How to choose a chatbot or AI developer: 10 questions, red flags, and proposal comparisons
When choosing a chatbot or AI developer, assess three things: whether they understand your process, can show comparable work, and separate development, recurring costs, and support in their proposal. An hourly rate says little on its own. Good questions about goals and exceptions help define the work before anyone commits to a price. This guide covers what to ask, what to watch for, and what to put in the contract.
Provider types: builder, freelancer, or agency
The right option depends on the complexity of the job. A no-code builder can suit a basic FAQ or lead form when its limits on logic and integrations are acceptable. A freelancer may be a good fit for a small bot with one integration, provided you agree on support and availability. An agency or in-house team may fit a process involving multiple channels, CRM, AI, and ongoing support; expect a higher rate or more coordination overhead.
Compare providers by their ability to deliver the scenarios, integrations, and post-launch responsibility you need, not by their label. Builder subscriptions and custom development costs also depend on scope and recurring service fees.
10 questions to ask a provider
Ask for written answers:
- What similar projects have you delivered? Request case studies with the problem and scope, not just screenshots.
- What do you need to learn about our process? Metrics, exceptions, and integrations should be discussed before estimating.
- What is included in the first release, and what is excluded?
- How are one-time and recurring costs split: development, hosting, AI API, services, and support?
- Who will work on the project, and who will be our point of contact?
- How do you test scenarios and evaluate AI quality?
- Who will own the source code, data, and accounts?
- What happens after launch: warranty, support, and response time?
- How do you handle integration failures: queues, retries, logs, and alerts?
- How do you protect data and access to AI tools?
Red flags in a commercial proposal
An exact price with no questions about your process may mean the provider priced a standard template or added a large uncertainty buffer. Other warning signs include calling something an “AI agent” without listing its actions; promising to replace most of a team without baseline data; omitting recurring costs; keeping code and accounts without a transfer plan; or skipping requirements discovery on an integration-heavy project.
Ask for detail behind case studies: the original problem, what was built, and the scope. A technology label is not proof of functionality. Gartner forecast that more than 40% of agentic AI projects would be canceled by the end of 2027. This is a forecast about the category overall, not a measured outcome for a particular provider or chatbot project. For terminology, see chatbot or AI agent.
How to compare proposals from different providers
Give every provider the same brief or requirements document first. Then compare how well they understand the task, relevant case studies, cost transparency, first-release boundaries, support, and ownership of code and data. Compare prices only after the scope is aligned.
| Criterion | What to check | Priority |
|---|---|---|
| Understanding the task | Clarifying questions tied to a business metric | High |
| Relevant work | Similar scenarios and integrations | High |
| Cost breakdown | Development, recurring fees, and support shown separately | High |
| First release | Included work and deferred items | Medium |
| Support | Warranty, terms, and post-warranty cost | Medium |
| Ownership | Repository and access handover | High |
| Price | Compared for the same scope | Medium |
BotLabs case studies for W8 Shipping and UA Made publish hours by workstream. Use this as an example of scope transparency, not as a prediction for your project: W8 Shipping and UA Made.
What to include in the contract
Record the first-release scope and boundaries, acceptance criteria and correction process, transfer of source code and access, ownership of third-party accounts, payment for AI APIs, warranty, and ongoing support. Include confidentiality and personal-data handling terms. Turn verbal promises into verifiable contract terms or an attached statement of work.
How BotLabs works
BotLabs starts project discussions by defining the first release: its primary scenario, metric, and minimum integrations. That provides a basis for estimating the work and showing development, recurring costs, and support separately. See the BotLabs portfolio for examples of the company’s work.
Conclusion
A capable provider asks questions before estimating, shares comparable case studies, and separates one-time from recurring expenses. Compare proposals using the same brief, check how code and accounts will be transferred, and put first-release boundaries, acceptance, and support in the contract.
Frequently asked questions
Review relevant case studies and ask about process, first-release scope, costs, testing, ownership, and support. Compare providers against the same brief.
A freelancer may suit a simple bot with one integration. A team may be a better fit for a process involving CRM, multiple channels, AI, and ongoing support.
An exact price with no process questions, an AI agent with no list of actions, missing recurring costs, and no clear handover plan for code and accounts.
The client or their company should own them. The contract should explain transfer of source code, bot accounts, server access, and third-party services.
For a project with integrations and multiple scenarios, discovery helps define the first release, risks, and scope before estimating.
Align the scope first, then compare one-time development, recurring fees, support, first-release boundaries, and transfer of rights and access.
Sources
Discuss chatbot development with BotLabs
In 15 minutes with a BotLabs implementation specialist, discuss your workflow, integrations, and support needs. Identify the inputs for a proposal that separates development, service costs, and ongoing support.
Scope first.
Then a proposal.
15 minutes with an implementation specialist: your workflow, integrations, and inputs for a BotLabs proposal.
