Bots that take work
off your hands

Useful for: Estate agents and small property teams.

Why it helps: Repeated questions about location, rooms and budget take time before a viewing is even booked.

What exists: The demo filters a fictional property list in your browser.

What we need: Your current listings, selection rules and handoff process.

How we check it: Agreed sample enquiries return matching listings or a clear no-match message.

Your real listings, messaging channels and CRM connection need a separate build.

Possible foundations: grammY, Activepieces

Useful for: Services, shops and teams with repeated enquiries.

Why it helps: Customers ask the same opening hours, delivery and booking questions.

What exists: Try a small rule-based FAQ demo. It does not call an AI model.

What we need: Your FAQ, policies, tone and support channel.

How we check it: Test questions get the approved answer; unknown questions reach the handoff.

AI search, your knowledge base and real support channels are configured for your project.

Possible foundations: Chatwoot, AnythingLLM

Useful for: Studios, agencies and services with repeatable pricing.

Why it helps: A useful first estimate can replace several rounds of basic questions.

What exists: The demo calculates a sample website quote using fictional units.

What we need: Your services, price rules, exceptions and enquiry destination.

How we check it: Test combinations match your agreed price table.

Demo numbers are examples, not FollowAI prices or a binding quote.

Useful for: Teams preparing CSV files for a CRM or mailing tool.

Why it helps: Repeated and incomplete rows make customer records harder to use.

What exists: The working demo processes pasted CSV entirely in your browser.

What we need: Your column format, matching rules and import requirements.

How we check it: Known duplicates are removed, invalid rows are reported and the result opens correctly.

A live CRM integration and additional column rules are scoped separately.

Useful for: Businesses handling enquiries across forms and inboxes.

Why it helps: Requests get missed when someone has to copy every detail manually.

What exists: FollowAI has an enquiry intake component; the complete CRM route is built and tested per client.

What we need: Your form, CRM, qualification rules and team responsibilities.

How we check it: A test enquiry reaches the agreed destination once, with the right fields.

The existing component is not proof of a finished connection to your CRM.

Possible foundations: Activepieces

Useful for: Owners improving a service website.

Why it helps: A site can look good while leaving the next step unclear.

What exists: FollowAI’s public website audit is available now.

What we need: Your public website and the action you want visitors to take.

How we check it: The report points to observable issues and useful next steps.

The free audit does not change your website. Implementation is a separate project.

Useful for: Small customer service teams.

Why it helps: Separate inboxes make it difficult to see who owns the next reply.

What exists: An open-source inbox is available as a starting point.

What we need: Your channels, users and access requirements.

How we check it: Test messages arrive, are assigned and can be answered through supported channels.

Hosting, channel availability, paid features and integrations are agreed before setup.

Possible foundations: Chatwoot

Useful for: Support teams answering product or policy questions.

Why it helps: Finding the right policy is often slower than writing the reply.

What exists: A document assistant foundation can be adapted to your material.

What we need: Current documents, access rules and example questions.

How we check it: Sample answers cite the relevant source or flag missing information.

Your team reviews replies; source links do not guarantee every answer is correct.

Possible foundations: AnythingLLM

Useful for: Teams with useful knowledge spread across many files.

Why it helps: Staff lose time searching folders or asking the same internal questions.

What exists: Document parsing and search foundations are available.

What we need: Representative files, permissions and update frequency.

How we check it: Agreed questions retrieve the right source within the permitted collection.

We test your file formats, access requirements and retrieval quality before launch.

Possible foundations: RAGFlow, Docling

Useful for: Studios, local services and appointment-based businesses.

Why it helps: Scheduling involves repeated questions about time, service and contact details.

What exists: Bot and workflow foundations can be connected for your process.

What we need: Services, availability, timezone, calendar and cancellation rules.

How we check it: Test bookings respect availability and avoid duplicate reservations.

Calendar access, message charges and edge cases are part of the agreed scope.

Possible foundations: grammY, Activepieces

Useful for: Small shops and services selling through Telegram.

Why it helps: Orders in free-form chats often miss items, quantities or contact details.

What exists: A Telegram bot framework provides the starting point.

What we need: Catalog, stock process, options and order destination.

How we check it: A sample order includes the chosen items and reaches the right person.

Payments, stock synchronization and delivery services are separate connections.

Possible foundations: grammY

Useful for: Operations teams entering invoice or document data.

Why it helps: Copying totals, dates and supplier names is repetitive and easy to get wrong.

What exists: A document parsing foundation is available.

What we need: Representative documents, target fields and review rules.

How we check it: Sample files show correct fields or clearly flag items needing review.

Accuracy depends on your documents. Accounting integration and tax decisions are outside the basic extraction scope.

Possible foundations: Docling

Useful for: Teams with frequent calls and follow-up work.

Why it helps: Decisions disappear when nobody has time to write clear notes.

What exists: Speech transcription and workflow foundations can be connected.

What we need: Sample recordings, languages and the destination for approved tasks.

How we check it: A reviewed example preserves decisions and assigns only agreed actions.

Participants, recording permissions, transcription costs and task integrations are agreed per project.

Possible foundations: faster-whisper, Activepieces

Useful for: Experts, creators and small marketing teams.

Why it helps: Useful ideas remain in recordings because repurposing them takes time.

What exists: Transcription and workflow tools provide a starting point.

What we need: Your recordings, examples of your voice and content formats.

How we check it: Drafts stay grounded in the recording and are easy for you to edit.

Human editing, fact checks and approval remain part of publication.

Possible foundations: faster-whisper, Activepieces

Useful for: Creators and teams maintaining several social accounts.

Why it helps: Posting becomes inconsistent when drafts and dates live in separate documents.

What exists: An open-source publishing foundation is available.

