Gartner Agentic AI Platforms Magic Quadrant: Where to Look

TL;DR

• Gartner publishes no single “Agentic AI Platforms Magic Quadrant”; agentic capability is scored across several overlapping 2026 markets.

• The closest literal match is Gartner naming Google the “Company to Beat” in its Enterprise Agentic AI Platforms vendor race.

• Gartner’s 2026 Hype Cycle for Agentic AI places agentic AI at the Peak of Inflated Expectations, and Gartner predicts over 40 percent of agentic AI projects will be canceled by the end of 2027.

• Gartner names Leaders separately for building AI applications, coding agents, conversational AI, and AI governance.

• In the related 2026 Magic Quadrant for AI-Augmented Low-Code Application Platforms, Mendix and OutSystems have each announced a tenth straight year as a Leader.

If you arrived searching for one chart that ranks agentic AI platforms, the short answer is that it does not exist yet, and the longer answer is more useful. Gartner spreads agentic evaluation across markets because agents now appear inside coding tools, conversational systems, data science platforms, and governance consoles.

A platform strong in one market can be absent from another. Reading only one report would miss most of the agentic landscape.

Here, a CIO scoping an agentic program first gets the evaluation landscape, then the adjacent low-code quadrant, side by side.

Where Gartner evaluates agentic AI

Gartner evaluates agents in several related markets. Each one below gives its edition, the scope it covers, and the Leaders Gartner names, with a link to the source.

• AI Application Development Platforms (Magic Quadrant, November 2025, updated April 2026): platforms for designing, building, testing, deploying, and governing applications that embed AI agents and assistants. The November 2025 edition named AWS, Google, IBM, and Microsoft as Leaders.

• Enterprise AI Coding Agents (Magic Quadrant, May 2026): agents that turn a developer’s intent into multistep plans, then run and check those steps across code and tests. Anthropic, Cursor, GitHub, and OpenAI are the Leaders, with AWS and Google among the Challengers.

• Conversational AI Platforms (Magic Quadrant, July 2026): platforms for building AI assistants and conversational agents. The 2026 edition names Google, Salesforce, Kore.ai, and SoundHound AI as Leaders.

• AI Governance Platforms (Magic Quadrant, June 2026, first edition): platforms that set, approve, and enforce responsible AI policies across applications and agents. IBM, ServiceNow, and Truyo are the Leaders, with OneTrust a Visionary and SAP also evaluated.

• AI Platforms for Data Science and Machine Learning (Magic Quadrant, June 2026): platforms for building and deploying models, now with native AI agents. Of its 18 vendors, Databricks, IBM, and Dataiku have announced Leader placements.

• Hype Cycle for Agentic AI (2026, Gartner’s first): places agentic AI at the Peak of Inflated Expectations. Gartner reports that only 17 percent of organizations have deployed AI agents, and more than 60 percent expect to within two years.

Gartner’s AI Vendor Race research, announced in December 2025, comes closest to the literal query: it calls Google the company to beat for enterprise agentic AI platforms, citing its integrated agent technology stack. No single report carries the title buyers search for.

What Gartner’s cancellation forecast tells buyers

Gartner predicts that over 40 percent of agentic AI projects will be canceled by the end of 2027, citing escalating costs, unclear business value, or inadequate risk controls. The last of those three causes is the one to plan for first. Agents that act on real approvals, payments, or records need the same audit trail, access control, and change history as any production system. When those controls are retrofitted after a pilot works, the pilot can fail its first compliance review.

That is why governance belongs on the checklist in every agentic market above. An agent that can execute is easy to demonstrate. An agent whose every action is attributed, reversible, and inspectable is what reaches production.

The governance gap Gartner’s agentic criteria expose

Read across those markets and one question stays with the buyer: who builds the governed workflows around the agents, under IT control, at the edges of the systems of record where the real data lives.

The Leaders sit in neighboring markets. AWS, Google, IBM, and Microsoft lead in AI application development; IBM, ServiceNow, and Truyo lead in AI governance; coding and conversational specialists lead their own reports. Gartner scored each one on its own market’s terms. Ask every vendor on your shortlist who can build the governed workflows at the edges of SAP, Oracle, Salesforce, and Workday, the structured work the core systems were never shaped to handle.

