Table of Contents
- What is Stack AI?
- Who owns Stack AI?
- How does Stack AI work?
- What features does Stack AI offer?
- Workflow builder
- Knowledge bases and integrations
- Interfaces: chat, forms, website chatbot, Slack, and API
- Security, governance, and compliance
- Deployment options
- How much does Stack AI cost in 2026?
- What counts as a Stack AI run?
- What does the Stack AI Free plan include?
- How do you budget for Stack AI Enterprise?
- What are Stack AI's pros and cons?
- Pros
- Cons
- What do users say about Stack AI?
- Who should use Stack AI?
- The stack: which tool does which job
- What are the best Stack AI alternatives?
- Stack AI vs SiteGPT: which one fits?
- How to choose between Stack AI and its alternatives
- Frequently asked questions
- Basics
- What is Stack AI?
- What is Stack AI used for?
- Can non-technical teams use Stack AI without writing code?
- Who owns Stack AI?
- Pricing
- How much does Stack AI cost?
- Is Stack AI free?
- What is a run in Stack AI?
- Does Stack AI have a plan between Free and Enterprise?
- Features and security
- Is Stack AI good for customer service chatbots?
- Can Stack AI combine several AI models and APIs in one workflow?
- Is Stack AI HIPAA compliant?
- Can you self-host Stack AI?
- Does Stack AI integrate with existing business tools?
- Alternatives
- What are the best Stack AI alternatives?
- What is the difference between Stack AI and Botpress?
- Is Stack AI or SiteGPT better for customer service?
- Sources
Created by
Do not index
Created time
Feb 7, 2026 06:14 AM
Stack AI is an enterprise platform for building, deploying, and governing AI agents with a no-code workflow builder, and it fits IT and operations teams in regulated industries that automate multi-step internal processes. It has a Free plan with 500 runs a month and one paid plan, Enterprise, sold by custom quote. Teams that only need a chatbot that answers customer questions from their own content can launch faster on a tool with a published price.
TL;DR
- Stack AI (the vendor writes StackAI) is a drag-and-drop builder for AI agents and workflows. Asana acquired it in May 2026, and it still runs as its own product and brand.
- Pricing has two plans. Free costs $0 for 500 runs a month, 2 projects, and 1 seat. Enterprise is a custom quote, and no paid price is published.
- Enterprise holds the controls most regulated buyers need: SSO, access control, on-premise and VPC deployment, and a HIPAA BAA. None of them are on Free.
- Stack AI holds 4.5 out of 5 on G2 from 39 reviews. Reviewers praise ease of use and integrations, and some report usage limits and a learning curve.
- Stack AI is a workflow layer. If the job is customer service answers from your own content, SiteGPT does that from $39 a month billed yearly. It has no workflow builder and no free plan.
Aspect | Verdict |
Best for | IT and operations teams in regulated industries that automate multi-step internal processes |
Free plan | Yes. 500 runs a month, 2 projects, 1 seat, community support on Discord |
Paid pricing | Enterprise only, by custom quote. No published price |
Biggest strength | A no-code workflow builder with 300+ integrations and on-premise deployment |
Biggest limitation | No self-serve paid plan. Production use starts with a sales conversation |
Compliance | SOC 2 Type II, HIPAA, GDPR, and ISO 27001 listed. BAA on Enterprise |
Owner | Asana, since May 2026 |
G2 rating | 4.5 out of 5 from 39 reviews |
What is Stack AI?
Stack AI is an enterprise platform for building, deploying, and governing AI agents. Its documentation says IT and operations teams use it to deploy internal applications that automate business processes.
The vendor spells the name StackAI. Most people search for it as Stack AI, and both names refer to the same product.
The use cases in the documentation range from a chatbot that retrieves information from Microsoft SharePoint to an automation that researches a company and drafts an investment memo. Stack AI says its customers include banks, hospitals, and defense contractors.
Stack AI sits on the workflow side of the AI tools market. A content-trained chatbot answers questions from your documents. A workflow platform runs a process with several steps, tools, and approvals. The difference between those two levels is covered in AI agent vs chatbot.
Who owns Stack AI?
Asana owns Stack AI. The company announced the acquisition on May 27, 2026, in a letter signed by its two co-founders.
The announcement makes four commitments to existing customers:
- Stack AI continues to operate as its own product and brand.
- Existing workflows keep running as they are.
- Workflows do not have to connect to Asana.
- On-premise and VPC deployments are not affected.
G2 now lists the product as StackAI by Asana.
