Freshsales Review: Cheap to Start, and Three Ceilings You Cannot Pay Past

Freshsales is Freshworks' CRM, and before anything else you need to be sure which product you are reading about. Freshworks also sells Freshsales Suite, a different product under a nearly identical name, and its own app marketplace lists the two as separate filters^[2]. Several of the review pages and help articles you will find by searching the name belong to the Suite rather than to the CRM. We separated them, and it changed the numbers more than once.

Once that is settled, the product is easy to describe. It is a CRM with a phone, a chat widget and a chatbot builder inside it, starting at nine dollars a seat, with a free plan for up to three people that the pricing page does not mention.

OUR TAKE: The capability is real and unusually broad for the price, and the vendor documents its plans better than anyone else in this category. Two things keep the score down, and both are structural rather than cosmetic. Support is identical on every plan, including the free one, which means the thing reviewers complain about most is the one thing you cannot fix by spending more. And the cheap plan is a shop window: automation, AI, reporting and relational data all begin on the middle plan, at more than four times the advertised entry price.

## What it actually is

A sales CRM that has swallowed a contact centre. Contacts, accounts, deals and a pipeline sit at the base, as they do everywhere. What is unusual is what comes with them at the bottom of the ladder: a built-in phone with call recording, call masking and voicemail drop, a chat widget, and a chatbot builder with intent detection, multilingual bots and handover to a human^[2].

Rivals at this price sell you a pipeline and expect you to bolt telephony on from somewhere else. Freshsales treats the phone as part of the product. For a team whose day is calls, that is a real difference and it is the main argument for choosing it.

What it is not is a system of record for a company with complicated data. The relational pieces arrive late, and we will come to exactly where.

## The four plans, including the one you will not see

Every figure is per user, per month. The second column is the rate when the year is paid up front, which is the figure the page loads with.

Plan	Month to month	Billed annually	What it is
Free	nothing	nothing	up to three users
Growth	$11	$9	the first plan with workflows and an API
Pro	$47	$39	automation, AI and reporting start here
Enterprise	$71	$59	governance and audit

The free plan is the interesting one, for two reasons.

The first is that it is not on the pricing page at all. The pricing page shows Growth, Pro and Enterprise. The free plan appears on a separate comparison page, and in the pricing page's own search description, which states that a free plan is available^[2]. A buyer who arrives at the obvious URL and reads it carefully will come away believing the product starts at nine dollars.

The second reason is what the free plan contains, and here the vendor's own documentation lets us be precise rather than impressionistic. Freshworks publishes a feature comparison with 136 rows, each carrying an explicit included-or-not flag for each of the four plans. The free plan is included on 44 of those rows. On 92 of them, the three paid plans differ from one another. Those two numbers add to 136 exactly^[2].

That arithmetic has a meaning. The free plan carries every feature common to all four plans and not one single feature that distinguishes any paid plan from any other. It is not a sample of the product. It is the intersection of the product: the floor, defined precisely.

Concretely, that means no workflows of any kind, no API access, no file storage, no sales dashboard, and a custom field allowance the vendor gives three different answers for^[2][4]. You can keep contacts and deals in it, move a deal along one pipeline, sync email both ways and make calls. Nothing happens on its own and nothing can be connected to anything else.

Where the product actually begins

The honest reading of the ladder is that there are two meaningful steps and only one of them is advertised.

From free to Growth, nine dollars a seat buys 38 further features, and the important ones are basic workflows, API access, the marketplace, the sales dashboard and curated reports^[2]. That is the step from a list to a system, and at nine dollars it is good value.

From Growth to Pro, thirty more dollars a seat buys 42 features; and it is here that everything a buyer is usually shopping for arrives at once. Every Freddy AI insight: contact scoring, deal insights, next best action, deduplication, out-of-office detection. Advanced workflows. Multiple pipelines. Sales sequences. Custom reports. Territory management. Multi-currency. Sales forecasting. The webhook and data APIs. The power dialer^[2].

That is 4.3 times the price, and it is the real product boundary. If you are comparing Freshsales against rivals on the strength of its entry price, you are comparing a contact database with twenty workflows against their full product.