That is the layer Kissflow is built for. Kissflow AI Builder, generally available as of May 2026, takes a single prompt or an uploaded requirements document and generates a complete working app, refined through conversation. Its output is blueprints: structured metadata describing business logic, not code. Because blueprints are deterministic and inspectable, every artifact is a standard Kissflow component a developer can open and read, and a compliance lead sees exactly what the AI built. Governance is built into the platform. AI changes use the same audit log as manual edits, governed by role-based access control and SSO.

Said plainly: this is AI-assisted development with the human in the lead, and humans own the blueprints. Today, Kissflow AI generates apps, forms, workflows, and integrations from a prompt for people to review, while learning from outcomes and self-correction are still in active development. And Kissflow stays outside every system of record, serving as the governed execution and exception layer that surrounds them.

Kissflow is included in the Gartner Hype Cycle for Enterprise Process Automation, 2025, and was named an Established Vendor and Customers’ Choice in Gartner Peer Insights 2024, a recognition based on customer reviews. Kissflow holds SOC 1 Type II, SOC 2 Type II, SOC 3, ISO/IEC 27001, HIPAA, GDPR, and CCPA certifications.

For how IT teams pair generative AI with low-code, see GenAI and low-code for IT productivity.

Which workflows this layer is built for

Start with the approval or exception process that gates your agentic program:

• Procurement and vendor management approvals

• Onboarding and exception routing that spans HR, IT, and finance teams

• IT service requests, alongside the ITSM tool you already run

• Finance approvals and financial consoles

• Compliance audits with audit-ready evidence on demand

The related market: Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms

A CIO scoping agents also has to compare the platforms that build the surrounding business applications, which Gartner positions in the Magic Quadrant for Enterprise Low-Code Application Platforms (LCAP). On 28 September 2026 Gartner retitled that report the Magic Quadrant for AI-Augmented Low-Code Application Platforms. Gartner’s summary describes these platforms as tools that “help developers and business users build enterprise apps faster through AI generation, agentic workflows, and governance.” For a critical read of that report, see what Gartner’s low-code quadrant missed.

What Gartner requires of an enterprise low-code platform

Gartner publishes the full 2026 inclusion criteria only inside the report. Its public descriptions of earlier editions set a minimum:

• A low-code way to build a complete application, including user interfaces, business logic, workflow, and data services

• A model-driven or graphical programming approach, with scripting

The 2026 edition adds AI generation, agentic workflows, and governance to how Gartner describes the market.

How the quadrant plots vendors

The Magic Quadrant plots vendors across four categories on Ability to Execute and Completeness of Vision.

• Leaders excel in both vision and execution, with scalable products and the financial viability to invest in continued development.

• Challengers execute well with a stable customer base, but their roadmap is more incremental than forward-looking.

• Visionaries anticipate emerging needs and innovate, but their ability to execute may trail the Leaders.

• Niche Players focus successfully on a specific segment or product set and deliver real value within it.

For scaling programs, this guide on composable enterprise applications covers architecture and governance for larger low-code rollouts.

The 13 vendors in Gartner’s 2026 AI-augmented low-code quadrant

Gartner’s 2026 Magic Quadrant for AI-Augmented Low-Code Application Platforms evaluates 13 vendors: Appian, Creatio, Globant (GeneXus), Huawei, Microsoft, Oracle, OutSystems, Pega, Retool, Salesforce, ServiceNow, Siemens (Mendix), and Zoho. Gartner’s public summary lists the vendors but not their positions, so an entry states a 2026 position only where the vendor has announced one.

Mendix

• Position: Leader in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms, its tenth consecutive year as a Leader. The Mendix Platform is a Siemens business where domain experts and developers both work on the visual models its AI produces.

• Strengths: connects to ERP, CRM, operational technology, financial, and supply chain systems without replacing them; records every agent decision, approval, and data lineage event in a traceable log.

• AI and governance: Maia produces visual models instead of raw code; admins enforce policies centrally across every app, agent, and workflow.

