Digital transformation is bought as a programme and paid for as a set of seats. The slide deck talks about outcomes. The invoice counts agents. That gap is where transformation budgets actually get spent, and it is wider than most procurement documents show.
This article puts published price ladders next to each other, then works out the annual bill for a 12-person transformation team so you can put a number against the deck. It ends with the part that decides whether the programme lands: which system becomes the record, what the others become, and the manual bridge that appears wherever two of them disagree.
The pricing model that shapes the whole thing
Every tool in this market sells per agent or per user. The licence is the product. What you do with the licence is your problem. That sounds simple, but it creates a specific distortion: a transformation programme that adds eight agents to handle the new workflow also adds eight full-price seats to every tool in the stack.
The tools in a transformation programme are not interchangeable. They do different jobs. One of them becomes the record of truth. The others become feeders, mirrors or dead ends. Getting that wrong is expensive, and the expense is not on the invoice.
Published monthly prices, annual billing. Asyntai is per workspace
The record of truth decides the programme
Asyntai
Conversations and decisions live here
CRM
System of record. Finance depends on it
Support Ticketing
Case history trusted by customers
The manual bridge
Wherever CRM and support ticketing disagree, someone manually reconciles them. That reconciliation work is part of the programme budget. It is not on the slide deck.
What Asyntai does instead
When a thread escalates, it routes the conversation straight to your team with the full transcript and context attached. Decisions made in Asyntai sync to the CRM, so the bridge covers less ground.
Where transformation work actually happens
A transformation programme touches at least four categories of tool, and each category has its own pricing logic. Understanding what each one does in a transformation tells you where the real costs land, because the invoice never shows the integration work.
| Tool category | What it does in transformation | Priced per |
|---|---|---|
| CRM | Customer record, pipeline, account history. Often the record of truth if sales is the trigger. | User |
| Service desk | Tickets, cases, internal requests. Usually the record of truth for IT-led transformation. | Agent |
| Internal comms | Slack, Teams, email. Becomes the shadow record for anything urgent. | User or flat |
| Analytics layer | Reporting, dashboards, CSAT. Tells you whether the transformation worked. | Per seat or flat |
| Website chat layer | Answers the questions that arrive before a case is opened, and routes what it cannot answer to a named person. | Message. Asyntai is $39, $139 or $449 a month, whatever the agent count |
Each of these categories is priced by a different unit, which is why a programme budget built from one of them rarely predicts the others. We make Asyntai, which is the AI chat layer on the website in front of this stack, priced per message rather than per agent, so read any mention of it as a disclosure.
The published price ladders
Here are the self-serve price ladders for four tools that appear in transformation programmes. Prices are per month, billed annually, in dollars. These are the figures you use in the planning software and the business case. They are the starting point, not the total.
| Vendor | Tier 1 | Tier 2 | Tier 3 | Tier 4 | Tier 5 | Tier 6 | Add-ons |
|---|---|---|---|---|---|---|---|
| Freshdesk | $19 | $55 | $89 | -- | -- | -- | $29+ |
| Zendesk | $19 | $55 | $115 | -- | -- | -- | $50+ |
| Salesforce | $0 | $25 | $100 | $195 | $395 | $550 | AI on Core+ |
| Zoho CRM | $0 | $14 | $23 | $40 | $52 | -- | -- |
| Asyntai | $0 | $39 | $139 | $449 | -- | -- | -- |
Asyntai plans are priced per workspace, not per agent, and each carries a set number of dashboard seats. The Free plan includes 50 messages and 1 seat. Starter at $39 includes 800 messages and 2 seats. Standard at $139 includes 15,000 messages and 3 seats. Pro at $449 includes 50,000 messages and 10 seats. Enterprise is a signed order form with a negotiated message allowance and seat count, which is the route for a programme team larger than ten.
What a 12-person team costs annually
The table above shows tiers. Here is what those tiers cost in practice for a team of 12, which is a realistic size for a departmental transformation programme. This is the number that goes into the platform business case, before the integration costs.
The record of truth problem
None of the numbers above tell you whether the transformation will actually work. That depends on a decision no vendor makes for you: which system holds the record, and what happens when the others disagree with it.
