Ask any sales head in Mumbai, Ludhiana or Coimbatore where the week went, and the answer is rarely “with clients”. It went on updating the CRM, rewriting the same proposal, chasing an internal approval for a discount and scrolling back through a WhatsApp thread to find what the client asked for on Tuesday. That is where AI for sales earns its place: not as a robot salesperson, but as an agent that takes the reading, sorting and drafting off your team’s plate.
The numbers back this up. In Salesforce’s State of Sales report, seventh edition, based on a survey of 4,050 sales professionals in 22 countries (India among them) run in August and September 2025, reps said they spend 40% of an average week selling and 60% on everything else, including prospecting, creating quotes and entering data by hand. It is a vendor’s survey, so read it as a signal rather than gospel, but few sales heads would argue with the direction.
Indian sales adds its own layers. Conversations run on WhatsApp, not email. A deal in manufacturing or education often has three or four decision makers, from the owner to the purchase head to the finance controller. Every quote needs the right GST treatment, and the follow-up after a site visit has to reach the right person in the right language. Below are 25 tasks an agent can draft this quarter, grouped by sales stage, each with who signs off and what they check.
AI for sales: how to read this list
Each task follows the same pattern. The agent drafts: it reads, sorts and writes a first version. You sign off: a named person approves before anything reaches a client. The check is the one thing the reviewer must verify every time. If a task has no obvious person to sign off, it is not ready to automate.
Gartner predicted in November 2025 that by 2028 AI agents will outnumber human sellers tenfold, yet fewer than 40% of sellers will report that agents improved their productivity. Its analyst Melissa Hilbert put it plainly: “Beyond a certain point, more AI does not mean more productivity.” The answer is a few well-chosen tasks, not more bots.
Prospecting and research
1. Prospect company summaries
Before a first call, the agent reads the prospect’s website, recent news, LinkedIn page and annual report (if there is one) and writes a one-page brief: what they make or sell, where they operate, recent expansions, likely pain points.
Agent drafts: a one-page brief with sources · You sign off: the salesperson · Check: facts are current and every claim links to a source.
2. Ideal client list building
Give the agent your best ten clients and it can describe the pattern (sector, size, city, trigger events) and draft a list of similar companies from public directories and trade fair exhibitor lists.
Agent drafts: a shortlist with the reason each company fits · You sign off: the sales head · Check: the fit logic matches how you actually win deals.
3. Trigger event monitoring
New factory, new campus, funding round, a fresh tender, a leadership change. The agent watches a list of target accounts and flags events worth a call, with a suggested opening line.
Agent drafts: a weekly digest of triggers per account · You sign off: the account owner · Check: the event is real and recent, not a recycled press release.
4. Decision-maker mapping
For a mid-sized manufacturer or a school group, the agent drafts a map of likely roles in the decision (owner, operations head, purchase, finance, IT) from public information and your past deals, with what each one usually cares about.
Agent drafts: a role map with gaps marked · You sign off: the salesperson · Check: nothing personal or invasive is collected, only business roles, in line with your DPDP Act obligations.
5. Lead list clean-up
Exhibition visitor lists, scanned visiting cards and old Excel sheets are full of duplicates, half numbers and “Mr. Sharma, Pune”. The agent standardises, de-duplicates and flags records that need a human look.
Agent drafts: a cleaned list and an exceptions sheet · You sign off: sales operations · Check: no records were merged wrongly and opt-in status is kept.
First contact and qualification
6. First-touch email and WhatsApp drafts
The agent writes a short, specific first message using the research brief: why you are reaching out, one relevant example, one question. It drafts in English, Hindi or Hinglish, as the prospect writes.
Agent drafts: two or three message variants · You sign off: the salesperson · Check: tone, no exaggerated claims, and the prospect has opted in to WhatsApp.
7. Inbound enquiry replies
When a lead arrives from your website or a Click-to-WhatsApp ad at 10 p.m., the agent replies within the WhatsApp service window, asks a few qualifying questions and books a slot. We cover the platform rules in our guide to WhatsApp automation for business in India.
Agent drafts: replies within a script you approved · You sign off: the sales head approves the script; a rep takes over at handover · Check: no prices or dates are promised by the agent.
8. Lead qualification notes
From the form, the chat and the first call, the agent scores the lead against your criteria (budget band, timeline, authority, need) and writes two lines on why.
