Fluid Topics vs Senso (2026)

Ready made connectors bring content in from authoring and repository systems such as Adobe Experience Manager, Paligo, Heretto, MadCap, Confluence, SharePoint, Salesforce, Zendesk and GitHub. The helpdesk integration surfaces documentation and AI assisted reply suggestions inside the support tool. These connectors read content into the knowledge hub and deliver it outward. Fluid Topics names none that lets an agent write to or take an action in those systems.

SourceFluid Topics, fluidtopics.com homepageread 2026-09-21

Content reaches Senso by upload, the API and connected agents, and the review-publish skill publishes approved drafts to a destination. Reading a website on a schedule and connecting Slack are both coming. Shopify, Stripe and website crawling come as code recipes. There is no native, authenticated connector that lets an agent read and act in outside systems.

SourceSenso, docs.senso.ai start, agent-skills and recipe pagesread 2026-09-21

The RAG chatbot ships with prebuilt orchestration over embeddings and vector search. The Agentic Chatbot handles multi step queries by reasoning over intent and coordinating several retrieval tools across the documentation. Both run as fixed pipelines inside Fluid Topics' own assistants. Fluid Topics names no workflows that customers build themselves to mix deterministic nodes with agent steps, and no built in branching, retries or fallbacks.

SourceFluid Topics, capabilities/agentic-ai and fluidtopics.com homepageread 2026-09-21

A fixed six stage loop (ingest, evaluate, remediate, verify, publish, re-observe) runs over a pipeline that ingests, compiles, queries and generates. A quickstart skill sequences the other skills for a customer's coding agent. The pipeline is Senso's own, and customers cannot build workflows with branching, retries, fallbacks or deterministic nodes.

SourceSenso, senso.ai llms.txt and docs.senso.ai concepts and agent-skillsread 2026-09-21

Each incoming customer inquiry is analyzed and reply suggestions for support agents are generated automatically, so the AI starts work as soon as an inquiry arrives. Otherwise the chatbots, in product help and helpdesk integration answer when a user asks, across portals, embedded help, the helpdesk and MCP clients.

Sourcefluidtopics.com/capabilities/artificial-intelligenceread 2026-09-21

An organization's tracked prompts run against AI models on a schedule, and the results feed the gap report. Otherwise, people and agents start work through the application, API, CLI and MCP server.

SourceSenso, docs.senso.ai analytics and mcp-serverread 2026-09-21

Product content from many sources and formats is centralized into one hub kept aligned with product updates. Fluid Topics provides core RAG services, including embeddings, semantic search and prompt management. Visibility rules by role, group and metadata are applied when feeding LLMs, and chatbot answers carry citations and links to the referenced content.

Sourcefluidtopics.com/capabilities/artificial-intelligenceread 2026-09-21

Files (PDF, Word, Excel, PowerPoint, HTML, CSV, JSON, XML) and raw text come in through the API. Senso compiles each source before it can be queried and keeps sources in a knowledge base folder tree with permissions, tags and version history. Search returns a grounded answer that cites the source chunks behind it, with their relevance scores, and unanswered questions are filed as gaps.

SourceSenso, docs.senso.ai concepts, knowledge-base and mcp-serverread 2026-09-21

The Agentic Chatbot handles multi step queries. Fluid Topics names no session, conversation or long term memory for its assistants or for agents using its MCP server. The unified product knowledge hub holds the customer's content. The analytics suite keeps a record of user preferences and history, which personalizes content delivery.

Sourcefluidtopics.com/capabilities/agentic-airead 2026-09-21

Shared context in Senso is the organization's knowledge base, which people and their agents query and add to. There is no separate session, conversation or agent memory.

SourceSenso, docs.senso.ai shared-context and conceptsread 2026-09-21

For support teams, Fluid Topics generates reply suggestions to incoming inquiries. Support agents review, validate and customize each suggestion before sending, so AI drafted replies reach customers only after a person approves them. This review step is fixed and belongs to that one feature. Fluid Topics names no approval steps that a customer can insert at a node, workflow or policy layer, and no autonomy modes that treat low risk and high risk actions differently.

Sourcefluidtopics.com/capabilities/artificial-intelligenceread 2026-09-21

In the verify stage, only failures and close calls from the automated check escalate to a human reviewer. Generated drafts wait in review until a person approves or rejects them, and nothing is published before that. The loop sets out which stages require a human.

SourceSenso, senso.ai llms.txt and docs.senso.ai agent-skillsread 2026-09-21

Fluid Topics holds ISO 27001:2022 certification and offers role based access control, single sign on and encryption. Its MCP server enforces content access rules through OAuth2, so agents receive only what their credentials allow, and admins can enable or disable individual tools. Content used for embeddings is processed internally and not exposed to external systems.