What we need: Your accounts, publishing roles and formats.

How we check it: An approved test post reaches a supported connected channel.

Supported networks, account permissions, hosting and license obligations are checked for your setup.

Possible foundations: Postiz

Useful for: Businesses with a clear B2B offer.

Why it helps: A long contact list is less useful than a smaller list with a reason for each match.

What exists: A web research foundation can support the collection process.

What we need: Your offer, market, qualification criteria and allowed sources.

How we check it: Each accepted company has the required evidence and matches the agreed criteria.

Research does not guarantee interest, meetings or sales. Outreach is a separate scope.

Possible foundations: Crawl4AI

Useful for: Teams losing track of warm enquiries.

Why it helps: A good conversation can stall when nobody owns the next step.

What exists: A workflow foundation can connect your existing tools.

What we need: Your enquiry source, team and follow-up rules.

How we check it: Test enquiries produce the right reminder without duplicates or unwanted messages.

Sending rules, consent and approval are agreed before any automatic messaging.

Possible foundations: Activepieces

Useful for: Businesses with repeated incoming phone questions.

Why it helps: Routine calls can interrupt work while customers still need a timely answer.

What exists: A voice-agent framework is available; FollowAI’s client-specific service still needs to be built.

What we need: Your scenario, languages, phone setup and escalation rules.

How we check it: Test calls complete the agreed task or reach a person safely.

Telephony, AI usage, recording rules and real-call testing are included in project planning.

Possible foundations: LiveKit Agents

Useful for: Individuals and small teams who want a personal AI helper.

Why it helps: A familiar chat can be a simple place to request work and receive a useful summary.

What exists: OpenClaw is an available platform. We configure and test your assistant as a custom project.

What we need: Your task, preferred hosting, model account and the information the assistant may use.

How we check it: The assistant handles agreed test messages, runs the chosen task and can be paused or updated.

Hosting and AI usage are separate. We agree access and approval rules before connecting real accounts.

Tools we can use: OpenClaw ↗

Useful for: Businesses and freelancers repeating the same work across apps.

Why it helps: A clear workflow can handle routine steps while you focus on the exceptions.

What exists: n8n provides the workflow platform. Your connections and business rules are built and tested for you.

What we need: The apps you use, example inputs, destination fields and what should happen when a step fails.

How we check it: A normal request reaches the right destination once; failure and retry examples behave as agreed.

Use your own n8n workspace or an agreed deployment. Hosting, paid plans and connected service costs are quoted separately.

Tools we can use: n8n ↗

Useful for: Office workers, freelancers and small teams with recurring document tasks.

Why it helps: Saving a process makes it easier to repeat a task without rewriting the brief every time.

What exists: Claude supports skills and connectors. We prepare your task, files and any agreed integration.

What we need: One recurring task, approved source files, a sample result and your Claude account features.

How we check it: New test material produces a draft in the agreed format; you can check sources and update the instructions.

Claude, Claude Code and Cowork features depend on your plan. MCP connections are scoped individually; finished drafts still need your review.

Tools we can use: Claude skills and plugins ↗

Useful for: Shops, service businesses and creators with an offer to show.

Why it helps: A short visual demonstration helps explain an offer and gives you creative options to test.

What exists: We produce your video to order using tools such as Kling, Veo / Flow and Runway.

What we need: Product references, message, audience, examples you like and where the video will be used.

How we check it: The agreed scenes, product details, captions, sound and export formats pass the final review.

Length, scene count, revisions, music rights and generation costs are agreed in the quote. Advertising spend is separate.

Tools we can use: Google Flow ↗, Runway workflows ↗

Useful for: Experts, educators and businesses explaining products or services.

Why it helps: A reusable presenter and translated content can help you update lessons or reach another audience.

What exists: HeyGen provides avatar and translation tools. Your script, production and language review are prepared for your project.

What we need: Your script or original video, target language and an approved avatar or permission to use the person shown.

How we check it: You approve the words, voice, lip sync, subtitles and final format in the agreed language.

We use your own likeness or approved material. Available languages, voice options, duration and review scope are confirmed before work.

Tools we can use: HeyGen ↗

Useful for: Freelancers, agencies and businesses with a clear target customer.

Why it helps: Daily prospect research takes time when the same sources need checking again and again.

What exists: We build the search, qualification rules and delivery around your offer. The complete bot is a custom project.

What we need: Your offer, target market, selection criteria, agreed sources and preferred delivery channel.

How we check it: Scheduled test runs find matching records with evidence, exclude known duplicates and report unavailable sources.

The bot finds potential customers; replies, meetings and sales depend on the offer and follow-up. Sending messages is a separate agreed task.

Possible foundations: Crawl4AI, Activepieces

Useful for: Business owners, creators and marketers who need to stay close to their niche.

Why it helps: Useful changes are easy to miss when news, product launches and competitor updates are scattered across sites.

What exists: We connect the chosen sources and build a digest for your niche as a custom project.

What we need: Your niche, competitors, topics, approved sources and how often you want updates.

How we check it: Test updates appear once with a source and date; summaries separate observed changes from suggested ideas.

The bot monitors the agreed sources. A signal is something to investigate, not proof of market demand or a prediction of success.

Possible foundations: Crawl4AI, Activepieces

Useful for: Small businesses, freelancers and people keeping track of recurring expenses.

Why it helps: Small recurring payments and price changes are easy to overlook across files and invoices.

What exists: We build the import, checks and report around your file formats as a custom project.

What we need: Sample expense exports, subscription details and your rules for flagging unusual or repeated charges.

How we check it: Known repeated charges and price changes are flagged against the original records; legitimate repeats can be marked as expected.

Savings depend on what your records reveal. You approve any cancellation or payment change; the bot does not move money.

Possible foundations: Docling, Activepieces