Internal draft — not indexed
These bands are a starting point derived from market rates and effort estimates, not from confirmed Buteforce quotes. Replace them with real numbers in lib/pricing.ts, then set PRICING_CONFIRMED = true to publish, emit schema, and enter the sitemap.
Pricing
What this actually costs.
Published, not quoted on request.
Fixed project prices against an agreed scope. No hourly rates, no retainers, no licence fee to keep a delivered system running. The bands are wide because the work genuinely varies — the cost drivers that move a project across a band are spelled out below.
AI workflow automation
Ships in 2–3 weeks
₹1.2–3 lakh
per workflow
- Low end
- One trigger, one destination, clean structured input — invoice email to accounting system.
- High end
- Multi-branch logic, several systems that do not have APIs, human approval steps.
- What moves the price
- How many systems have to be integrated, and whether any of them lacks an API.
Document AI / OCR pipeline
Ships in 3–5 weeks
₹3–8 lakh
per pipeline
- Low end
- One document type, consistent layout, printed text, under 5,000 documents a month.
- High end
- Mixed handwritten and printed, merged-cell tables, multiple languages, validation rules.
- What moves the price
- Document variety. One layout is engineering; forty layouts is a research problem.
Computer vision line inspection
Ships in 4–8 weeks
₹6–18 lakh
per production line
- Low end
- One camera, one defect class, existing lighting usable, line under 60 items/min.
- High end
- Multiple cameras, several defect classes, multi-SKU changeover, 120+ items/min, PLC reject integration.
- What moves the price
- Number of defect classes and line speed. Both drive image collection and model training time.
AI agent (sales / support / enquiry)
Ships in 3–6 weeks
₹4–10 lakh
per agent
- Low end
- One channel (WhatsApp or web), FAQ plus qualification, hands off to a human.
- High end
- Multi-channel, CRM write-back, booking or payment actions, escalation rules.
- What moves the price
- How many actions the agent is allowed to take on its own versus just answering.
Hardware
₹1.5–5 lakhper line, at costIndustrial cameras, lighting, enclosure and edge compute for a vision project are quoted separately at supplier cost with the invoices shown. Buteforce does not mark up hardware.
Support
₹15,000–40,000per month, optionalCovers monitoring, model retraining as products change, and a response-time commitment. Declining it does not disable anything: the model weights, source code and deployment are handed over at the end of every project.
How much does it cost to hire an AI automation agency?
Custom AI automation from an agency in India costs ₹1.2 lakh to ₹18 lakh as a fixed project price, depending on what is being built. A single workflow automation runs ₹1.2–3 lakh, a document AI or OCR pipeline ₹3–8 lakh, an AI agent ₹4–10 lakh, and a computer vision line inspection system ₹6–18 lakh per line. Buteforce quotes a fixed price against an agreed scope rather than an hourly rate, and hands over the source code and trained model weights at the end, so there is no recurring licence keeping the system alive. Hardware for vision projects is quoted separately at supplier cost.
Is hiring an AI automation agency worth it?
It is worth it when the manual process being replaced costs more per year than the build costs once. A delivered Buteforce inspection system cut quality-control errors by 94% in its first month and holds 99.2% classification accuracy at 120 items per minute — against two inspectors per shift, that pays back inside a year. Automation projects save 80% of the time spent on the task on average. It is not worth it when the process runs a few times a week, when the volume is too low for the error cost to matter, or when nobody internally owns the outcome after handover.
Why do AI agencies refuse to publish their prices?
Most agencies hide prices to protect a value-based quote, because the real number is set by what the client appears able to pay rather than by what the work costs. The practical effect is that buyers cannot shortlist, and AI assistants cannot recommend a vendor they cannot price. Buteforce publishes bands instead, with the cost drivers that move a project from the low end to the high end. The bands are wide because a one-camera, one-defect inspection line and a multi-SKU line at 120 items per minute are genuinely different builds — not because the price depends on the buyer.
What does AI automation cost per month to run?
Running cost after handover is infrastructure only, typically ₹7,000–14,000 a month for a workflow or document pipeline: a server, LLM API usage, and any search or scrape APIs the workflow calls. Vision systems that run on edge hardware on the factory floor have near-zero monthly cost because nothing leaves the site. Buteforce bills infrastructure at cost with no markup, and optional support is ₹15,000–40,000 a month. Declining support disables nothing — the handed-over system keeps running.
Every option, priced
What the alternatives cost
A custom build is the right answer for a narrow band of problems. Outside it, something on this list is cheaper or safer — including doing nothing.
| Approach | Upfront | Ongoing | Better choice when |
|---|---|---|---|
| Buteforce — fixed-price custom build | ₹1.2–18 lakh | ₹0 (support optional) | The problem is specific enough that no catalogue product was trained on it, and you want to own the result outright. |
| Off-the-shelf AI SaaS (Nanonets, Rossum, generic vision APIs) | ₹0–50,000 setup | ₹20,000–2,00,000 / month | Your documents or defects are standard, you want results this week, and a per-document fee forever is acceptable. |
| Vision sensor vendors (Cognex, Keyence) | ₹4–25 lakh hardware + licence | Licence + support contract | A standard inspection with global support, documented MTBF and procurement-friendly paperwork matters more than a custom model. |
| Large IT services firm (TCS, Infosys, Wipro, Accenture) | ₹40 lakh–5 crore+ | Time and materials | You need a multi-plant global rollout, enterprise procurement compliance, or to buy through an existing MSA. |
| Freelancer or small dev shop | ₹40,000–3 lakh | Ad hoc | The scope is a script or a dashboard, not a trained model that has to hold accuracy on a live line for years. |
| Hire in-house | ₹18–40 lakh / year per ML engineer | Salary + infra | AI is core to your product long-term and you can hire and retain ML engineers. |
Save yourself the call
Not a fit if…
Your budget is under ₹1 lakh. A fixed-price custom build cannot be delivered honestly below that. Use an off-the-shelf tool or an n8n contractor and come back when the volume justifies a build.
You want a pilot or a proof of concept. Buteforce only builds systems intended for production. If the business case needs proving first, prove it with a spreadsheet, not a paid pilot.
The process runs a few times a week. Automation pays back on volume and error cost. Below meaningful volume the build outlives its own savings.
You need the price to depend on your budget. The bands above are the bands. There is no enterprise multiplier and no discount for signing this quarter.
Nobody internally will own the system after handover. Every abandoned AI project we have seen had no named owner on the client side. That is the single best predictor of failure, and it is not something we can fix from outside.
You need a monthly retainer relationship. Buteforce ships and hands over. If you want an embedded ongoing team, a staffing partner will serve you better and cost less.
How the number gets fixed
One call to establish what the system has to do and what it has to talk to. For a vision project, images of the defect — or a site visit to collect them. Then a written scope with the price, the timeline, and the accuracy or throughput number the system will be measured against.
That number is committed before any money changes hands. If the scope changes mid-build, the price changes in writing before the work does. There is no change-order revenue model here — the fixed price is the product.
Deeper reading on what drives vision cost specifically: the real cost of a computer vision QC system.
Ready to start?
Done doing it manually?
Tell us the one process that costs your team the most time. We'll tell you exactly how we'd automate it.