Sourcefluidtopics.com/capabilities/agentic-airead 2026-09-21

Admin, collaborator and viewer roles sit alongside custom roles built from granular permissions, each an action on a resource. API keys are scoped to knowledge base folders, and each folder has its own access list. OAuth connections to the MCP server act only with the approving person's permissions, using PKCE and rotating refresh tokens. There is no published certification or audit report.

SourceSenso, docs.senso.ai permissions and mcp-serverread 2026-09-21

The AI comes with what Fluid Topics calls full audit logging, including audit logs for the chatbot. The analytics suite tracks how users interact with AI features such as chatbots and summarizers. Fluid Topics names no step by step view of the prompts, retrieved passages, tool calls and outputs of an AI run.

Sourcefluidtopics.com/capabilities/artificial-intelligenceread 2026-09-21

Every query is recorded as a query turn in search history. Each turn is tagged by source, such as mcp, so agent traffic shows separately in analytics, and unanswered queries go into a gap report. The record shows that each request was made and whether it found an answer. There is no step by step view of the prompts, retrieved chunks, tool calls and outputs in an agent's run.

SourceSenso, docs.senso.ai mcp-server and analyticsread 2026-09-21

The AI is a cloud native solution designed for enterprise environments. No self hosted, dedicated cloud, VPC or region selection option is published.

Sourcefluidtopics.com/capabilities/artificial-intelligenceread 2026-09-21

Customers reach Senso as a hosted API, MCP server and application at senso.ai. There is no self hosted, dedicated cloud or VPC option, and customers cannot choose a region.

SourceSenso, docs.senso.ai start, concepts and mcp-serverread 2026-09-21

The platform ships its own AI tools, ready to use. They are a RAG chatbot with prebuilt orchestration and a customizable UI, an Agentic Chatbot, AI reply suggestions for support agents, AI search and translation on the fly. These built in assistants are a fixed set over one knowledge hub, and none of them is a template, blueprint or role specific agent that a customer picks from and adapts. The agentic use cases for onboarding and RFP replies show what a customer's own agents can do through the MCP server.

Sourcefluidtopics.com/capabilities/artificial-intelligenceread 2026-09-21

Teams can install a set of agent skills (quickstart, context-layer, evaluate-remediate, generate-verify, publish, search and review-publish) all at once or one at a time, including from the Shipables catalog, for Claude Code, Codex, Cursor, Copilot, Gemini and Cline. Reusable content type templates and recipes, such as a help center crawler and a blog to social converter, sit alongside them.

SourceSenso, docs.senso.ai agent-skills, content-types and recipe pagesread 2026-09-21

The RAG chatbot ships with prebuilt LLM connectors and fine tuned prompts. OpenAI is among the platform's ready made connectors, so the customer connects the language model its assistants use. Customers bring their own model providers. Fluid Topics does not say which providers are supported or whether the choice can be set per workspace.

SourceFluid Topics, fluidtopics.com homepageread 2026-09-21

Customers choose which AI models Senso monitors, and a raw context endpoint lets them feed Senso's retrieval into their own LLM. Customers cannot choose the model Senso itself uses to generate answers and drafts.

SourceSenso, docs.senso.ai analytics and conceptsread 2026-09-21

A fully operational MCP server lets any MCP compatible agent call Fluid Topics for its product knowledge, authenticated with OAuth2. APIs, alongside semantic search and RAG services, connect product content to AI applications without replacing authoring systems.

Sourcefluidtopics.com/capabilities/agentic-airead 2026-09-21

A REST API at apiv2.senso.ai covers ingesting, compiling, querying and generating, authenticated with scoped organization API keys. An MCP server's tools query, read and ingest into the knowledge base over OAuth or an API key, and a CLI serves scripts and CI.

SourceSenso, docs.senso.ai concepts, authentication, mcp-server and CLIread 2026-09-21

The analytics suite measures documentation performance. It covers search analytics and queries with no result, user journeys, topic ratings and comments, case deflection, and GenAI metrics that track how users interact with chatbots and summarizers. None of these scores the quality of the AI's answers over time or tests an agent before release.

Sourcefluidtopics.com/capabilities/content-analyticsread 2026-09-21

Before anything Senso generates is published, an automated gate checks each draft for source support, faithfulness, freshness and permission. Drafts are checked against the organization's own sources, and the verification results are listed for review. Evaluation runs can be batched in parallel across tracked questions.

SourceSenso, senso.ai llms.txt and docs.senso.ai use-cases/test-ai-answersread 2026-09-21

Content goes out through portals, embedded help, the helpdesk and the MCP server. Fluid Topics names no browser, desktop or computer control by an agent.

SourceFluid Topics, fluidtopics.com homepage and capabilities/agentic-airead 2026-09-21

Work happens through uploads, the API, the CLI and the MCP server, and no agent controls a browser, desktop or computer.

SourceSenso, docs.senso.ai start and recipe pagesread 2026-09-21