AI in sales and marketing: a use-case map for B2B and B2C

In short: AI in sales and marketing is best at repetitive, text-heavy work: drafting content and outreach emails, preparing draft quotes from the price list, reading tender documents, summarising calls and updating the CRM. For B2B and project-based sales, quoting and tender preparation are usually the best places to start, because that is where salespeople lose the most hours. Anything AI drafts still needs a human check before it reaches a customer.

Most writing about AI marketing is about retail: ad copy, product photos, chatbots on social pages. Companies that sell to other companies face a different problem: long sales cycles, several decision makers, complex quotes and tender packs dozens of pages long. Walking through the sales funnel step by step shows where AI helps in both models.

What is AI in marketing?

AI in marketing means using artificial intelligence to understand customers, create content, personalise messages and measure campaigns faster than by hand. In sales, AI handles the paperwork around selling: looking up prices, drafting quotes, taking notes, sending reminders and suggesting who to call first.

Two different technologies are involved. Generative AI writes and summarises: emails, posts, draft quotes, call notes. Machine learning predicts from numbers: which lead is likely to buy, which deal is at risk, which product to suggest alongside another. Smaller companies usually start with generative AI because it needs little historical data.

AI across the sales funnel

Every stage of the funnel has a job AI can do, a type of data it needs and a person who must check the result.

Stage	What AI does	Data needed	Who checks
Attract	Drafts posts, video scripts, ad variants, topic ideas	Product facts, brand voice, customer profile	Content owner
Nurture	Writes email sequences per segment, summarises sign-up needs	Opted-in contact list, interaction history	Marketing
Advise and quote	Finds matching products, drafts quotes, reads tender documents	Price list, product codes, past quotes, discount policy	Salesperson, sales manager
Close	Suggests terms, flags stalled deals, prepares meetings	CRM data, communication history	Head of sales
After sale	Sorts feedback, suggests upsells, answers common questions	Orders, feedback, product documentation	Customer service

The first two stages work with off-the-shelf tools. The last three need AI to read real company data, so they take more preparation and save more hours.

AI in marketing: where it actually helps

AI is most useful for producing content and analysing feedback, and least useful as a substitute for strategy. It cannot decide which customers you should target; your marketing lead has to.

Content and ads

AI drafts blog posts, social posts, video scripts and many headline variants for testing. Drafts improve sharply when the input is good: real product specs, the questions customers actually ask, a few past pieces as a style sample. Someone still has to check every number and promise, because AI tends to invent product features.

Customer and competitor research

AI can group hundreds of comments, reviews and call notes into recurring themes. For competitors, it reads product pages, public price lists and brochures to build a first comparison table. Few teams ever had time to do this regularly.

Personalised email

Instead of one email to everyone, AI drafts messages by industry, company size or need. Only send to people who have agreed to receive them. In Vietnam, the Personal Data Protection Law (No. 91/2025/QH15), in force since 1 January 2026, sets clearer rules for using personal data in marketing.

Campaign reporting

AI can answer questions such as “which channel brought in the most quote requests this month?” once ad, website and CRM data sit in one place. If your numbers are still scattered, see building a data-driven business first.

AI in B2B and project-based sales

B2B and project sales are where AI saves the most time, because salespeople spend much of their week on paperwork rather than with customers.

Faster quotes from the price list and past quotes

A salesperson describes the customer's need in plain words. AI finds matching items in the catalogue, applies prices and discounts according to policy, pulls similar past quotes for reference and drafts the quote. The salesperson reviews and adjusts it before sending. One condition is non-negotiable: there must be a single, correct price list. AI reading three different spreadsheets will give three different prices.

Tender documents

Tender packs are long, with technical and commercial conditions spread across many sections. AI reads the full pack, lists every requirement, flags unfavourable clauses and drafts the technical narrative from your existing capability documents. The bid owner must still check each requirement against the original, because missing one condition can disqualify the bid.

Call notes and CRM updates

Salespeople dislike data entry, so CRMs (customer relationship management systems) are often incomplete. AI summarises a call or meeting recording, extracts need, budget, decision maker and next step, and proposes a CRM update the salesperson confirms with one click. Better CRM data makes every later analysis more reliable.

Prioritising leads and chasing stalled deals

With enough CRM history, AI suggests who to call first, which deals have gone quiet and which customers are due to reorder. An AI agent can run the whole chain: check the CRM, draft a reminder email, create a task for the salesperson. See What is an AI agent? for how these assistants work.

