces confident, wrong prioritization. That’s arguably worse than no scoring, because the team trusts it.
The fix isn’t a smarter model reading your CRM. It’s something that writes to your CRM without a human remembering to, and that acts on time and on events rather than when someone opens a chat window.
Where an outside AI assistant fits alongside Freddy
This is the shape of the useful setup: keep Freddy for what it’s genuinely good at — scoring and in-CRM drafting — and put an agent outside Freshsales that handles the logging and follow-through Freddy was never designed to do.
Carly is an AI executive assistant you set up by conversation instead of code. Freshsales isn’t one of Carly’s native connectors, but because Freshsales exposes that open REST API, it’s a bring-your-own-key integration: you paste your Freshsales API key at carlyassistant.com/integrations and Carly works against the same contacts, deals, notes, tasks, and sales activities endpoints described above.
What that unlocks for a Freshsales team:
- Meeting notes become CRM records. After a call, the summary lands as a note on the right deal and the next step becomes a task — the same pattern we describe in having an AI agent update your CRM after meetings.
- Email threads get logged. A reply from a prospect turns into a sales activity plus a stage nudge, rather than living only in someone’s inbox. See an AI email agent that updates your CRM.
- Stale deals get chased. A recurring sweep across your pipeline surfaces deals with no activity in N days and drafts the follow-up — and Carly sends across Gmail and Outlook, so the follow-up actually leaves.
- It runs on events, not on you. Carly fires when the email arrives or the meeting ends, in the cloud, without your laptop open.
Pricing: free, unlimited Zapier-style workflows; AI agents from $35/month. For a Freshsales team weighing $49 per 100 Freddy Agent sessions on top of Pro seats, it’s worth pricing both. If you’d rather build the logging layer yourself, we wrote up how to build an AI CRM agent and why we built CRM handling into Carly in the first place.
FAQ
Is Freddy AI included in Freshsales, or is it an add-on? Both, depending on the feature. Contact scoring, deal insights, and AI email writing come with Pro ($39/user/month) and Enterprise ($59). Freddy AI Agent is a separate metered add-on at $49 per 100 sessions, and Freddy Copilot is a per-seat add-on starting around $29/agent/month that you can buy for only some of your users.
Can I use Freshsales AI on the $9 Growth plan? Generally no — the Freddy features sit on Pro and above. Accounts created before 4 March 2024 are grandfathered into some scoring functionality on Growth, which is why answers vary depending on who you ask.
Does Freshsales have an API I can use myself? Yes. Go to Profile Settings → API Settings to get your API key and bundle alias. The API supports read and write on contacts, accounts, deals, notes, tasks, appointments, and sales activities, at roughly 1,000 requests per hour per account.
Is there an official Freshsales MCP server for Claude or ChatGPT? No. Freshworks’ MCP beta covers Freshservice and Freshdesk only, on Enterprise plans through an Early Access Program. Freshsales connections go through third-party MCP wrappers or a direct API integration using your own key.
Will Freddy AI log calls and update deals for my reps? No. Freddy scores, flags risk, recommends next actions, and drafts emails. Data entry is still on the rep, which is why the scores degrade when logging slips.
What happens to Freddy’s deal insights if our CRM data is stale? They degrade quietly. Freddy analyzes recorded activity, so a deal with no logged activity reads as gone cold whether or not the rep spoke to the buyer yesterday. Fixing the logging is a prerequisite for trusting the scoring.
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