From Pro to Enterprise, twenty more dollars buys nine features: field-level permissions, audit logs, record types, custom modules, automatic profile enrichment, forecasting insights, deal teams, the reporting extract APIs and an extendable API rate limit^[2]. Measured per feature gained, that is the most expensive step in the ladder by a factor of three. It is also the only place the governance features live, and those are not the kind you can add later without moving the whole account.

## The ceilings that do not move

Most pricing ladders have at least one limit that stays flat no matter what you pay. Freshsales has four, and together they describe the shape of the product more honestly than the feature lists do.

Support is 24x5 on every plan. The support section of the vendor's 136-row matrix contains exactly one row, and it is included identically on the free plan and on the most expensive one^[2]. There is no premium tier, no named contact and no faster response time purchasable at any price on the published ladder.

The per-minute API limit is 400 on every plan. The hourly ceiling does climb, 1,000 calls an account an hour, then 2,000, then 5,000; but the per-minute figure is 400 across the board^[5]. For any burst shorter than an hour, the fifty-nine-dollar plan behaves exactly like the nine-dollar one.

The chatbot allowance is 500 sessions, and it is not monthly. The comparison matrix lists it as a plan feature on each paid tier, beside genuinely recurring limits such as daily email caps. The pricing page's own FAQ says those 500 sessions are limited to once per account^[1]. It is a trial grant presented in the same column as a quota. Worse, the vendor's documentation confirms that previewing a bot while you build it consumes sessions, published or not^[7]; so a careful team can spend its one-time allowance before a customer ever sees the bot.

Workflows are capped and then sold back. Twenty on Growth, fifty on Pro, a hundred on Enterprise; and every plan, including the top one, lists additional workflows as a paid add-on at five dollars for ten^[2]. The cap is not a boundary that pushes you upward. It is a meter.

The data model arrives late

This is the limit that will decide whether Freshsales fits, and it is the one furthest from the pricing page.

Custom fields are counted per entity, separately for contacts, accounts, deals and custom modules; and the ladder is 150 on Growth, 250 on Pro and 630 on Enterprise^[4]. In raw volume that is generous; several rivals cap custom fields at a single flat number across the whole account.

The catch is in the breakdown. Lookup fields and formula fields, the two field types that let one record refer to another, do not exist below Pro at all^[4]. On the nine-dollar plan your records cannot be related to each other. For a simple pipeline that does not matter. For anything where a deal needs to point at a project, or an account at a parent company, it is the end of the road, and you will discover it after your data is already inside.

Downgrading, if you ever need to, is blocked by your own data: fields above the lower plan's limit must be deleted before the move, and the vendor spells out the arithmetic with an example^[4].

## What the reviews say, and why they disagree so violently

Freshsales produces the widest disagreement between rating sources we have measured anywhere.

Where buyers review software, it does well. Capterra gives 4.5 across 623 reviews, with 404 of them at five stars^[9]. G2 gives 4.5 across 1,234^[11]. TrustRadius gives 8.5 out of ten across 137.

Where consumers review companies, it does badly. Trustpilot gives 1.3 across 120 reviews^[10]. That is a gap of 3.2 points on the same five-point scale, and it is the widest in our corpus.

We did not average it away. Trustpilot has no page for Freshsales: only one for Freshworks, which also sells Freshdesk, Freshservice, Freshchat and Freshcaller. So we read a hundred of those reviews and attributed each one to a product. Twenty-two name Freshsales. Three name Freshdesk. Three name a telephony or chat product. Seventy-four name no product at all.

The twenty English-language reviews that name Freshsales and no sibling product average 2.9: a long way above the company's 1.3, and still a long way below the software aggregators. Twenty is a small sample and we are not going to pretend otherwise. But the direction is consistent, and so is the subject.

Sixty per cent of the English-language reviews we read are about support. That is the highest single-theme share we have measured for any product in any category, and those reviews average 1.9 stars. Set that beside a support entitlement of 24x5 on every plan, with nothing better available at any price, and the two facts stop being separate observations. The vendor's own pricing page confirms what its reviewers are saying: there is nothing to buy.

The one genuinely cheerful signal is setup, which is the best-received theme in the product-level subset at 4.0 stars. People get it running. What happens afterwards is where the complaints start.