How does Stack AI work?
Stack AI works in four steps:
- Start a project. Open a blank project or pick a template such as Chat with Knowledge Base.
- Build the workflow. Drag nodes onto a canvas and connect them. Nodes cover inputs, AI models, knowledge bases, tools, logic, and outputs.
- Pick an interface. Choose a form, a chat assistant, a website chatbot, a Slack app, or an API in the Export view.
- Publish and monitor. Publish the workflow, then track runs, tokens, and conversations in the Analytics and Manager views.
Answers from your own documents come from a Knowledge Base node. It handles indexing and retrieval, the pattern known as retrieval-augmented generation (RAG). The documentation says its default settings fit about 90% of use cases.
What features does Stack AI offer?
Stack AI offers five feature groups that matter for a buying decision: the workflow builder, knowledge bases and integrations, interfaces, governance, and deployment options.
Workflow builder
The Stack AI workflow builder is a 2D canvas with a sidebar of nodes, described in the platform overview. You connect nodes to define what the agent reads, which model it calls, which tools it can use, and what it returns.
The pricing page lists Python and JavaScript code nodes on both plans. The platform is also model-agnostic. The integration catalog includes Anthropic models and models hosted on AWS Bedrock, among others.
An Evaluator view tests an agent on a batch of inputs from a CSV file. A second model can grade each output or compare it with a reference answer.
This is where Stack AI is the stronger pick. A content-trained tool such as SiteGPT has no workflow canvas. If a process must read a document, check a record, ask a person to approve, and write the result to another system, Stack AI is built for that job. If the workflow is a customer conversation on chat or phone, the Voiceflow pricing and review covers a builder made for that.
Knowledge bases and integrations
Stack AI connects agents to more than 300 integrations, according to its integrations catalog. It also supports MCP servers when a team needs a custom tool. The catalog groups integrations by type.
Category | Examples named by Stack AI |
Enterprise systems | SharePoint, SAP, Workday, Salesforce |
Data and storage | AWS S3, AWS Athena, AWS RDS, Azure Blob Storage, Snowflake |
Project and work tools | Asana, Airtable, Atlassian (Jira and Confluence), Miro |
AI models | Anthropic, AWS Bedrock |
Knowledge base sources | Document upload, web scraping, Google Drive, Notion |
Integrations do not all support the same actions. Stack AI's own guidance is to check the operations listed on each integration page before you design a workflow around one.
Interfaces: chat, forms, website chatbot, Slack, and API
Stack AI offers six ways to put an agent in front of users, listed in its interface documentation:
Interface | What it does | Typical use |
Form | A web form that takes inputs and returns a result | Business users running a repeatable task |
Chat Assistant | A hosted chat page with message history | Internal knowledge assistants |
Website Chatbot | A chat icon embedded with HTML or React code | Visitors on a public website |
Slack App | A bot that team members message in Slack | Internal team tools |
API | A REST endpoint that runs a published workflow | Embedding the agent in your own product |
Batch Run | Runs the workflow over a dataset of inputs | Bulk document processing |
The documentation also mentions a Microsoft Teams integration, a voice interface, and deployment through WhatsApp or SMS.
Stack AI can serve customers directly. Its customer service page describes Penny, a public-facing assistant the YMCA Retirement Fund built to answer participant questions.
One plan limit matters here. The plan comparison lists API Access as an Enterprise feature, so the Free plan is for testing interfaces Stack AI hosts.
Security, governance, and compliance
Stack AI lists four certifications on its security page: SOC 2 Type II, HIPAA, GDPR, and ISO 27001. The same page lists these controls:
- AES-256 encryption at rest and TLS 1.3 in transit
- Customer-controlled data retention
- Data processing addendums that bar AI model providers from training on your data
- Custom SSO through SAML
- Vulnerability scans tracked to resolution
Governance features go further than login. The documentation describes role-based access control over who can edit knowledge bases or publish workflows, approval flows before an agent goes live, and version control of every change.
The plan matters again. The pricing page lists SSO, access control, and a BAA under Enterprise only. A healthcare team cannot put patient data through the Free plan under a BAA. How BAA terms differ across chatbot vendors is covered in the guide to HIPAA-compliant AI chatbots.
Deployment options
Stack AI offers four deployment models, from multi-tenant cloud to air-gapped on-premise. The on-premise documentation says the platform can run inside AWS, GCP, Azure, or your own servers, and can orchestrate local AI models.
On-premise and VPC deployment are Enterprise features. Enterprise customers also get forward-deployed engineers and weekly co-building sessions, according to the documentation.
