How to chat with Lampi?
Lampi AI Chat helps you ask questions, work with your selected context, search files or the web, and use reusable skills for specific workflows. You can choose the right assistant, context, and capabilities before sending your message so Lampi answers with the right level of depth.
Understanding the Chat
To make the most of Lampi AI, it helps to understand what Lampi can use when answering your message.
In chat, you can combine:
- Context - files, notes, meetings, or integrations selected in the context panel
- Search - file search or web search, with a Fast or Pro mode
- Skills - reusable ways of working for specific tasks
- Deep Search - a deeper research workflow for complex questions
- Additional tools - integrations and actions available from the tools panel
You can select multiple tools before sending a message. Lampi will use the selected context, searches, skills, Deep Search, and MCP tools when they are relevant to the request. More tools can give Lampi broader access, but they can also make the answer slower or less focused, so select the capabilities that match the task.
For file and web search, Fast runs a direct lookup for targeted answers. Pro uses a deeper workflow for research, analysis, comparison, verification, or summaries.
Tools give Lampi access to a specific action, such as searching the web or using an integration. Skills guide Lampi through a reusable way of working, such as researching a company, reviewing a document, or producing a specific output format.
What can you use in chat?
🔍 Files Search
What the assistant will do: Search across the files you selected in the context panel and use the most relevant content to answer your question.
How to use it: Select the relevant documents in the context panel, activate Files Search in the chat bar, then choose Fast or Pro.
- Fast - best for quick, targeted questions, keyword-style searches, and specific facts from your files.
- Pro - best for deeper document work, such as summaries, comparisons, cross-document analysis, and questions that require a more thorough review.
🌐 Web Search
What the assistant will do: Search public web sources and use the results to answer questions that require current or external information.
How to use it: Activate Web Search in the chat bar, then choose Fast or Pro.
- Fast - best for quick web lookups, recent facts, and simple public information.
- Pro - best for deeper web research, source comparison, verification, and more complete answers.
⚡ Skills
What the assistant will do: Use a reusable workflow that gives Lampi extra guidance for a specific type of task.
Skills can include instructions, examples, preferred output formats, and access to relevant tools. They are useful when you want Lampi to follow a consistent process without rewriting the same long prompt each time.
How to use it:
Open the tools panel and select a skill from Built-in skills or My skills. You can also type / in the chat input to search and activate a skill quickly.
For more details, see Skills.
🔬 Deep Search
What the assistant will do: Run a comprehensive research workflow for complex questions. Deep Search plans the research, gathers evidence from multiple sources, browses relevant pages, tracks key findings, and produces a structured, well-supported answer or report.
Deep Search is designed for deeper synthesis than a standard chat answer. It can work with selected files, web sources, or both when relevant. It is especially useful when you need to compare sources, identify gaps or uncertainty, build a market or strategic report, or prepare a diligence-style memo.
How to use it: Open the tools panel in the chat bar and activate Deep Search before sending your request. Write a clear brief that includes the topic, expected output, audience, language, and any specific sections you want.
🔧 MCP Tools
What the assistant will do: Access additional tools and integrations, such as specialized databases, external services, document actions, or custom capabilities available in your workspace.
How to use it: Open the tools panel and select the relevant tool or integration. Available tools depend on your workspace and subscription.
For more details, see Tools & MCP Library.
How to know what to use?
When to use Files Search
Use Files Search when the answer should come from documents, notes, uploaded files, or other context selected in Lampi.
- Choose Fast for specific questions, exact facts, definitions, names, figures, or sections.
- Choose Pro for summaries, document reviews, comparisons, or questions that require Lampi to inspect the documents more deeply.
When to use Web Search
Use Web Search when the answer depends on public information, recent information, or sources that are not in your selected context.
- Choose Fast for quick public facts or simple current information.
- Choose Pro when you need a more reliable answer based on multiple sources, comparisons, or verification.
When to use Skills
Use Skills when you want Lampi to follow a repeatable way of working.
Skills are best for:
- Standard company research
- Investment memo or diligence workflows
- Contract or document review checklists
- Spreadsheet, slide, or document generation workflows
- Any task where the format, process, or quality rules matter
If you often write the same long prompt, turn that process into a skill or select an assistant that already includes the relevant skills.
When to use Deep Search
Use Deep Search when you need a complete, evidence-backed answer rather than a quick lookup.
Deep Search is best for:
- Market research, company research, and competitive analysis
- Due diligence notes, investment memos, and strategic reports
- Questions that require multiple sources, structured synthesis, or explicit uncertainty
- Workflows where you want Lampi to produce a reusable report
For simple factual questions, use Web Search or Files Search. Use Deep Search when the task needs deeper planning, broader evidence gathering, and a more complete final deliverable.
When to use MCP Tools
Use MCP Tools when your task needs a specific integration, database, action, or external system. These tools vary by workspace and can extend Lampi beyond standard file and web search.
Maximize Performance
To get the best results from Lampi, consider the following strategies:
🛠️ Break down complex tasks
Rather than asking an LLM to tackle a large, multifaceted task in one go, break it down into sequential steps, addressing each part individually. This method allows the LLM to focus on one aspect of the task at a time, ensuring clarity, accuracy, and a higher quality of output.
✍️ Craft clear prompts
Providing clear and detailed instructions is essential for effective tool performance. For more, see Writing a prompt.
If you have activated file or web search, tell Lampi which sources matter most and whether it should rely only on selected context. Focusing the request helps reduce irrelevant answers and hallucinations. For more, see Reducing hallucinations in AI outputs.
🎯 Select data sources wisely
Choosing the right data sources for each action is crucial to achieving high-quality, relevant results.
The assistant will perform better when you carefully select specific data that is pertinent to the task at hand, rather than relying on a vast amount of unrelated information.
If you are working on market intelligence, select only the relevant reports and choose the search mode that matches your question: Fast for targeted facts, Pro for deeper analysis.
🤖 Choose the relevant AI Assistant
When engaging with Lampi, you can select an AI assistant configured with specific skills, guidelines, tools, and knowledge.
Assistants are useful when you want a recurring role or working style to apply by default. For example, you can use an assistant tailored for financial analysis, customer support, translation, or document review. For one-off workflows, you can also select a skill directly in chat.
For more, see Create AI assistants.
Next: Manage context in chat →