AI assistants

10 AI assistants, from retrieval to agents

Retrieval chatbots, a healthcare compliance assistant, a meeting copilot and agents that work inside the tools people already use.

Get a plan like thisSee the Automate practice

What the set covers

The set spans retrieval over private documents, a compliance assistant, a meeting copilot, a writing agent in Telegram, a caption generator and model predictions read with GPT-4 Vision. Open a row for what each one does and the stack it runs on.

The projects 10

Open a row for what was built, the details that mattered and the stack.

  • Healthcare regulations and compliance assistantHealthcare. Automation.

    A question answering surface over healthcare regulation, where being confidently wrong is the failure mode that matters. Answers are grounded in the regulatory text and cite the clause, and the assistant declines rather than guesses.

    • Answers grounded in regulatory source text with the clause cited
    • Explicit refusal when the corpus does not support an answer
    • Regulated-source ingestion with the effective date carried through
    • Query and answer log retained for review

    Stack: PythonOpenAILangChainVector searchFastAPI

  • Meeting copilot, a SaaS assistantB2B SaaS. Custom software.

    A product rather than a script: meetings are transcribed, summarized into decisions and actions, and made searchable across the history, with the multi-tenant plumbing a SaaS needs behind it.

    • Transcription and speaker separation per meeting
    • Summary split into decisions, actions and open questions rather than one blob
    • Search across the meeting history, not only within one call
    • Multi-tenant data isolation with per-workspace access control

    Stack: PythonOpenAIFastAPIReactPostgreSQL

  • Retrieval chatbot on Azure OpenAI and Azure AI SearchEnterprise IT. Automation.

    The same retrieval pattern built inside Azure, for an organization whose data was not permitted to leave its tenant. Indexing, embedding and inference all run on Azure services under the customer's own subscription.

    • Indexing and retrieval through Azure AI Search
    • Inference through Azure OpenAI inside the customer tenant
    • Document ingestion pipeline with incremental reindexing
    • Access scoped through the tenant's own identity provider

    Stack: Azure OpenAIAzure AI SearchPythonAzure Functions

  • Document question answering over a private document setKnowledge management. Automation.

    A chat surface over a company's own documents. Questions are answered from the document set rather than from the model's general knowledge, and every answer carries the passages it came from so a reader can check it.

    • Retrieval augmented generation over a private corpus, not a public index
    • Chunking and embedding pipeline rebuilt on document change
    • Citations returned with every answer
    • Refusal path when the corpus does not contain the answer

    Stack: PythonOpenAILangChainVector searchFastAPI

  • DocTalk, a document conversation surfaceKnowledge management. Automation.

    Upload a document, then ask it questions. Built for people working through long documents one at a time rather than searching a whole library, so the retrieval is scoped to the file in front of them.

    • Per-document session scope, so answers cannot drift to another file
    • Handles long documents by chunk with overlap rather than truncating
    • Passage highlighting against the source view
    • Supports the document formats people actually send

    Stack: PythonOpenAILangChainStreamlit

  • LangChain question answering botKnowledge management. Automation.

    A question answering service built on LangChain, with the retrieval chain, prompt templates and evaluation harness kept as code so a change to the prompt is reviewable in a pull request.

    • Retrieval chain defined in code, versioned with the repository
    • Prompt templates parameterized rather than pasted per environment
    • Evaluation set run against changes before release
    • Swappable model provider behind one interface

    Stack: PythonLangChainOpenAIVector search

  • Telegram agent for writersPublishing. Automation.

    A writing assistant that lives where the writer already is. Drafting, rewriting and research run as a Telegram conversation, with per-user state so a thread picks up where it left off.

    • Conversational state held per user across sessions
    • Draft, rewrite and summarize as distinct commands rather than one prompt
    • Long output split to fit the platform's message limits
    • Rate limiting per user to keep model cost bounded

    Stack: PythonTelegram Bot APIOpenAIRedis

  • Caption generator on GPT-4oMarketing. Automation.

    A tool that writes post captions from an image and a short brief, in a house voice supplied as examples rather than as adjectives. Built for a team producing a high volume of social posts against a fixed tone.

    • Vision input, so the caption describes the actual image
    • Tone carried by supplied examples rather than instruction words
    • Variant generation with length and platform constraints applied
    • Batch mode for a content calendar rather than one post at a time

    Stack: OpenAI GPT-4oPythonFastAPIReact

  • Machine learning predictions augmented with GPT-4 VisionApplied machine learning. Data pipelines.

    A conventional model handled the structured features while a vision model read the images the structured data described. The two signals were combined so the prediction used what the photograph showed as well as what the record said.

    • Structured model and vision model combined at the feature level
    • Vision output constrained to a fixed schema rather than free text
    • Fallback to the structured model when the image is absent or unusable
    • Per-source contribution reported for each prediction

    Stack: PythonGPT-4 Visionscikit-learnpandas

  • Forex assistantFinancial markets. Automation.

    A conversational surface over currency market data: rates, movement over a window, and the definitions traders ask about, answered from live data rather than from the model's memory of it.

    • Live rate lookup through a market data API rather than model recall
    • Historical windows computed on request
    • Terminology answered from a curated reference set
    • Explicit statement that nothing returned is trading advice

    Stack: PythonOpenAIMarket data APIFastAPI

More of the work

Next step

Need something in this shape?

Say what the system has to do in a few sentences. The free plan says whether to build it, what it would cost and how long it would take.

Get a plan like thisBook a technical walkthrough

We usually reply within 2 business days with a free plan for what to do first. No call needed.

Both are optional. The site and every form work either way, and nothing here runs until you switch it on.

Google Analytics also receives the 17 on-site actions this site records, among them a request delivered through the form and the need, budget range and timing picked on it. Turning a category off later deletes the cookies it set on this site. Our cookie policy lists every one.