Agent drafts: a score with reasons · You sign off: the salesperson · Check: the reasoning, not just the score; overrule it freely and note why.
9. Lead routing suggestions
Which rep should take this lead? The agent suggests one based on territory, language, sector experience and current load.
Agent drafts: a suggested owner · You sign off: the sales manager · Check: key accounts and existing relationships are not reassigned by mistake.
Human in the loop: the agent can draft the first hello, but the salesperson owns the relationship from the first reply. If a prospect asks “who am I speaking to?”, the honest answer should be a named person who will be on the next call.
Meetings
10. Pre-meeting briefs
Ten minutes before a meeting, the rep gets a single page: who is attending, what was discussed last time, open questions, the client’s last order or enquiry, and three things worth asking.
Agent drafts: the brief from CRM, email and chat history · You sign off: the salesperson · Check: the history is complete and the right client was matched.
11. Meeting notes and action items
From a recorded call (with consent) or the rep’s voice note in the car afterwards, the agent writes the summary, decisions, objections and next steps.
Agent drafts: notes and a task list · You sign off: the salesperson · Check: commitments are worded exactly as agreed, nothing added.
12. Objection libraries
The agent groups the objections your team hears (“your price is higher than the local vendor”, “we already have a supplier”) and drafts answers from what has worked before.
Agent drafts: an objection-and-answer sheet · You sign off: the sales head · Check: the answers are true and fit your positioning.
13. Demo and site-visit prep
For a factory demo, a show flat visit or a campus tour, the agent drafts the agenda, the checklist (samples, brochures, parking, gate pass) and the confirmation message to the client.
Agent drafts: agenda, checklist and confirmation · You sign off: the salesperson · Check: dates, addresses and attendee names.
Proposals and quotes
14. Proposal first drafts
The agent assembles a proposal from your templates, past winning proposals and the meeting notes: understanding of the need, scope, approach, timeline. Salesforce’s report lists creating quotes among the top uses of agents in sales.
Agent drafts: a full first draft in your format · You sign off: the salesperson, then the sales head for large deals · Check: scope matches what the client asked for, not what the template assumed.
15. Quotes with GST worked out
From your price list, the agent prepares line items, discounts within approved bands, HSN or SAC codes and GST (CGST and SGST, or IGST for interstate supply).
Agent drafts: a quote ready for review · You sign off: the salesperson and finance · Check: GSTIN, place of supply, tax rate and totals, every time.
16. Case study and reference matching
The agent picks the two most relevant case studies or references for this prospect by sector, city and problem, and drafts the paragraph that introduces them.
Agent drafts: a short list of matched examples · You sign off: the salesperson · Check: the reference client has agreed to be named.
17. Tender and RFP compliance checks
For government or large corporate tenders, the agent reads the RFP, lists every mandatory document and eligibility condition, and flags gaps in your draft response.
Agent drafts: a compliance checklist with gaps · You sign off: the bid owner · Check: deadlines, EMD amounts and format rules against the original document.
18. Proposal translations and summaries
A two-page summary in Hindi or Marathi for the owner, the full English proposal for procurement, both from one source.
Agent drafts: summary and translated versions · You sign off: a fluent team member · Check: numbers and commitments are identical across versions.
Human in the loop: no agent decides a price, a discount or a delivery date. It can prepare the quote and show the margin; a person with authority sends it, because the promise in that document is yours to keep.
Follow-up and closing
19. Follow-up sequences after a proposal
Day 1, confirm receipt and offer a walkthrough. Day 3, one useful thing. Day 7, a plain question about timing. The agent drafts each one in context and stops when the client replies.
Agent drafts: the sequence, using approved templates outside the WhatsApp window · You sign off: the salesperson · Check: the sequence pauses for holidays, festivals and live negotiations.
20. Stalled deal nudges
The agent flags deals with no activity for a set number of days and drafts a reason-specific nudge for the rep, not the client: “Purchase head went quiet after the revised quote; consider a call to the owner.”
Agent drafts: a stalled-deal list with suggested next steps · You sign off: the account owner · Check: the deal is genuinely stalled and not waiting on your side.
21. Payment link and advance reminders
Once a deal is agreed, the agent sends the order confirmation, the proforma and a UPI or bank payment link for the advance, then a polite reminder if needed.
Agent drafts: confirmation and reminder messages · You sign off: finance approves the list · Check: amounts match the signed quote; long-standing partners get a call, not a template.
