How to use AI in business operations: inside WhatsApp, your CRM, Tally and Sheets, and what stays with a person, shown on Xale AI, which answers in Malayalam.
- Published
- Updated
- Reading time
- 13 minutes
Drafted with AI tools; researched, edited and approved by Althameem Ali. Our editorial policy
On this page · 9 sections
Our co-founders are two of Xale's three founders, and we sell AI automation: we build workflows with AI where it pays for businesses in Kerala and the Gulf. So we have a stake in this topic. Every outside fact below links to its source, read on 8 October 2026, and Xale AI is shown as xale.in describes it, on the product our team built, not on a client's system.
This guide is for owners and operations heads who want AI to take work off their team without a big project. It covers what AI in operations looks like in a real system, how to add it to the tools you already pay for, and what should stay with a person. Three of our guides go deeper on the rest: which tasks to automate first, how to build AI into business software and what AI automation costs in India.
Where does AI help operations first?#
Where the work is reading and retyping, not deciding. Three questions sort any task or business process: how often it happens, how clear the rule is and what a mistake costs. To automate a process that runs many times a day on a clear rule, you need a plain automation, not AI. Frequent work that means reading messy text, such as chats, call notes, invoices and enquiries, is where AI drafts and a person checks. Anything where a mistake costs money or trust stays with a person, with AI doing the legwork. Our guide to AI agents turns those questions into a what-to-automate-first matrix, and its strongest advice holds here too: rules before agents.
What does AI in operations look like in a real system?#
The example we know best is Xale AI, the assistant inside Xale CRM, which is made by Xale Private Limited in Kozhikode and built by the team behind Marketlube, as our Xale case study tells. Everything in this section is as xale.in's AI CRM page (opens in a new tab) describes it (updated 8 October 2026), with the telecaller details from its AI telecaller page (opens in a new tab) (updated October 2026).
- Switch it on, then ask. An admin turns Xale AI on for the workspace in Settings. You then ask it by chat, from Ask AI in the header or Ctrl+K on any screen, or start a live voice call that speaks Malayalam by default and English on a click, can be interrupted mid-sentence and shows a running transcript.
- Ask what a manager asks every morning. xale.in's own example is a branch head in a Kozhikode counselling office who starts a voice call at 9.30 am and asks, in Malayalam, how many visa-team follow-ups are overdue and due today, then which belong to one counsellor, then who is working right now. The answers come from the same follow-ups, calls and reports the team already works in.
- Check the working. As xale.in describes it, every answer lists what it looked up and which filters it applied, so anyone can check a number. Lead, deal and follow-up answers cover only what the person asking is allowed to see; call counts and the user list cover the whole workspace.
- Get the report without building it. It runs reports with the CRM's own report measures, draws the chart and gives a day report per person, so "What did Priya do today?" is a question, not a spreadsheet.
- Describe a rule; a person saves it. Describe an automation in plain English or Malayalam and the draft opens in the automation builder (opens in a new tab), with its trigger, conditions and actions, for someone to preview, change and save. Once saved, every run is queued, retried and logged like any other automation.
- Transcribe the call nobody wrote up. A recorded call is transcribed on demand, tuned for calls that mix Malayalam and English, and the transcript is saved on the call.
- Hand it the phone, not a list. In telecaller mode someone on the team opens it on one student, chooses what the call is for and hands the phone over. By default it says up front that it is an AI; it writes the note and the agreed callback, and suggests a status for a counsellor to decide. xale.in puts it plainly: "It is a call with that person, not a list being dialled."
- Use it from the AI app you already have. Xale works as an MCP server (the Model Context Protocol, an open-source standard (opens in a new tab) for connecting AI applications to external systems), so Claude, ChatGPT, Gemini CLI, Cursor, VS Code or Codex can answer from the CRM "with the same access rules as inside Xale". They can add a note or update a status only when changes are switched on for that connection, and each connection can be revoked at any time.
Notice what it never does: show a lead the asker isn't allowed to see, save a rule by itself, or decide a student's status. That is the design to copy, whatever system you run.
Our own website is a smaller example, and its automations use no AI model at all. Every WhatsApp link to us on marketlube.in opens a chat whose first message says it came from our site and, where the page knows it, what the visitor was reading, so whoever answers starts with the context. Blog posts are written ahead and go live by themselves at 09:00 IST on their day, and a daily job then refreshes the pages that list them and tells Bing and other search engines through IndexNow. Where AI does help is in the writing: every post, this one included, is drafted with AI tools and researched, edited and approved by its named author, as our editorial policy explains.
How do you connect AI to the tools you already use?#
Most businesses don't need a new system to start. These are the AI features already inside tools that Kerala businesses run every day, as each vendor describes them:
The pattern repeats in every row: the AI drafts, sorts or answers, and a person keeps control of what goes into the books, the CRM or a customer's chat. Our guides to AI in CRM and AI in ERP compare the vendors' AI features and prices in more depth.
You are past built-in features when the AI has to work across tools: a WhatsApp enquiry checked against stock in the ERP and logged in the CRM, say. That is AI integration work, a project of its own, and our LLM integration guide explains the three patterns it can follow. If you want your own assistant on the WhatsApp Business Platform rather than Meta's agent in the app, read what Meta allows for AI chatbots on WhatsApp before anyone builds one.
