Pricing and quotation guide
Internal AI assistant development cost
An internal AI assistant can be inexpensive to start if it is a general-purpose tool licensed per user with clear guidance, or a significant project if it connects to your documents and systems with permissions that must be designed and tested. Costs depend on the number of tasks, how many sources and systems are connected, how complex permissions are, whether a pilot and training are included, how licensing or usage fees scale with staff numbers, and how much support is needed afterwards. Data protection work is part of the cost either way.
This guide covers cost drivers, typical scopes, pricing approaches, recurring fees and comparing quotes. It gives no prices, because they depend on scope, staff numbers and suppliers; LetGetGrow sets out the cost for your project in a custom quotation.
Illustrative diagram
What drives internal assistant cost
- Tasks
- Each task may need its own sources or tools.
- Sources and systems
- Each connection needs setup and testing.
- Permissions
- Mirroring access rules is careful work.
- Pilot and training
- Running a pilot and training staff take time.
- Licensing
- Per-user or usage fees scale with your team.
Costs paid to third parties (separate from our fees)
- AI tool per-user licences
- AI usage fees
- Search or storage services
- Adviser fees for data protection
Ask for licensing costs projected for your full team, not just the pilot group.
Approaches compared
| Approach | Typical scope | Relative cost |
|---|---|---|
| General tool with guidance | Licensed AI tool, policy, training | Lower to set up; per-user fees |
| Connected to documents | Approved sources, permissions, citations | Moderate |
| Connected to systems | Documents plus tools such as ticketing or CRM | Higher |
How internal assistants are priced
Pilot then rollout
Suits you when
- Value and risk uncertain
Watch out for
- Rollout price agreed in advance
Fixed build plus licences
Suits you when
- Scope settled
Watch out for
- Licence growth with headcount
Managed service
Suits you when
- Ongoing changes expected
Watch out for
- What the monthly fee covers
Licensing and usage fees
Some AI tools charge per user per month; others charge by usage. Per-user fees are predictable but grow with headcount, while usage fees vary with how much staff use the assistant. Project both for your full team over a year before deciding.
Permission work
Designing and testing permissions is often the most important part of a connected assistant. It takes time to map who can see what and to test with accounts in different roles. Cutting this work to save money risks exposing restricted information.
Example: comparing approaches
Illustrative example — not a client case study
Situation: A professional services firm of around forty staff considers an internal assistant.
- A general AI tool with a usage policy is priced per user per month.
- A connected assistant drawing on the firm's procedures adds build and permission work.
- The firm pilots the general tool first, then adds connected procedures for one team.
What this shows: The firm learned how staff used AI before investing in connections. This example is illustrative.
Pilot and training
A pilot with one team and short training sessions add cost but reduce the risk of a poorly received rollout. Training should cover what the assistant is for, what not to enter and how to check answers.
Comparing quotes
Check tasks, sources, system connections, permission design and testing, pilot, policy and training, licensing projections, data protection support and ongoing support.
Keeping the cost down
Start with a few tasks and one team, connect read-only sources before systems, use existing permission structures and run training in-house after the first session.
Data protection work
ICO guidance says AI use usually requires a data protection impact assessment. Preparing it, checking supplier terms and writing staff guidance take time and possibly adviser fees.
Ongoing support
Sources change, staff join and leave and suppliers update their tools. Budget for someone to maintain sources and permissions, whether internally or through a provider.
Judging the value
Ask pilot users how much time they save and on which tasks, and check the quality of outputs. Weigh that against licences, build and support, and against the risk of staff using unmanaged public tools instead.
Connecting to systems
Letting the assistant read from or act in ticketing, CRM or scheduling systems adds development, permission design and testing for each connection. Starting with documents only, and adding systems once the assistant proves useful, spreads the cost and the risk.
Preparing sources
If the assistant draws on internal documents, those documents need checking for currency, contradictions and sensitive content before they are connected. Where documentation is messy, this preparation can be a meaningful part of the cost, and it improves the documents for everyone.
Questions about internal AI assistant development cost
What affects internal assistant cost?
Tasks, sources and systems, permissions, pilot, training and licensing.
Is a general AI tool cheaper?
Usually to set up, but per-user fees grow with headcount.
Why does permission work cost so much?
Mapping and testing access by role takes careful effort.
Is a pilot worth paying for?
It reduces the risk of a poor rollout.
What ongoing costs are there?
Licences or usage, source and permission upkeep and support.
How do we judge value?
Time saved by pilot users, output quality and risk reduced.
Request a Custom Quote
Ask about Internal AI assistant development cost
Thank you — your enquiry has been sent.
We have received your message and will reply by email. If you need to add anything, email info@letgetgrow.com.