Question 1

What does a typical engagement look like?

Accepted Answer

We start with a 2-week discovery to map use cases, data and integrations, then move into a 6–12 week production build with weekly demos.

Question 2

Do you work fixed-price or time-and-materials?

Accepted Answer

Both. Pilots are typically fixed-scope; long-running platform work runs on milestone-based T&M.

Question 3

Is the BAI Agent Platform self-hosted?

Accepted Answer

Yes — we support fully managed, dedicated single-tenant, and customer-managed deployments.

Question 4

Which LLM providers are supported?

Accepted Answer

OpenAI, Anthropic, Google, Mistral, plus self-hosted open models via vLLM or TGI.

Question 5

How is the platform priced?

Accepted Answer

Tiered subscription with usage-based add-ons for inference and connectors. See the pricing page for details.

Question 6

Are there discounts for non-profits or research?

Accepted Answer

Yes — please contact us with details about your project.

Question 7

Where is data stored?

Accepted Answer

By default in the customer's chosen region. EU, US, India and Singapore are available; other regions on request.

Question 8

Do you support SSO and audit logging?

Accepted Answer

Yes — SAML, OIDC, SCIM, plus comprehensive audit trails exportable to your SIEM.

Question 9

How do you prevent hallucinations?

Accepted Answer

Retrieval grounding, structured outputs, eval-driven development, and human-in-the-loop for high-stakes flows.

Question 10

How do you measure quality?

Accepted Answer

We instrument every agent with offline evals and live quality metrics — accuracy, helpfulness, safety, latency and cost.

Question 11

What support is included?

Accepted Answer

Email and Slack for all customers; dedicated success engineering on Professional and Enterprise tiers.

Question 12

What's the on-call SLA?

Accepted Answer

24/7 for Enterprise (P1 < 30 min response); business hours for Professional.

Frequently asked questions

Services