• Limitations: the license price leaves out compute, so every deployment option on the Standard and Premium plans costs extra; high availability, horizontal scaling, and multi-region failover come only with the quote-based Premium plan; Standard support runs 9am to 5pm.

• Pricing: per app, monthly: the Standard plan for one app starts at €918 per month, and Premium is by quote. Mendix’s site shows euros only.

Microsoft

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Microsoft has not announced its 2026 position. Power Apps sits in Microsoft Power Platform with Power Automate, Power Pages, Dataverse, and Copilot Studio.

• Strengths: runs inside Microsoft Teams; stores app data in Microsoft Dataverse.

• AI and governance: brings Microsoft 365 Copilot and AI agents into custom apps; admins set access, auditing, data loss prevention, and environment policies centrally.

• Limitations: agentic features work only in model-driven apps; Dataverse database capacity beyond the included allowance is a separate add-on.

• Pricing: per user, paid yearly: Power Apps Premium is US20.00peruserpermonth,orUS12.00 per user per month with a 2,000-seat minimum; the Developer Plan is free for building and testing.

OutSystems

• Position: Leader in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms, its tenth consecutive placement as a Leader. The OutSystems Agentic Systems Platform runs in the OutSystems managed cloud or self-hosted on-premises.

• Strengths: more than 400 prebuilt connectors, plus MCP support to call any API.

• AI and governance: Mentor turns business requirements or prompts into application components, and Agent Workbench coordinates agents and apps; role-based access, audit trails, and governance policies apply to each app and agent.

• Limitations: the standard subscription includes ISO 27001 only, with SOC 2 Type II and PCI DSS sold as add-ons; 24x7 support and the 99.95 percent uptime SLA are paid upgrades over 8x5 support and 99.5 percent; self-hosting OutSystems Developer Cloud is a paid add-on.

• Pricing: by custom quote; the subscription starts with 100 internal users, and extra users, runtimes, and add-ons raise the price; the Personal Edition is free for testing apps, not for production.

Zoho

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Zoho has not announced its 2026 position. Zoho Creator is its low-code platform for business users and developers.

• Strengths: more than 1,000 prebuilt connectors, including SAP and Oracle; Deluge scripting and VS Code for custom logic; ISO/IEC 27001 certification and SOC 2 compliance.

• AI and governance: Zia agents turn a description into a requirements document, build the app, and generate test cases; agents and automations run inside the access controls an admin sets.

• Limitations: licensed per user; customer and vendor portals are a paid add-on sold in blocks of portal users; the Free edition covers one custom app.

• Pricing: per user per month, billed annually, on Standard, Professional, and Enterprise plans, plus a custom Flex plan; prices vary by region; a 15-day free trial and a Free edition are available.

Retool

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Retool has not announced its 2026 position. Retool is for teams that build internal tools on their own databases and APIs.

• Strengths: native mobile apps for iOS and Android; connects to any database, API, or LLM; unlimited web and mobile apps on every plan.

• AI and governance: an AI app builder generates and edits apps from chat prompts, and Retool Agents run tasks with LLMs; audit logs start at the Business plan.

• Limitations: SAML or OpenID Connect single sign-on and source control come only with the Enterprise plan; AI agents are billed by the hour on top of the plan.

• Pricing: per builder and per internal user, billed annually: Team is US10perbuilderandUS5 per internal user per month, Business is US50perbuilderandUS15 per internal user per month, and Enterprise is by quote; the Free plan covers up to five users.

Oracle

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Oracle has not announced its 2026 position. Oracle APEX is the low-code platform built into the Oracle Database.

• Strengths: no caps on the number of apps, developers, or users; inherits the database’s security, with access control down to rows and columns.

• AI and governance: the APEX AI Application Generator builds apps from declarative metadata, and APEXlang gives a definition people can read, review, and version; activity auditing is built in.

• Limitations: runs only on the Oracle Database; support depends on an active Oracle Database support contract; each release is supported for about 18 months, so upgrades are regular.

• Pricing: no separate license fee; APEX is included with every supported edition of the Oracle Database, so the cost is the database or Oracle Cloud service underneath it.