What the search data says about who is looking

We measure demand before we write, and Freshsales produced a result worth repeating because it says something about the product that no feature list does.

Across three hundred keyword candidates built from the product's name, only five survived with any volume at all. The largest term in the entire commercial tail is freshsales pricing, at seventy searches a month. freshsales review has twenty. freshsales alternatives has ten.

The largest tail term of any kind is freshsales login, at a hundred and seventy: more than twice the pricing term and more than eight times the review term.

There is a parallel trap in the name. The term freshworks carries 14,800, nearly ten times the product's own 1,600. But Freshworks is the company, and it also sells Freshdesk, Freshservice, Freshchat and Freshcaller. That volume is not Freshsales demand; it is four products' worth of people, most of whom are not looking for a CRM.

Two conclusions follow. First, most of the people searching this product's name already own it, and are trying to get in. Second, if you are comparing this product against rivals on the strength of how much is written about it, you are measuring the wrong thing: there is comparatively little written about it because comparatively few people are shopping for it.

◆Our scorecard

3.3/5

AI 3.2/5

Breadth is genuine and the placement is honest about itself. Freddy AI covers contact scoring, deal insights, deduplication, next best action, out-of-office detection and email rewriting, with forecasting insights held back for the top plan. The chatbot builder is unusual in this category for being on the free plan, with ten of its features including intent detection, multilingual bots and agent handover. Three things hold the mark down. Every Freddy insight begins on the middle plan, so the advertised entry price includes no AI at all beyond the chatbot. The 500 chatbot sessions that the comparison matrix lists as a plan feature on each paid tier are, by the vendor's own FAQ, a one-time grant per account rather than a monthly allowance — and building a bot spends them, because previewing an unpublished bot consumes sessions. And the AI agent beyond that is metered at $49 per hundred sessions, where a session is a whole 24-hour window with one person, with the documentation stating explicitly that an unlimited number may be purchased and no spending ceiling documented anywhere.

Value for money 3.2/5

The published ladder is attractive at the bottom and the bottom is not what it appears to be. $9 per seat per month on annual billing is among the cheapest real entries in the category, and a free plan for up to three users exists on top of that. But the free plan carries 44 of the vendor's 136 published features and not one of the 92 on which the paid plans differ from each other, with no workflows and no API at all. And the entry paid plan has no sales sequences, no custom reports, no Freddy insights and a single pipeline, so the features a buyer is shopping for all sit on the middle plan at 4.3 times the price. Two further deductions, both measured: the badge beside the billing toggle claims a flat 20% annual saving and the real figures are 18.2%, 17.0% and 16.9%, so the claim is met on no plan; and price is the most negatively received theme in the reviews we read, at 1.2 stars across a quarter of the product-level subset, the worst price sentiment we have measured in this category. The heaviest deduction is the contract that neither pricing page mentions: payment obligations are non-cancelable and non-refundable, seats cannot be reduced nor plans downgraded during a term, renewal is automatic unless cancelled thirty days ahead, and the agreement reserves an explicit right to raise the fee at renewal. A monthly price displayed beside a per-seat figure suggests a flexibility the terms do not grant.

Capability & depth 3.7/5

Unusually broad for the price, and the breadth is documented rather than claimed: the vendor publishes a 136-row feature matrix with an explicit included-or-not flag for every plan, which is the clearest plan documentation in our CRM corpus. Telephony, chat and a chatbot are built into the CRM rather than integrated, with call recording, call masking and voicemail drop from the first paid plan and a power dialer, warm transfer, conference calling and bring-your-own-carrier from the middle. Depth at the top is governance rather than capability: field-level permissions, audit logs, record types and ten custom modules. Two real ceilings. Workflows are capped on every plan — 20, then 50, then 100 — and additional ones are sold back at $5 for ten, which makes the cap a meter rather than a plan boundary. And the data model is generous but late: 150, 250 then 630 custom fields per entity is more room than rivals that cap at one flat number, yet lookup and formula fields — the two that let one record reference another — do not exist below the middle plan, so the cheap plan's records cannot be related to each other.

