Moonhub implementation cost: what AI hiring tools cost instead

In June 2025 a hiring startup announced it was joining Salesforce, and the story was quickly told as an acquisition. A year later, someone in your talent team asks what it would cost to "put Moonhub into our Salesforce org". The short answer is that it costs nothing, because there is no Moonhub product to put anywhere. The longer answer is more useful: what AI hiring tools cost for a business that runs on Salesforce, and what a sensible project looks like.

#What actually happened to Moonhub

TechCrunch carried Salesforce's own correction of the story. In its words, "A Salesforce spokesperson told TechCrunch that Moonhub has not, in fact, been acquired (by the company’s definition of the term). Rather, Moonhub is shutting down, and only some of the company’s team, not the whole team, as MoonHub implied in its post about the news, will be joining Salesforce."

That sentence settles the cost question for the name. Salesforce did not buy a product. It took on some people from a company that was closing. There is no Moonhub licence on any Salesforce price list, no Moonhub package from Salesforce on AppExchange, and no Moonhub feature inside Sales Cloud, Service Cloud or Agentforce that you can enable. A quote that prices a Moonhub implementation is pricing a product that does not exist, and you should send it back.

It is also worth separating two things that often get blurred. People who join Salesforce from a closing startup bring skills and ideas, and those may shape future Salesforce products. That is not the same as Salesforce selling the startup's product, and nothing you can buy today carries the Moonhub name or a published Moonhub price. Plan on what is on the price list now, not on what a team might build later.

What the talent team wanted from Moonhub is still a fair thing to want. Moonhub built AI tools for hiring, and the category it worked in is crowded. Some of those tools can be connected to Salesforce, and some of the job can be built on Salesforce directly.

#Salesforce does not sell a recruiting system

This is the fact most budgets miss. Salesforce publishes prices for HR service desks, employee agents and customer-facing agents, but it does not sell an applicant tracking system. Recruiting on Salesforce therefore takes one of three shapes:

1. Buy a recruiting platform that has its own AI features, and integrate it with Salesforce for the records you need there, such as hiring requests, contractor placements or onboarding cases.
2. Build a candidate-facing agent on Agentforce, usually on your careers site, to answer questions, collect details and book screens, with the recruiting system remaining the record of truth.
3. Extend HR Service in Salesforce to pick up the new starter once the offer is signed, so onboarding requests, equipment and access run through one case.

Most organisations end up with the first and third, and add the second only when candidate volume justifies it. Each shape has a different cost profile, and only some of the prices are public.

#What AI hiring tooling costs, where the price is published

The figures below come from each vendor's own pricing page, as at October 2026, in US dollars unless stated. Salesforce shows Australian dollar figures on its pages, given in brackets. Check the Agentforce pricing page, the HR Service pricing page and the Workable pricing page before you rely on any figure, because each can change.

Workable is in the table because it publishes its prices, not because it is the right choice for you. Its AI recruiting agent runs on credits: one credit to screen a candidate against your ideal profile, two to source a passive candidate, and ten each time a candidate chats with it. The page adds that "Every paid account starts with 3,000 free AI credits to try." Its subscription plans are priced by employee band, so the plan price depends on the size of your organisation rather than on the number of recruiters. Many other recruiting platforms ask you to contact sales, and for those the price is quote-only. Expect it to be set by employee count or hiring volume, the modules you take, the number of recruiter seats, and the integrations included.

On the Salesforce side, the useful detail is on the HR Service page: "Agentforce (flex credits) and Mule for Flow (workflows & connectors) are included in IT and HR Service products, and are unmetered." If you already run HR Service, a new-starter agent built inside it does not add a separate Flex Credit bill for employee use. A public careers-site agent is different. Candidates are not employees, so that agent is customer-facing, and Salesforce meters it per conversation or per action. The Agentforce page frames the second option this way: "Flex Credits are a new unit of payment for Agentforce that align cost to the business value your AI agents create."

#Working out the consumption bill yourself

Vendors will forecast consumption for you, but it pays to understand the arithmetic before you read their numbers, because every input is something you already know better than they do.

For a recruiting agent on credits, start with applications per year by role family. Multiply by the screening charge, since every inbound application is screened once. Add the passive candidates you expect the agent to source, at the sourcing charge, and the share of candidates who will chat with it, at the chat charge. The chat line is the one that surprises people, because at ten credits an interaction it costs ten times a screen, and candidates who are unsure about a role tend to ask several questions.

For a careers-site agent on Agentforce, the question is which meter suits your traffic. Per-conversation pricing is simple: each candidate conversation is one charge, however many questions it holds. Flex Credits charge per action, so a conversation where the agent looks up three roles and books a call uses several actions. Short, single-question conversations favour Flex Credits; long, multi-step ones favour per-conversation pricing. Ask Salesforce which buying model your contract allows, because pre-purchase, pay as you go and pre-commit arrangements are all described on the pricing page.