In any transformation programme that lasts longer than six months, you will have at least two systems each believing it is the source of truth. One holds the customer record. One holds the case history. One holds the internal note. They were not designed to agree, and they will not agree without deliberate effort.
The effort is a manual bridge. Someone reads from one system, writes into another, and tries to keep the versions consistent. In a 12-person team, that person costs about a quarter of their time. In a 30-person programme, it is often a dedicated role. The salary for that role does not appear in any software invoice, but it is a direct consequence of the software choices made upstream.
When two systems disagree on the status of a case, the customer experiences it as confusion. When the CRM says the account is active and the service desk says the contract has lapsed, the agent on the phone sees both and has to decide which one to believe. That decision is made under pressure, it is made differently every time, and it generates a record that will need reconciling later.
This is the part of digital transformation that does not appear in the planning software. The planning software tells you what the new system will do. It does not tell you what to do with the data that already exists in the old system, or who will do it, or how long it will take, or what happens when the two versions of the truth are both true according to the tools that hold them.
The integration tax
Every tool in the ladder above charges per seat or per agent. Every integration between those tools charges separately. The integration cost is where the transformation budget actually goes, and it rarely appears in the initial platform planning.
A typical transformation programme that touches Zendesk, Salesforce, and a custom database will budget for the Zendesk seats, the Salesforce seats, and the database licence. It will then discover that moving a case from Zendesk to Salesforce requires either a paid connector or a developer, and that the connector has its own pricing model that was not in the original plan.
Freshdesk publishes a connector price: $80 per 5,000 tasks. That is a reasonable figure. But a programme that moves 200 cases a day will exhaust that allowance in 25 days and then pay again. The programme that budgets $80 once and moves on will be surprised by the second invoice.
Salesforce's own integration layer can push data into other systems, but it is designed for Salesforce-to-Salesforce integration, and anything outside that perimeter requires either a paid integration tool or a developer who knows both systems. The developer costs more per hour than any of the software prices in this article, and they are not optional once you have data that needs to live in two places at once.
Zoho has its own integration ecosystem through Zoho Flow, and the pricing there is separate from the CRM pricing. A team that chooses Zoho CRM because the per-user price is low will discover that the full Zoho stack required to run a transformation programme has a different total cost than the CRM alone suggests.
What the AI layer actually costs
The AI add-on is where the per-agent pricing model shows its teeth. Every vendor charges for it separately, on top of the base licence, and the AI layer is where the volume sensitivity lives.
Freshdesk charges $29 per agent per month for Freddy AI Copilot. That is the AI layer that answers questions without a human in the loop. The additional AI agent, which handles sessions automatically, is priced separately at $0.49 per session after the first 500. For a team handling 1,000 sessions a month, the cost of the sessions above the included 500 is:
Zendesk charges $50 per agent per month for Copilot and another $50 for Workforce Engagement. Together, that is $100 per agent per month in add-ons. For 12 agents, that is $1,200 per month in add-ons alone, or $14,400 per year. The base licence at $115 per agent is already expensive, and the add-ons double the cost again.
Salesforce AI is available only on Core and above. That means you cannot use Salesforce AI on the Starter Suite or the Pro Suite. The entry point for Salesforce AI is $195 per user per month, and then the AI features add to that. The published pricing page does not break out the AI cost separately, which makes it hard to compare against vendors who price AI as a distinct line item.
Asyntai prices the AI layer per workspace, not per agent. The Standard plan at $139 per month includes 15,000 messages. The Pro plan at $449 includes 50,000 messages. There is no per-agent charge for the AI layer, and there is no add-on for the AI features themselves. The variables are the message count and the seat block. Asyntai Pro is $5,388 per year and carries 10 dashboard seats, so a 12-person programme takes the seat count to an Enterprise order form rather than paying twelve licence fees. There are no per-agent AI fees and no tier restrictions on AI features.
Building the transformation toolkit
The tools in a digital transformation programme do not operate in isolation. They form a tech stack, and the stack is only as reliable as its weakest integration. When you are planning which software to buy, you are also planning which integrations you will need to build and maintain.
A transformation toolkit for a mid-sized programme typically includes at minimum a CRM, a service desk, a knowledge base, and some form of analytics or reporting layer. Each of these has its own pricing, its own user model, and its own way of handling data. The cost of the individual tools is visible. The cost of making them work together is not.