How much does Stack AI cost in 2026?
Stack AI costs $0 a month on the Free plan and a custom amount on Enterprise. The Stack AI pricing page publishes no price for Enterprise and lists no plan between the two.
Plan | Price | Runs a month | Projects | Seats | Support |
Free | $0 a month | 500 | 2 | 1 | Community support on Discord |
Enterprise | Custom quote | Custom | Unlimited | Custom | Dedicated support and solution engineers |
What counts as a Stack AI run?
A Stack AI run is one execution of a workflow. The pricing page defines it as "the execution of a workflow instance."
The 500 free runs work out to about 16 runs a day across a 30-day month, or about 22 a day across 22 working days. That is enough to build and test a workflow. It is thin for a tool a whole team uses every day.
What does the Stack AI Free plan include?
The Stack AI Free plan includes the core builder. The plan comparison shows agents, flows, storage, vision, speech, code nodes, data loaders, knowledge bases, and the Slack bot on both plans.
The Free plan leaves out the features a production deployment usually needs.
Feature | Free | Enterprise |
Runs a month | 500 | Custom |
Team members | 1 | Custom |
Projects | 2 | Unlimited |
Agents, flows, knowledge bases, code nodes | Yes | Yes |
API access | No | Yes |
Analytics | No | Yes |
Custom domain | No | Yes |
Email, priority, and dedicated support | No | Yes |
SSO and access control | No | Yes |
On-premise and VPC deployment | No | Yes |
BAA | No | Yes |
How do you budget for Stack AI Enterprise?
You budget for Stack AI Enterprise by getting a quote, because the pricing page gives no number to start from. The plan card says runs and seats are both custom, so the quote depends on at least four inputs:
- How many workflow runs you expect each month
- How many people will build or use agents
- Whether you need cloud, VPC, or on-premise deployment
- How much help you want from Stack AI's engineers
Bring those four numbers to the first call. Ask what happens when you pass the run allowance, since the pricing page does not say.
How published prices compare across chatbot vendors is in the AI chatbot pricing comparison.
What are Stack AI's pros and cons?
Pros
- No-code workflow builder. Non-developers can connect models, knowledge bases, and tools on a canvas.
- Code when you need it. Python and JavaScript nodes are listed on both plans.
- 300+ integrations. The catalog covers enterprise systems, data stores, and AI model providers, plus MCP servers.
- Several interfaces from one workflow. The same agent can ship as a form, a chat page, a website chatbot, a Slack app, or an API.
- Deployment choice. Multi-tenant cloud, VPC, and on-premise are all offered, including air-gapped setups.
- Governance built in. Role-based access, approval flows, version control, and audit logs are part of the platform.
- A free plan to test on. 500 runs a month cost nothing.
- Hands-on help. Enterprise customers work with forward-deployed engineers.
Cons
- No published paid price. You cannot estimate cost without a sales conversation.
- No plan between Free and Enterprise. A small team that outgrows 500 runs has one option.
- The Free plan is a sandbox. It has 1 seat and no API access, analytics, custom domain, or SSO.
- Compliance controls sit on Enterprise. SSO, access control, and the BAA are not on Free.
- A learning curve for advanced work. G2 reviewers mention it 6 times.
- Usage limits. G2 reviewers mention them 7 times.
- You build the experience yourself. A customer service bot on Stack AI is a workflow someone has to design, test, and maintain.
What do users say about Stack AI?
Stack AI holds 4.5 out of 5 on G2 from 39 reviews. 30 of the 39 are five-star reviews.
G2 tags the themes that reviewers mention most often.
Theme on G2 | Mentions | Type |
Ease of use | 24 | Praise |
Integrations | 10 | Praise |
Ease of creation | 9 | Praise |
Usage limitations | 7 | Complaint |
Learning curve | 6 | Complaint |
Limited features | 3 | Complaint |
G2's summary of the reviews says users praise the drag-and-drop builder for letting non-coders build complex workflows quickly. The same summary notes a learning curve for more advanced features.
39 reviews is a small base. Read the individual reviews from companies of your size before you rely on the average.
Who should use Stack AI?
Stack AI is a good fit if you:
- Automate internal processes with several steps, systems, and approvals
- Work in a regulated industry and need on-premise or VPC deployment
- Have an IT or operations team that will own the agents
- Want one platform for many use cases across departments
- Can buy through a sales process and a custom contract
Stack AI is a poor fit if you:
- Need a published monthly price before you commit
- Want a website chatbot answering from your help content this week
- Have no one to design and maintain workflows
- Need more than 500 runs a month but are too small for an enterprise contract
The stack: which tool does which job
Stack AI and a customer service chatbot do different jobs, and many companies run both. The table shows where each tool sits.