Pegasystems

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Pega has not announced its 2026 position. Pega Platform orchestrates agents, people, and systems across enterprise workflows.

• Strengths: one set of rules and processes governs AI agents alongside people; connects existing AI, data, and cloud systems without replacing core systems.

• AI and governance: Pega Blueprint turns an idea into an application design; Pega describes its orchestration as fully auditable.

• Limitations: Pega aims the platform at large enterprises and the processes they cannot afford to get wrong; its site publishes no prices.

• Pricing: sold by quote; a free Community Edition and a Platform Trial are available for evaluation.

Globant / GeneXus

• Position: included, as Globant, in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Globant has not announced its 2026 position. GeneXus, a Globant company, generates web, mobile, and cloud applications from a single model.

• Strengths: one technology-agnostic model targets several platforms; runs on-premises, in one cloud or several, or in a hybrid setup.

• AI and governance: combines generative and deterministic AI, with agents that build on the project knowledge base; GeneXus describes control over data, compliance, privacy, and operations.

• Limitations: Pro and Enterprise prices are not published; the pricing page refers to token-based usage but does not show how tokens are counted.

• Pricing: Free, Pro, and Enterprise plans; Free is for building a first product, and Pro and Enterprise are by quote.

Huawei

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Huawei has not announced its 2026 position. Huawei Cloud Astro Zero offers no-code and low-code development on Huawei Cloud.

• Strengths: separate no-code and low-code tracks in one service; every edition lets teams try an app in a sandbox before release.

• AI and governance: an assistant component can call foundation models through connectors; each role carries its own page and data permissions.

• Limitations: a low-code runtime, IoT access, and mobile app release come only with the Professional edition; the Professional edition’s low-code runtime is shared among tenants; the Free edition is capped at ten users; the debugging sandbox is an extra purchase on Professional.

• Pricing: billed by edition (Free, Standard, Professional) and by user, on yearly or monthly terms; Standard is US36permonthandProfessionalUS159 per month, each for up to 30 users; extra users cost US1.2peruserpermonthonStandardandUS5.3 per user per month on Professional.

Appian

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Appian has not announced its 2026 position in this report. The Appian Platform combines process orchestration, a data fabric, and AI agents.

• Strengths: a data fabric joins records from separate systems without migrating them; security rules for individual rows and fields are set once and applied to every app, API, report, and agent.

• AI and governance: Agent Studio places AI agents inside business processes, where their work is orchestrated and auditable.

• Limitations: AI agents come with the advanced and premium capability tiers, and usage limits may apply; self-managed customers need Appian on Kubernetes, version 25.3 or later, for AI features.

• Pricing: by quote, priced per user, per month, per app across Standard, Advanced, and Premium.

Creatio

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Creatio has not announced its 2026 position. Creatio pairs ready-made CRM apps with a no-code studio for building AI agents.

• Strengths: works with any large language model; connects agents to enterprise systems over MCP, APIs, and webhooks.

• AI and governance: Creatio AI Studio builds, tests, and publishes agents through prompts and visual designers, with policies, approvals, and an audit trail across each agent’s life cycle.

• Limitations: its packaged applications are CRM modules for marketing, sales, and service; on-site deployments are self-managed in the customer’s own data center.

• Pricing: US40peruserpermonthonGrowthandUS75 per user per month on Enterprise; Unlimited is by custom quote, with a minimum purchase of US$10,000 per year on a standard three-year term.

Salesforce

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; Salesforce has not announced its 2026 position. The Salesforce Platform builds apps and AI agents on Salesforce data.

• Strengths: Testing Center generates test cases and shows whether agents stay within their guardrails before release; the Einstein Trust Layer keeps prompts and responses out of third-party model storage and training.

• AI and governance: Agent Builder drafts a full Agentforce configuration from a plain description of what a team wants; security, privacy, and compliance controls cover each agent and workflow.

• Limitations: Agentforce is sold as add-ons to the Enterprise and Unlimited Editions; AI use is metered through consumption-based Flex Credits.

ServiceNow

• Position: included in the 2026 Gartner Magic Quadrant for AI-Augmented Low-Code Application Platforms; ServiceNow has not announced its 2026 position. App Engine builds custom workflow apps on the ServiceNow AI Platform.