Cloud-native tools make integration simpler in theory, because they expose APIs that other cloud tools can call. In practice, the API access may be limited to certain tiers, or may require a paid add-on, or may work differently than the documentation suggests. Before you commit to a cloud platform for your transformation, check which API features are available on your target tier and which are locked behind an enterprise agreement.
The transformation toolkit also needs to account for the tools your team already uses. If your organisation runs on Microsoft, the platform choice will be different than if it runs on Google. If you use a custom database, the integration requirements will be different than if you use a standard CRM. These factors do not appear in the vendor comparison table, but they determine which table row is actually cheapest for your situation.
Choosing the record before choosing the tools
The price ladder tells you what a seat costs. It does not tell you which seat should be the record. That decision is made by the business, and it has to be made before you finish the software selection, because the choice of record system determines which integrations you need, which data has to move, and which manual bridges you will build.
If the CRM is the record, then the service desk is a mirror. Every case in the service desk must reflect a customer record in the CRM software holding it, and when they disagree, the CRM wins. That means the service desk team needs access to the CRM, or a way to see the CRM record from inside the service desk, or a person whose job is to keep them in sync.
If the service desk is the record, then the CRM is fed from it. Customer data comes in through the service desk, and the CRM is updated from there. That means the service desk team is the source of truth for everything, and the CRM is downstream. That arrangement works well for IT-led transformation, where the service desk is already the record of every customer interaction.
If neither is the record, you are building a third system that holds the truth and feeds both. That is the most expensive option. It requires the most integration, the most manual bridges, and the most ongoing maintenance. It is also the most common option in large organisations, where different departments have different definitions of the customer and no single system can hold them all.
What to measure before you buy
The numbers in this article are published prices. They are the starting point, not the final answer. The final answer depends on what you actually need the tool to do, how many people will touch it, and how many cases you will run through it.
Before you choose a tier, measure three things. First, how many agents will actually use the tool. Not how many are in the team, but how many will log in more than once a week. A team of 12 might have 8 active users and 4 occasional users, and the occasional users might not need full agent seats.
Second, how many cases you handle per month. Not just tickets, but the full volume of interactions that will flow through the new system. If you are moving from email to a service desk, the email volume is a reasonable proxy. If you are adding a new channel, estimate from comparable teams in your organisation.
Third, how many integrations you need. Every integration is a connection that must be maintained, and every maintenance task is a cost that is not on the invoice. Count the systems that need to share data with the new tool, and add a line for each one in your budget. If you do not know yet, budget for two integrations and add 50 percent.
These three numbers determine which tier you need, whether you need the AI layer, and whether the total cost of ownership is different from the per-seat sticker price. They are not in any vendor's comparison table, but they are the numbers that will appear on your actual invoice.
The manual bridge that decides whether it lands
Every transformation programme generates at least one manual bridge. This is the thing nobody budgets for at the start, and everybody discovers six weeks in. It appears wherever two systems disagree, and it is maintained by a person who spends their day reconciling the difference between what the CRM says and what the service desk says.
The manual bridge is not a software problem. It is a governance problem. It exists because nobody decided, before the tools were chosen, which system would be the authoritative record. The decision was left until after go-live, when it became a person who gets paid to sit in the gap.
Preventing the manual bridge requires a decision that is made before the software selection, not after. The question is not which tool is best. The question is which tool holds the truth. The answer shapes every integration, every data migration, and every process that touches the new system.
If the CRM is the record, the service desk is a mirror. If the service desk is the record, the CRM is a feeder. If neither is the record, you are building a third thing to hold the truth, and that third thing will have its own cost, its own maintenance, and its own bridge to the other two.
The transformation programme that lands is the one where this question was answered before the first seat was purchased. The programme that struggles is the one where the question was never asked, and the answer emerged from the first argument about whose case record was correct, six weeks after go-live, when the manual bridge was already in place.
Prices and availability checked September 2026. Freshdesk pricing from its own pricing page. Zendesk pricing from its own pricing page. Salesforce pricing from its own pricing page. Zoho CRM pricing from its own pricing page. Asyntai pricing from asyntai.com/pricing.