Then add the seasonal shape. Graduate programmes, retail peaks and annual intakes can concentrate a large share of a year's applications into a few weeks. A forecast that divides the annual total by twelve will understate the bill in the months that matter and invite an awkward conversation with finance.

#When building on Salesforce makes sense, and when it does not

Building recruiting on Salesforce is attractive when hiring is part of something you already run there. Staffing and contractor businesses often treat candidates, placements and clients as one book of business, and keeping them in one platform has real value. Organisations that already run HR Service, and want the hiring manager's request, the approval and the onboarding case in one place, also benefit.

It makes less sense when the recruiting team is happy with a dedicated platform and Salesforce is mainly a sales or service system. Rebuilding job boards, scheduling and offer management on Salesforce to replace a working recruiting system is a large custom project with a long maintenance tail. In that case, integrate the two at the points where records genuinely need to meet, and leave candidate management where it already works.

#What an implementation is made of

An AI hiring project touches more systems than most people expect, and the work packages reflect that.

Discovery. Map the hiring process from requisition to first day. Who raises the request, where it is approved, which system holds the candidate, which holds the employee, and where Salesforce fits. Many organisations find that Salesforce holds the hiring manager's request and the onboarding case, while the candidate record never touches it.

Platform choice. If you are buying a recruiting platform, run a short evaluation against your own roles. Ask each vendor what its AI does with a candidate's data, how it explains a screening recommendation, and whether you can switch the AI off for particular roles.

Integration. Connect the recruiting system to Salesforce through its API or a published connector, or through MuleSoft if you already run it. Typical flows are requisition to job, offer accepted to onboarding case, and placement to contact or account for staffing firms. Decide which system owns each field, and do not let both write it.

Candidate agent build. If you add an Agentforce agent to your careers site, define its subagents narrowly: answer questions about roles and process, collect contact details, book a screening call. Keep screening decisions out of its hands.

Onboarding. Configure HR Service so that an accepted offer opens the right case, assigns equipment and access tasks, and gives the new starter an agent to ask questions in their first weeks.

Testing, training and hypercare. Test with real but anonymised candidate journeys, train recruiters and hiring managers, and watch the first hiring cycle closely.

#The risks that are specific to hiring

Hiring is not like service. A wrong answer from a service agent costs a refund; a wrong screening recommendation can cost a candidate a job and the employer a complaint. Several costs follow from that, and they belong in the budget from the start.

Human decision points. Decide where a person must review AI output before it affects a candidate. Screening, rejection and shortlisting are the usual places. Each review point is ongoing recruiter time.

Explanation and audit. Keep a record of what the tool recommended, what the recruiter decided and why. Ask vendors how their logs work and how long they keep them.

Privacy and consent. Candidates share personal information. Tell them what is processed by AI, keep data only as long as you need it, and get legal advice for each country you hire in. Employment and privacy law differ between markets, and some places regulate automated screening specifically.

Fairness testing. Check outcomes across groups before and after go-live. If the vendor cannot help you do that, factor in the cost of doing it yourself.

None of these are reasons to avoid AI in hiring. They are reasons a fixed-price quote that covers only configuration is incomplete.

#Getting comparable quotes, and what gets left out

Give each bidder the same pack: annual hires by role family, the countries you hire in, your current recruiting system, whether you run HR Service, and which of the three shapes you want. Ask for a quote broken into the work packages above, plus a twelve-month forecast of AI credits or Agentforce conversations based on your candidate volumes.

The items most often missing are:

• Consumption forecasts. Credits for a recruiting agent and conversations for a careers-site agent rise with applications, so a busy graduate intake costs more than a quiet quarter.
• Integration upkeep. Recruiting vendors change their APIs. Someone must own the connector after go-live.
• Sandboxes. Testing candidate journeys needs a sandbox with realistic data. Salesforce lists sandboxes in HR Service Desk Unlimited Edition, not in Enterprise Edition.
• Legal review, in each jurisdiction, before go-live and again when the AI features change.
• Recruiter time for review points, which does not shrink to zero.
• Contract overlap while the old recruiting system runs alongside the new one.

#Where to go next

For the history, what happened to Moonhub sets out the 2025 announcement and Salesforce's correction. Moonhub and Salesforce covers what the team's arrival does and does not mean for Salesforce products. The licence question on its own is on Moonhub Salesforce licence cost. If an architecture diagram still shows Moonhub, integrating Moonhub with Salesforce explains why that box should come out, and should I migrate from Moonhub helps former Moonhub users plan a move. Before you sign with a firm, use the questions on hire a Moonhub Salesforce partner.