Inflectra.ai Frequently Asked Questions (FAQs)
Technical Capabilities
What AI models and technologies do you use?
- We use the Amazon Bedrock Generative AI platform as our core AI infrastructure
- For all regions where they are available, we use the Amazon Nova LLM models
- For regions where Amazon Nova is not available, we use Anthropic Claude Haiku
- We Amazon Bedrock Guardrails to monitor and ensure the safety and compliance of our AI infrastructure.
How do you train and fine-tune your models?
- We do not currently train or fine tune our own models.
Can your solution be customized to our specific use case?
- The user is not able to customize the models used or the system prompts. The user is able to determine what AI actions to take and what data to send to the service.
What is your model’s accuracy, and how is it measured?
- We are using Amazon’s Nova models and Amazon carries out extensive accuracy and benchmarking. You can read more about that, for example, here: https://aws.amazon.com/blogs/machine-learning/benchmarking-amazon-nova-and-gpt-4o-models-with-flotorch/.
How do you handle model drift and updates?
- This is not applicable as we do not own or create any of our own models.
Data Privacy & Security
How is customer data handled, stored, and protected?
- Customer data is encrypted at rest and in transit. Once it enters our cloud environment, the data does not leave the region of the hosted site. Information about AI interactions is stored for logging and debugging for a short period of time after which it is purged.
Are you compliant with regulations like GDPR, HIPAA, or CCPA?
- Yes. Inflectra.ai is compliant with GDPR, CCPA, HIPPA and the EU AI Act.
- Inflectra Corporation regularly follows patch maintenance, security operations, monitoring protocols, and infrastructural configuration enforcement standards under our certified SOC 2 regulatory requirements.
Do you use customer data to train your models?
- This is not applicable as we do not train any models.
What security certifications do you hold (e.g., ISO 27001, SOC 2)?
- Inflectra Corporation is SOC 2 Type 2 certified. Other Security information is available here.
Performance & Scalability
What is your system uptime and SLA (Service Level Agreement)?
- Here is our Hosting Terms of Service with a guarantee of 95% uptime or a refund. Our actual uptime is 99.999% excluding the normal monthly maintenance and 99.985% including scheduled maintenance.
How do you ensure scalability during peak usage?
- Inflectra.ai is 100% serverless, running on AWS first party services including Amazon Bedrock and Lambda. It is built by design to be horizontally scalable to meet variability in usage.
What is your latency for real-time applications?
- Our product is not considered a real-time system in the strictest definition. It is used for application lifecycle management, quality assurance and project management.
- In general, AI services will have latency due to the processing by the LLM. This can vary based on the model, the input data, and the nature of the operation. Our current AI functionality is asynchronous and may see latency of a few seconds. Latency may change over time. For example, agentic workflows can have significant latency because of the complex operations involved. Streaming response latency is very different from asynchronous responses.
Ethics & Bias
How do you mitigate bias in your AI models?
- This is not applicable as we do not own or create our own models. However, Amazon has the following policies and processes in place with respect to bias in its models: https://docs.aws.amazon.com/ai/responsible-ai/nova-micro-lite-pro/overview.html
Do you have an ethics board or review process?
- We have an AI Advisory Board. See our Responsible AI policy here.
Can you provide transparency or explainability in your AI decisions?
- It is the responsibility of the foundational model owner - in this case Amazon for its Nova model to provide details about its transparency and explainability. You can read the AI Service Cards for Amazon Nova models here: https://docs.aws.amazon.com/ai/responsible-ai/nova-micro-lite-pro/overview.html.
Support & Maintenance
What kind of support do you offer (e.g., 24/7, dedicated account manager)?
- Inflectra offers unlimited support via email and phone calls
- Typically tickets are answered within a couple of hours (normal support coverage is 24x5 and policy is that tickets will be answered no later than 1 business day)
- For cloud customers we are monitoring 24x7 for any outages
- Customer’s have a dedicated Account manager for escalations.
How often do you release updates or patches?
- Inflectra.ai receives regular updates and can be patched rapidly in the event of a critical bug. There is effectively zero disruption of service to customers during these updates. Information about all customer facing feature releases are included in our Inflectra.ai release notes: https://spiradoc.inflectra.com/About/release-notes-inflectra.ai-v1/
What is your incident response process?
- We have outlined that process here: https://inflectra.com/Support/KnowledgeBase/KB698.aspx
Integration & Compatibility
What APIs or SDKs are available?
- We do not currently provide any APIs or SDKs for Inflectra.ai. It is only accessible as a browser based service inside Spria
- Spira itself has a well-documented API that can be found here: https://spiradoc.inflectra.com/Developers/API-Overview/
Can your solution integrate with our existing systems (CRM, ERP, etc.)?
- Inflectra.ai has no third party integrations at this time, however the core Spira application integrates with over 80 different external systems: https://www.inflectra.com/Products/SpiraPlan/Integrations.aspx
What platforms or environments do you support (cloud, on-prem, hybrid)?
- Inflectra.ai is only supported with our cloud SaaS option.
- Our core Spira platform itself supports cloud, on-prem and hybrid.
Case Studies & References
Can you share case studies or success stories?
Architecture Diagrams
Core SaaS Application
The Spira cloud / SaaS platform hosted by Inflectra using Amazon Web Services (AWS) has the following core architecture diagrams:
The application architecture uses an n-tier web application delivered using the Microsoft IIS web server, with the transactional data resident in a single-tenant SQL Server database. These two components are hosted on Elastic Cloud Compute (EC2) virtual machines running in multiple different AWS availability zones and regions for redundancy and disaster recovery.
Inflectra.ai Architecture
The Inflectra.ai architecture uses a different stack to the core Spira platform. Instead, it uses a cloud-native, serverless architecture for maximum availability and scalability.
We are using Amazon Bedrock to host the Large Language Models (LLMs) being used to provide the Generative AI functionality, with Amazon Bedrock Guardrails and Prompt Management to ensure that common AI attack vectors are prevented:
In addition, we have a detailed monitoring and observability stack in place to ensure we can monitor the inputs and outputs to/from all of the AI models for quality control.