Ease of use 3.6/5

The evidence is modestly positive and the sample is small, so the mark is modest too. Setup is the best-received theme in the product-level reviews we read, at 4.0 stars, and complexity is raised in only one of them. Capterra's 623 reviewers give 4.5 overall with 404 of them at five stars. Migration from another CRM and data import are on the free plan, which lowers the cost of trying the product properly, and the trial is 21 days of the full product rather than a restricted version. Against that, the interface draws occasional criticism at 3.0 stars, and the product spans CRM, telephony, chat and a chatbot builder in one interface, which is more surface to learn than a focused sales tool presents.

Integrations 3.0/5

Scored on what we could read, which is less than the vendor implies. API access is included from the first paid plan, which is better placement than several rivals, and the Freshworks marketplace is available from that plan too. But webhooks, the extensive data APIs and the rich messaging APIs all begin on the middle plan, and the reporting extract APIs and an extendable rate limit are held for the top one — so the cheap plan has an API it cannot push events out of. The hourly rate limit does climb — 1,000, 2,000 then 5,000 calls an account an hour — but the per-minute limit is 400 on every plan including the most expensive, so for any burst shorter than an hour the $59 plan behaves exactly like the $9 one. The top plan's advertised extendable rate limit is nowhere defined, and the only documented way to raise a limit is to email support, which any plan can do. Two of the vendor's own articles also disagree on whether the limit is per account or per agent, a factor of ten on a ten-seat team. And the marketplace publishes no application count on any page we could read: its own filter lists Freshsales and Freshsales Suite as separate products, so even a headline figure would need splitting. We give no credit for a catalogue we have not measured. Four data centre regions, selectable from the first paid plan, are a genuine point in favour.

User sentiment 3.0/5

The widest disagreement between sources in our entire corpus, 3.2 points, and it is not noise. Where buyers review software, the product does well: 4.5 across 623 reviews on Capterra, 4.5 across 1,234 on G2, 8.5 out of 10 across 137 on TrustRadius. Where consumers review companies, it does badly: 1.3 across 120 reviews on Trustpilot. We measured the gap rather than averaging it away. The Trustpilot profile is for Freshworks, which also sells Freshdesk, Freshservice, Freshchat and Freshcaller; of the 100 reviews we read, 22 name Freshsales and 74 name no product at all. The 20 English-language reviews that name Freshsales and no sibling product average 2.9 — better than the 2.05 of the corpus as a whole, and still far below the software aggregators. Twenty reviews is a small sample and the mark reflects that uncertainty rather than resolving it.

Docs & learning resources 3.5/5

The plan documentation is the best we have read: a 136-row comparison matrix in which every row carries an explicit included-or-not flag for each of the four plans, which lets a buyer state what a plan lacks rather than infer it from a missing tick. Tooltips explain the governance features in the buyer's terms, and the help centre does publish the hard numbers the pricing pages omit: custom fields by type per plan, API limits per hour and per minute, free call minutes, and an article that states each plan's limits as the things you lose on downgrade. Three deductions. The matrix is served as collapsed accordions, so a reader sees two of twenty-four sections without clicking. The numbers live nowhere near the pricing page, so a buyer has to know they exist to look. And the documentation contradicts itself in three separate places we checked: how many custom fields a free account gets, whether the API limit is per account or per agent, and whether seats can be reduced mid-term — on that last one the help centre documents a procedure the master agreement forbids.

Support & reliability 2.2/5

The weakest axis, and the only one where the vendor's own structure and its reviewers point the same way. The support section of the 136-row matrix contains exactly one row — 24x5 support — included identically on all four plans. There is no premium tier, no named contact and no faster response time purchasable at any price on the published ladder: the free plan and the $59 plan get the same support. That matters because support is overwhelmingly what reviewers write about. It is the dominant theme in 60% of the English-language reviews we read, the highest single-theme share we have measured across any product, and it averages 1.9 stars across the company corpus and 2.6 across the Freshsales subset. A product whose most-discussed weakness cannot be remedied by spending more is being described accurately by its own pricing page.

Weighted across the axes above (weights reflect what actually drives the decision in this category). How we score →

## Small things that tell you how the product thinks

The currency depends on which page you open. Read from the United States, the pricing page quotes dollars and the comparison page quotes euros, for the same plans, at the same numbers^[2]. The vendor's own data confirms the euro and dollar figures are numerically identical on every plan. Whether that is a convenience or a nine-per-cent markup depends on the exchange rate on the day, and on which of the vendor's two pages you happened to land on.