• Strengths: admins work in ServiceNow Studio, professional developers in the ServiceNow IDE, and outside coding tools connect through the Build Agent SDK, all under one governance model; Creator Studio offers drag-and-drop building for simple request apps.

• AI and governance: Build Agent creates or updates an app’s workflows, screens, and embedded agents from a plain-language request; every app inherits the platform’s security rules, permissions, and audit logging, whichever tool built it.

• Limitations: App Engine apps run on the ServiceNow platform itself; Creator Studio covers simple request-based apps and comes with App Engine.

• Pricing: by quote, in App Engine Foundation and Prime packages.

Strengths, limitations, and prices above were checked on each vendor’s own site on 7 October 2026.

For enterprise planning, see the companion guide on application lifecycle management in low-code platforms.

Where Kissflow fits for governed app generation

Kissflow is an AI-powered enterprise application platform where business teams and IT build, automate, and govern their core work through business logic, not code. Business teams build on it, and IT approves and governs what they build. Founded in 2012, Kissflow serves over 1,200 customers worldwide. Kissflow is included in the Forrester AppGen and Low-Code Platforms Landscape, Q2 2026, and was named a Strong Performer in The Forrester Wave, Q1 2024.

The real contest is who can build. Developer-first platforms are built for professional developers. On Kissflow, business teams build on a platform IT approves and governs, people who are not developers can read the business logic in the blueprints, and AI and manual edits share one audit log. Godrej Consumer Products displaced Mendix. Production use spans new store opening and retail lifecycle management, oil rig shutdown and turnaround management, and end-to-end procurement and financial consoles.

Where Kissflow is not the fit: it does not replace a system of record, and teams that need AI which learns from outcomes will find that layer still in development.

Frequently asked questions

Does Gartner have a Magic Quadrant for agentic AI platforms?

No. In 2026 Gartner has no Magic Quadrant with that title. It assesses agent capabilities inside five neighboring reports, led by AI Application Development Platforms, and in a separate Hype Cycle for Agentic AI. A buyer has to read across those reports to see the full field.

Which company does Gartner call the one to beat in agentic AI?

Google. Gartner’s AI Vendor Race research calls Google the company to beat for enterprise agentic AI platforms. That finding comes from a vendor-race assessment, which does not position the wider field of agentic vendors against each other the way a Magic Quadrant does. The same release notes that Google has not yet built expert agents for specialized business problems.

What does Gartner predict about agentic AI project failure?

Gartner expects many agentic AI projects to be canceled by the end of 2027 because of rising costs, unclear business value, and weak risk controls. Read it as a governance warning: fund only projects where audit trails and access controls exist before the pilot.

How does Gartner judge agentic AI capability across these reports?

Gartner has no single scorecard for agentic AI. Each report judges agents inside its own market: building them in AI Application Development Platforms, coding with them in Enterprise AI Coding Agents, and policing them in AI Governance Platforms. A buyer should ask each vendor how agent actions are logged, who can approve or reverse them, and how agents are versioned and retired.

Who are the Leaders in the Gartner Enterprise Low-Code Magic Quadrant?

In the 2025 edition, the Leaders were Mendix, OutSystems, Microsoft, ServiceNow, Appian, and Salesforce. The Challengers were Oracle and Zoho, the Visionaries were Pegasystems and SAP, and the Niche Players were Retool and Creatio. The 2026 edition, retitled AI-Augmented Low-Code Application Platforms, evaluates 13 vendors; Gartner’s public summary does not show their positions, and Mendix and OutSystems are the two that have announced Leader placements.

Where does a governed execution layer fit alongside these platforms?

A governed execution layer handles the structured work at the edges of core systems such as SAP, Oracle, Salesforce, and Workday, without replacing them. Business teams build approval and exception workflows under IT governance, with every action logged under role-based access control, so the workflows around AI agents reach production with controls in place from the first release.

Where to start with agentic AI evaluation

Gartner’s cancellation forecast is the signal to act on: shortlist platforms where audit trails and access control exist before the first pilot. See how Kissflow governs every AI-generated app under one audit log.