How B2B and B2C use AI differently

B2C uses AI mainly to create lots of content and answer many customers at once. B2B uses it to handle fewer customers whose files are far more complex.

	B2C	B2B and project sales
Volume	Many customers, small orders	Few customers, large contracts
Where AI helps most	Content, ads, customer chatbots	Quotes, tenders, CRM notes
Most important data	Purchase behaviour, order history	Price list, past quotes, technical documents, CRM
Cost of an AI mistake	One complaint, one bad post	Wrong price, missed tender condition, wrong commitment
Review needed	Spot checks, approval before publishing	Approval of every quote and tender before sending

For automated customer replies in B2C, see AI customer service chatbots.

Choosing AI marketing software

Choose by the job and by where your data lives, not by a list of trending tools.

Group	Used for	Watch out for
General AI assistants on business plans	Content, email, summaries, research	Clear rules on which data may go in
AI features inside CRM and marketing automation	Lead scoring, email suggestions, deal summaries	Only as good as the CRM data
Specialist tools (design, video, SEO)	Images, video, content optimisation	Image rights and labelling of AI content
Custom solution linked to price list, CRM, inventory	Quoting, tenders, sales assistant	Needs a data owner and more design work

If you have no CRM, or nobody updates it, fix that before buying more AI tools.

Risks to control

The biggest risk is AI saying something wrong to a customer on your behalf. A quote with the wrong price or an ad promising a feature you don't have becomes a commitment.

Set three rules from day one. No AI-drafted content goes out without human approval. No customer data goes into free AI tools. And meet the transparency duties: Vietnam's Law on Artificial Intelligence (No. 134/2025/QH15) took effect on 1 March 2026, and its Article 11 requires clear disclosure when publishing AI-generated content that could mislead people about what is real, plus clear labels on content that imitates a real person's face or voice. Have legal review any ad that uses a synthetic face or voice. The wider framework is in AI governance and data security.

Where to start

Start with the task that eats most of your sales team's time, not with the tool everyone is talking about.

1. Measure the team's time. For two weeks, each salesperson logs time spent on quotes, tender documents, CRM entry, email and customer meetings.
2. Clean the data for the chosen task. For quoting, settle on one price list and one product catalogue. For content, write a one-page brand voice guide and a list of verified product facts.
3. Pilot with two or three people. Include your best salesperson and your newest. Record time before and after, how many quotes needed corrections, and the typical AI errors.
4. Write the approval rules, then scale. Define who approves what, and above which value the head of sales must sign off.

Measure results with numbers the sales team already cares about: time from request to quote sent, share of quotes corrected after review, tenders submitted per quarter, share of CRM deals with complete mandatory fields, and quote-to-order conversion compared with the same period last year.

Common mistakes and when to wait

The most common mistake is using AI to mass-produce similar content and then wondering why leads don't grow. AI content without your own facts reads exactly like your competitors'. The second is letting AI send bulk email to lists of unclear origin, which hurts your domain reputation and creates personal-data risk.

Hold off on AI for sales if you don't yet have a single price list, if every salesperson quotes differently, or if the team is a handful of people sending a few quotes a month. Standardising the quoting process in a spreadsheet or a simple CRM will pay off sooner.

FAQ

What is AI in marketing? Using artificial intelligence to create content, understand customers, personalise messages and measure campaigns. People still set the strategy and approve the content.

Where should a B2B company start using AI in sales? With quoting or CRM notes, because both are time-consuming, repetitive and measurable. The prerequisite is a single, agreed price list and product catalogue.

Can AI write tender documents? AI reads tender packs, lists requirements and drafts the technical narrative quickly. The bid owner must still check every requirement against the original and remains responsible for what is submitted.

Does Google penalise AI-written marketing content? According to Google Search Central (2023), Google rewards quality regardless of how content is produced, but mass-producing pages to manipulate rankings breaks its spam policies. Generic AI content with nothing original rarely ranks well.

Do AI-made video ads need a label? Content that imitates a real person's face or voice, or could be mistaken for real, must be labelled or disclosed under Article 11 of Vietnam's AI law. Have legal review it before publishing.

---

HAT Solutions helps B2B sales teams cut the time spent on quotes and tender documents with AI connected to their price list, CRM and existing documents. Send us a sample quote and your current process, and we will point out which step to automate first. Contact us for advice, or read the overview AI for business.