Data residency costs you features. Four regions are offered from the first paid plan. But the vendor's own documentation on residency notes that enabling it disables call masking and verified caller ID. For a European team that needs residency and also needs the phone, that is a trade nobody mentions on the way in.

File storage is the one limit that moves in your favour. Two gigabytes per user on Growth, five on Pro, and a hundred on Enterprise^[2]. Twenty times the storage for one and a half times the price is the single most generous step anywhere in the ladder, which is worth knowing mainly because it is so unlike the rest of it.

The mobile app has a button to book an Uber. It is on the free plan. We mention it because it is the clearest signal of who this product was designed for: a field sales person moving between meetings, not an operations team at a desk.

The contract nobody shows you

Neither pricing page mentions the terms, and the terms are stricter than the monthly rate beside them suggests.

The master agreement states that payment obligations are non-cancelable and non-refundable regardless of whether you use the service^[3]. During a subscription term you may neither reduce your user count nor move to a cheaper plan^[3]. Plans renew automatically at the same seat count unless you give written notice thirty days before expiry. And the agreement reserves an explicit right to raise the fee at renewal, naming inflation and product development as reasons^[3].

There is a wrinkle worth knowing: the help centre publishes an article explaining how to reduce your number of users, which is a procedure the agreement forbids mid-term. Both documents are the vendor's. We cannot tell you which one governs your account, and that is the point — ask before you sign, in writing.

## Where it sits in the category

We have now scored five CRMs on the same eight-axis rubric, and Freshsales lands at the bottom of them. That needs explaining, because on capability it is not the weakest of the five.

It scores well on breadth: the phone, chat and chatbot included at the entry price are things its rivals charge extra for or do not offer at all. It scores well on plan documentation, better than anyone else we have read — a 136-row matrix with an explicit flag per plan is genuinely unusual, and it is the reason this review can tell you what each plan lacks rather than guessing from a missing tick.

What pulls it down is concentrated in two places. Support, at 2.2 out of five, is the lowest mark we have given any product on any axis, and it is low for a structural reason rather than an anecdotal one: the vendor sells exactly one level of support and it is the same on every plan. And value, despite the low entry price, suffers because the entry price does not buy the product most people think they are buying, and because the contract behind it is considerably less flexible than a monthly rate implies.

A CRM that is cheap to enter and expensive to actually use is a familiar shape. What makes this one unusual is how precisely the vendor's own documentation lets you see it.

What we could not verify

Three things we would normally check, and could not.

How many integrations it has. The Freshworks marketplace filters by product, but publishes no application count on any page we could read — and its own filter separates Freshsales from Freshsales Suite, so even a headline figure would need splitting between them. We score what we measure, so this product gets no credit for a catalogue size we cannot see.

What an extendable API rate limit extends to. The top plan advertises one. No page defines the term or names a ceiling, and the only documented way to raise a limit is to email support and ask — which any plan can do^[5].

Whether call time is rounded. The telephony pages give an overage rate per minute and a renewal charge for phone numbers without a figure, and say nothing about per-second billing. For a calling team that is a material gap, and it is the question we would put in writing first.

There is a fourth, subtler one. Two of the vendor's own articles disagree about whether the API rate limit is applied per account or per agent^[5]. On a ten-seat team those two readings differ by a factor of ten. We have recorded both rather than choosing, because choosing would mean inventing.

## Who should buy it

A small sales team whose work is phone calls, who want a CRM and a dialer on one bill, and who will be on Pro. Not Growth. If the budget is nine dollars a seat, buy something else or accept that you are buying an address book.

A team of three or fewer who want to try the shape of the product before committing can genuinely use the free plan, provided they understand it has no automation and no API and that the clock on migrating out starts the day they put data in.

Who should not: anyone whose data has relationships, because lookup and formula fields start at thirty-nine dollars a seat. Anyone who needs an escalation path on support, because there is not one. And anyone who needs the flexibility a monthly price implies, because the contract does not grant it.