Architect-Programming BAS Small Business

It is necessary to configure the CRM system according to the company's business processes: import the contact database from Excel, set up sales funnels, lead and deal cards, statuses, automation, analytics, reporting, user access rights, and other necessary functions for the full operation of the company.

TECHNICAL TASKImplementation and Integration of BAS Accounting + BAS Trade Management with Websites and Services1. General Information It is necessary to combine and configure the joint operation of the BAS Accounting and BAS Trade Management systems, as well as connect 3 websites to the BAS with correct data synchronization.2. Project Tasks2.1. BAS Integration Combine BAS Accounting with BAS Trade Management Set up correct data exchange between the systems Ensure a unified logic for accounting sales, inventory, and finances2.2. Website Connection Connect 3 websites to the BAS: Website on WordPress Website on OpenCart Website on the Khoroshop platform Integration functionality: transfer orders from websites to the BAS synchronize order statuses account for payments synchronize customers (optional) synchronize products and prices2.3. Inventory Accounting Currently, there is 1 general warehouse in the BAS It is necessary to divide it into 3 separate warehouses Set up: correct accounting of balances linking orders from websites to the corresponding warehouses write-off of goods from the warehouse during sales2.4. Service Connection Checkbox: fiscalization of sales transfer of receipts correct operation with the BAS Nova Poshta (NP): creation of TTH from the BAS transfer of order data synchronize delivery statuses

It is necessary to set up a connection for vapi.ai through make.com to the site on prom.ua. The bot should inquire about the price, availability, and description of the product on the site.

Automation / AI Content Factory for Creating Short Videos Looking for a specialist to develop an automated system for mass creation of AI content for TikTok / Instagram Reels. The main task is to create not a template that generates similar videos each time, but a full-fledged content factory capable of regularly producing different videos: with new plots, scenes, locations, actions, and visual solutions. ### How the process should roughly work: 1. Idea and Script Generation - the system generates a new theme/idea for the video; - creates a script; - breaks it down into separate scenes/frames; - for each scene, a description of what should happen is formed. It is important that the videos differ from each other each time, and the system does not constantly repeat the same ideas, scenes, and structure. 2. Image Creation for Scenes - for each scene, an image is automatically generated; - the images must correspond to the overall story of the video; - there should be the ability to work in two modes: Mode 1 — Different Characters and Different Stories Each video can have new heroes, a new visual style, new locations, and a plot. Mode 2 — AI Blogger / Constant Character One character is fixed, who maintains a recognizable appearance between videos, but at the same time changes: - clothing; - locations; - situations; - actions; - plot; - environment. So, there should be a conditional AI blogger who constantly appears in new content. 3. Bringing Scenes to Life + Voice Preferably use AI models that can immediately: - bring the image to life; - create speech / lip-sync; - generate or synchronize voice. That is, it is not necessary to do voiceover as a separate stage if the chosen AI service can generate video with speech immediately. But the architecture should consider the cost of generation: if separate voice generation + subsequent animation turns out to be significantly cheaper or of higher quality, that option can also be used. Language of speech: - Russian; - Ukrainian. 4. Video Assembly The output should be a set of ready scenes or a full vertical video. Additionally, the following can be implemented: - automatic scene connection; - subtitles; - music; - sound effects; - transitions; - automatic editing. Editing is not a mandatory part of the first stage of the project — it depends on the cost and complexity of implementation. 5. Auto Posting Optionally: - automatic publication on TikTok; - automatic publication on Instagram Reels. Auto posting can also be separated into a distinct stage. ### Important Principle of the System It must be a scalable AI content factory, not a rigidly fixed workflow for generating identical videos. The system should be able to create a large amount of different content through: - generating new ideas; - different scripts; - different scenes; - different locations; - different character actions; - different visual compositions; - controlling repetitions of already used ideas. At the same time, there should be the ability to fix certain parameters, such as: - one constant AI character; - specific visual style; - channel theme; - video format; - language; - video duration. ### Example Logic: Theme / Niche → New Idea → Script → Scene Division → Image Generation → Scene Animation with Speech → Video Assembly → Saving / Publication ### Regarding Implementation Preferably use n8n, but open to other solutions if the performer can offer a more stable architecture. The system should be modular so that in the future it can replace: - LLM; - image generation model; - image-to-video / talking image model; - voice service; - editing service; - publication method. It is desirable to provide for the storage of the history of already created content so that the system can take into account previous videos and reduce the number of repeating ideas. ### What I want to receive from the performer in the response: - your experience with AI automation and n8n; - examples of similar projects; - what architecture you propose; - which AI models / APIs you suggest using; - how you propose to solve the issue of content diversity; - how you propose to ensure the consistency of the constant AI character; - approximate cost of one generated video; - cost of automation development; - what can be implemented within the first MVP. I am ready to consider phased implementation: Stage 1: working content factory until the generation of ready scenes / video. Stage 2: automatic editing. Stage 3: automatic publication on TikTok / Instagram.

There is a Corp. license. It is necessary to configure the control center for 1C and SQL to understand where the performance drops and where the hangs occur.