22. Handover notes to delivery
The agent writes the handover from sales to delivery or operations: what was promised, by when, special terms, who to contact.
Agent drafts: a handover note from the deal record · You sign off: the salesperson and the delivery lead · Check: every promise in the proposal and in chat appears in the note.
CRM and reporting
23. CRM updates from email and WhatsApp
The most disliked sales task. The agent reads the day’s conversations and proposes updates to deal stage, next step, amount and contacts, in Zoho, HubSpot, Salesforce or a sheet.
Agent drafts: proposed field updates for one-click approval · You sign off: the salesperson · Check: deal stage and amount, the two fields reports rely on.
24. Weekly pipeline summaries
Every Friday, the agent writes a one-page pipeline note for the founder: new deals, movement, risks, and what is likely to close this month, in rupees.
Agent drafts: the weekly summary · You sign off: the sales head · Check: the forecast reflects judgement, not just CRM stage percentages.
25. Win and loss analysis
At the end of the quarter, the agent reads won and lost deals and drafts patterns: where you lose on price, which sectors close fastest, which objections recur.
Agent drafts: a pattern report with examples · You sign off: the founder or sales head · Check: the conclusions are supported by enough deals to mean something.
What the data says about adoption, and about data
The same Salesforce report found that 54% of sales professionals say their teams already use AI agents, and another 34% expect to within two years. It also found that 46% of sales professionals with agents say data quality issues hurt their sales, and 51% of sales professionals say security concerns delayed their AI initiatives.
That last pair matters more than the adoption figure. An agent reading a CRM full of duplicates and blank fields writes confident nonsense, so tasks 5 and 23 are dull but essential. And because your CRM and chats hold personal data, decide early what the agent may see and store under the DPDP Act, and ask your counsel to confirm the position for your case.
What to keep human
- The relationship. Who calls the owner, who visits the plant, who shows up when something goes wrong.
- Price and discount decisions. The agent shows the margin; a person decides.
- Commitments. Delivery dates, scope, warranties and payment terms are promises your company has to keep.
- Sensitive conversations. Complaints, disputes, a client who is upset or a deal that is falling apart.
- Reading the room. Knowing that the finance controller’s silence means a no, or that the owner’s nephew is the real decision maker.
How to start this quarter
Pick three tasks. Choose one from prospecting, one from proposals and one from CRM. Tasks 1, 14 and 23 are a sensible default for most B2B teams; real estate and education teams often start with 7 instead of 1.
Run a two-week pilot. Name the person who signs off each task, give the agent access only to what it needs, and review every output. Expect to correct it often in the first week; that is how the instructions improve.
Measure before and after. Write down the baseline first: hours per rep per week on the task, time to first reply, days from meeting to proposal, CRM fields left blank. Compare after two weeks. If a task does not move a number, drop it and try the next one. We explain why this discipline separates the pilots that pay back from the ones that stall in why AI projects fail.
At Cannyworx, AI does the heavy lifting on research, drafting and data entry, and a senior person on our team designs the workflow and approval screens with yours, so every output has a named human who signs off. If you are unsure whether you need a simple assistant or a proper agent, our guide to chatbots, agents and agentic applications explains the difference.
If you want all 25 tasks on one printable page, with sign-off roles and risk levels, the free “25 sales tasks” guide below has it. And if you would rather talk through which three to start with, book a free consultation.
Questions people ask
What is the best first use of AI for sales in an Indian company?
Usually the work that eats a rep's evening: meeting notes into the CRM, follow-up drafts after a call, and proposal first drafts. They repeat every day, they are easy to check, and a mistake is caught before a client sees it.
Can an AI agent message prospects on WhatsApp by itself?
It can, but we would not let it make commitments. Keep the agent to replies inside the 24-hour customer service window and approved templates outside it, only for people who have opted in, and route anything about price, discounts or delivery dates to a person.
Will AI replace our sales team?
No. In complex B2B, real estate or education sales, the relationship, the judgement on price and the promise to the client stay with people. The agent takes the reading, sorting, typing and drafting, so your salespeople have more hours for conversations.
How do we measure whether an AI sales pilot worked?
Take a baseline before you start: hours per week on the task, time to first reply, days from meeting to proposal, CRM fields left blank. Run the pilot for two weeks with a person reviewing every output, then compare the same numbers.