Layer | Job | Tool |
Workflow and agent orchestration | Run multi-step processes across business systems, with human review where it matters | Stack AI |
Customer-facing answers | Answer customer questions from your website, help center, and files, and escalate to a person | |
Human queue | Tickets and live agents for the conversations the AI hands over | Your helpdesk, such as Zendesk |
Systems of record | Hold the customer, finance, and document data the workflows act on | Salesforce, SAP, Workday, SharePoint |
SiteGPT is the knowledge-trained answer layer in this stack. It trains on 12+ content sources, including websites, files, YouTube, Google Drive, Notion, and help centers. Human escalation is on every plan.
Its limits are plain. SiteGPT has no workflow builder, no on-premise deployment, and no free plan. It does not run back-office processes. It connects to Zendesk on the Scale plan and above, and it has no chat integrations on Starter.
Stack AI can also serve as the customer-facing layer through its Website Chatbot interface. The stack makes sense when the service team wants answers live without building a workflow, or when the Stack AI contract covers internal use only.
What are the best Stack AI alternatives?
The best Stack AI alternative depends on the job you need done. No tool here is best overall. Each row names the case it fits.
The ranking uses three criteria: how close the tool is to Stack AI's main job of workflow automation, whether it publishes a price, and how much building it asks of you.
# | Tool | Best for | Starting price | Free option | Main trade-off |
1 | Technical teams that want workflow automation with published plans | Published per-execution plans, listed in euros. The entry plan includes 2,500 executions a month | Self-hosted Community Edition on GitHub. Cloud trial with no card | Aimed at technical teams. SSO starts on the Business plan | |
2 | The knowledge-trained customer service answer layer | $59 a month, or $39 a month billed yearly | 7-day trial, card required | No workflow builder and no free plan | |
3 | Custom conversation flows with code on messaging channels | $150 a month billed annually | Free plan, 100 conversations | Extra conversations cost $65 per 100 on Plus | |
4 | Small teams that want a content-trained agent with a free tier | $40 a month on monthly billing | Free plan, 50 message credits a month | Removing branding costs $99 a month |
n8n comes first because it is the closest match for the reason most people look at Stack AI: automating multi-step work across systems. SiteGPT is second because it solves a different job, and it only fits if that job is yours. If you want a similar no-code agent platform with WhatsApp, Telegram, and LINE channels and usage billed in credits, read the GPTBots review.
More builders with visual editors are compared in the guide to no-code chatbot builders. Larger customer service platforms are covered in enterprise AI chatbot solutions.
Stack AI vs SiteGPT: which one fits?
Stack AI and SiteGPT solve different problems. Stack AI builds agents and workflows for internal processes. SiteGPT is a chatbot trained on your own content for customer service and lead generation.
Criterion | Stack AI | SiteGPT |
Core job | Build and govern AI workflows | Answer questions from your content |
Workflow builder | Yes, visual canvas plus code nodes | No |
Setup | Design, test, and publish a workflow | Add content sources and embed the widget |
Content sources | Document upload, web scraping, Google Drive, Notion, and 300+ integrations | 12+ sources: websites, sitemaps, files, YouTube, GitHub, Google Drive, Notion, Dropbox, OneDrive, SharePoint, Box, help centers |
Deployment | Cloud, VPC, or on-premise | Cloud only |
Free plan | Yes, 500 runs a month | No. 7-day trial with a card |
Paid pricing | Custom quote | $59, $129, or $429 a month on monthly billing. $39, $79, or $259 a month billed yearly |
Billing unit | Workflow runs | Messages |
Human handoff | Human review steps you design into a workflow | Escalation to a person on all plans |
HIPAA BAA | Enterprise | Enterprise only |
Choose Stack AI when the AI must act inside your systems, follow an approval path, or run inside your own infrastructure.
Choose SiteGPT when the job is answering customers from content you already have. Annual billing lowers each plan's yearly cost:
SiteGPT plan | 12 months on monthly billing | 12 months on yearly billing | Saving | Messages a month |
Starter | $708 | $468 | $240 (34%) | 4,000 GPT-4.1-mini messages |
Growth | $1,548 | $948 | $600 (39%) | 10,000 GPT-4.1-mini messages |
Scale | $5,148 | $3,108 | $2,040 (40%) | 40,000 GPT-4.1-mini messages |
The two tools count usage in different units. Stack AI counts workflow runs and SiteGPT counts messages. Estimate your own volume in both units before you compare.