When we build that kind of AI ourselves, the model is the last thing we add. An assistant that answers a call or a WhatsApp chat has done its job only when the enquiry sits as a lead with its context, the right person has it, and it shows in the next morning's report. So we build it on the platform behind Xale CRM (opens in a new tab), which already holds the leads, their follow-ups and call records, and the roles, automations and reports around them (reported by our team), and leave the AI one job: the conversation.
What should stay with a person?#
Anything that commits money, makes a promise or can't be undone. In practice:
- Money: payments, refunds, discounts and credit. AI can draft the entry or the reply; a person presses send.
- Anything legal or regulated: contracts, tax filings, GST and e-invoices, and claims about health or results.
- Decisions about people: hiring, rejecting an applicant, or changing a customer's status when it changes what happens to them next.
- Rules that run at scale: an automation that touches every lead needs a person to preview it before it's saved.
- Customers who ask for a person: the hand-off should be quick and carry the conversation with it.
The tools above already work this way, which is a good sign the pattern is right. Tally's AI drafts vouchers but posts nothing on its own. Xale AI's rule drafts wait for someone to press Save, and its telecaller leaves the status to the counsellor: "A counsellor decides what the student's status becomes," as xale.in's telecaller page puts it, and a student's "do not call" is recorded and honoured on the next call. Meta lets the owner take over any WhatsApp chat from its Business AI.
Keep a log as well. If nobody can see what the AI read and what it changed, nobody can fix it when it is wrong. Preview, then save, then log: if you take one rule from this guide, take that one.
What does it cost, and what rules apply?#
Starting inside tools you already pay for usually costs least, though each vendor's terms say what the AI adds to the bill. A build costs more. Innovatrix Infotech (opens in a new tab), an automation firm in Kolkata, publishes market ranges for n8n builds from ₹20,000 to ₹45,000 for a single simple workflow to ₹3 lakh to ₹10 lakh for a production AI agent, with retainers of ₹20,000 to ₹1.5 lakh a month for upkeep (updated 6 October 2026). Our AI automation cost guide sets those beside other sellers' prices and a person-month model, and the LLM guide works out what the model calls cost each month in rupees.
On the rules: India's Digital Personal Data Protection Rules were notified on 14 November 2025 with an eighteen-month period for phased compliance, and the Act's highest penalty, up to ₹250 crore, applies to failing to keep reasonable security safeguards (the government's backgrounder (opens in a new tab)). An AI tool that reads your customers' chats or leads is processing their personal data, so the business using it carries those duties. Our DPDP checklist for leads and CRM data turns them into steps, and the LLM guide covers where model providers process data.
How do you start?#
With one process and one number. Pick the task your team repeats most, write down how long it takes today and how often it goes wrong, switch on the AI feature in the tool that already holds the work, and keep a person approving every output for the first month. Then compare against the number you wrote down. Our 30-day pilot plan sets out the four weeks, from the baseline to the decision to expand, adjust or stop.
Frequently asked questions#
Can a small business use AI without a developer?#
Yes, for the features built into tools it already uses. Meta's Business AI is set up from the Tools tab of the WhatsApp Business app, the Sheets AI function is a formula, Tally's Docs by Ira is enabled on an eligible TallyPrime licence, and many CRMs switch their assistant on in settings. A developer becomes necessary when the AI must work across several systems, read live data under each user's permissions, or run at volumes where usage costs need watching.
Does AI work in Malayalam?#
More and more, but test it on your own data. Meta says its Business AI on WhatsApp is available in all native languages in India, and Xale AI speaks Malayalam by default, with call transcription tuned for calls that mix Malayalam and English. Support varies by tool and task: Google's help page says the Sheets AI function can be unavailable because of your language setting. Run twenty of your own chats or calls through any tool before you rely on it, and check names, places and amounts.
Is business data safe with AI tools?#
It depends on the plan and the set-up more than on the AI. Check where the data is processed, whether the vendor may use it to train models, and whether the AI answers inside the same permissions as the rest of the system. Under section 8(1) of the Digital Personal Data Protection Act, 2023 (opens in a new tab), the business stays responsible for personal data that a Data Processor, such as an AI vendor, processes on its behalf. Use business plans, give each person their own access, and keep customer data out of free consumer tools.
Where to go from here#
Write down the three jobs your team repeats most, and for each, name the tool that already holds the work and the person who will approve what the AI drafts. Start with one. If you'd like a second pair of eyes on where AI fits in your operations, talk it through with Althameem Ali, our co-founder, in a one-hour consultation, after which you get a written action plan within 24 hours. The answer may well be a setting in software you already pay for.
Changelog#
- 2026-10-12: Added a paragraph on how we build AI ourselves: as the top layer on the Xale platform, where leads, follow-ups, call records, roles, automations and reports already exist (reported by our team). One sentence we had quoted from xale.in is now in our own words, still credited to xale.in; the automation builder links xale.in's page on it; and the disclosure calls Xale the product our team built. The paragraph on how we build AI is now written for this guide, about AI that works across tools, with the same facts.
Related services
Written by
Co-founder, Marketlube
Althameem Ali researched, edited and approved this post. Marketlube is a growth partner in Calicut building brands, websites, CRMs and campaigns for businesses across Kerala, India and the Gulf.