A HIPAA BAA and a signed DPA are available on the SiteGPT Enterprise plan only, which has custom pricing. The SiteGPT HIPAA page has the program details.
How to choose between Stack AI and its alternatives
If you need... | Pick... | Why |
Multi-step internal workflows with approvals | Stack AI | The builder, integrations, and human review steps are made for it |
On-premise or air-gapped deployment with vendor help | Stack AI | Enterprise includes on-premise deployment and solution engineers |
Workflow automation with a published price | n8n | Plans and execution limits are listed on its pricing page |
Workflow automation you host and run yourself for free | n8n | A self-hosted Community Edition is on GitHub |
Customer service answers from your own content, live this week | SiteGPT | It trains on existing content with no workflow to build |
A flat, published bill for a website chatbot | SiteGPT | Plans have fixed prices and message limits |
Scripted conversation flows on WhatsApp or other messaging channels | Botpress | Its Studio combines visual flows with code |
A free tier for a content-trained agent | Chatbase | The free plan includes 50 message credits a month |
Frequently asked questions
Basics
What is Stack AI?
Stack AI is an enterprise platform for building, deploying, and governing AI agents. It combines a no-code workflow builder, knowledge bases, 300+ integrations, and interfaces such as chat, forms, and an API.
What is Stack AI used for?
Stack AI is used to automate business processes with AI agents. The vendor lists use cases in legal, risk and compliance, finance, customer service, IT service management, and business operations.
Can non-technical teams use Stack AI without writing code?
Yes. Workflows are built by dragging and connecting nodes, and the Knowledge Base node handles document retrieval with default settings. G2 reviewers mention ease of use 24 times and a learning curve for advanced features 6 times.
Who owns Stack AI?
Asana owns Stack AI. The acquisition was announced on May 27, 2026, and Stack AI continues as its own product and brand.
Pricing
How much does Stack AI cost?
Stack AI has a Free plan at $0 a month and an Enterprise plan with a custom quote. The pricing page publishes no price for Enterprise.
Is Stack AI free?
Stack AI has a free plan with 500 runs a month, 2 projects, and 1 seat. It does not include API access, analytics, a custom domain, SSO, or a BAA.
What is a run in Stack AI?
A run is one execution of a workflow. Each time a published workflow processes a request from start to finish, it uses one run.
Does Stack AI have a plan between Free and Enterprise?
No. The pricing page lists two plans, Free and Enterprise.
Features and security
Is Stack AI good for customer service chatbots?
Stack AI can run a customer service chatbot through its Website Chatbot interface, and the YMCA Retirement Fund uses it for a public assistant. You build that bot as a workflow. A content-trained tool such as SiteGPT is faster to launch when answering from existing content is the whole job.
Can Stack AI combine several AI models and APIs in one workflow?
Yes. A workflow can call different models at different steps, use tools and integrations, and run Python or JavaScript code. The API interface lets your own application trigger a published workflow.
Is Stack AI HIPAA compliant?
Stack AI lists HIPAA among its certifications, next to SOC 2 Type II, GDPR, and ISO 27001. The plan comparison lists a BAA on Enterprise only, so covered healthcare work needs the Enterprise plan.
Can you self-host Stack AI?
Yes, on Enterprise. Stack AI offers on-premise deployment in AWS, GCP, Azure, or your own servers, and a VPC option.
Does Stack AI integrate with existing business tools?
Yes. Stack AI lists more than 300 integrations, including SharePoint, SAP, Workday, Salesforce, Snowflake, and Asana. Supported actions differ by integration.
Alternatives
What are the best Stack AI alternatives?
n8n fits technical teams that want workflow automation with published plans. SiteGPT fits customer service answers from your own content. Botpress fits custom conversation flows with code. Chatbase fits small teams that want a free tier.
What is the difference between Stack AI and Botpress?
Stack AI is built for internal workflows and enterprise governance, and it sells its paid plan by custom quote. Botpress is built for conversational agents on chat channels and publishes per-conversation prices from $150 a month billed annually.
Is Stack AI or SiteGPT better for customer service?
Stack AI is better when service work involves multi-step processes inside your systems. SiteGPT is better when most questions can be answered from your website and help center and you want a published monthly price.
Sources
All pages opened on 5 October 2026.
Written by
