Sales & AI

AI in Sales: Realistic Uses for SMEs

What AI in sales realistically does today

AI in sales is not an autonomous seller but an assistant for the routine around selling. Language models (LLMs) and CRM automation take over writing, searching and sorting tasks that cost time and need little finesse. Value appears where repetition, text work and data upkeep dominate - not in building relationships or deciding a close.

  • Well suited: drafts, summaries, prioritization, data cleanup, research.
  • Not suited: negotiation, building trust, final pricing and closing decisions.
  • Rule of thumb: AI proposes, the human reviews and owns the result.

Lead qualification: priorities instead of gut feeling

AI can enrich incoming inquiries and sort them by fit: it pulls together public company information, matches it against your ideal customer profile and proposes an order for handling. Promising contacts reach your desk first, instead of in order of arrival. Important: scoring criteria must stay explainable and must not become an opaque automatic rejection of people.

  • Use signals: industry, size, region, concrete need in the inquiry text.
  • Mind the limit: automated individual decisions trigger revDSG requirements.

Follow-ups and email drafts in minutes

The biggest time sink in sales is following up. From a short instruction and the conversation context, AI produces personalized drafts - including a fitting subject line, tone and next step. It can deliver variants in several languages, which is especially useful in multilingual Switzerland. The salesperson reviews, corrects facts and sends. Never send unchecked: models can invent details or miss the tone.

Meeting notes and conversation summaries

From transcripts of video or phone calls, AI creates structured notes: key points, objections, commitments and open to-dos. This unburdens memory and enables cleaner handovers in the team. If you record calls, in Switzerland you need the consent of everyone involved - recording without consent is not allowed. Ideally the summary is filed directly as an activity in the CRM.

Proposal drafts and CRM hygiene

For proposals, AI builds a first draft from blocks and meeting notes: scope description, value arguments and structure. Prices, discounts and legal clauses, however, belong in maintained templates, not the model - invented numbers are a real risk. For CRM hygiene, AI helps detect duplicates, complete fields and flag stale records, so the pipeline stays reliable.

  • Always validate against approved pricing and contract templates.
  • Run CRM cleanup in batches, spot-check the results.

Data protection, rollout and the Swiss framework

Customer data is personal data. Anyone feeding it into AI tools must comply with the revised FADP (revDSG): purpose limitation, transparency and a record of processing activities. With providers outside Switzerland you need appropriate safeguards for data transfer; the supervisory authority is the EDOEB. In practice: share only necessary data, clarify responsibilities and start with a clearly scoped pilot whose benefit can be measured.

Frequently asked questions

Does AI in sales replace my salespeople?

No. AI takes over routine such as drafts, summaries and data upkeep. Building relationships, negotiating and deciding a close remain human tasks. Realistically, teams gain time they can invest in customer conversations.

Which sales task should I automate first?

Start where repetition and text work are high and mistakes stay easy to correct - usually follow-up drafts or meeting summaries. These cases deliver quick, measurable value at low risk.

May I enter customer data into an AI tool?

Only in compliance with the revised FADP (revDSG): purpose limitation, transparency and data minimization. Providers outside Switzerland require appropriate safeguards. Clarify the legal basis beforehand; the supervisory authority is the EDOEB.

How do I prevent false statements in AI proposals?

Always source prices, discounts and legal clauses from maintained templates, not the model. Use AI only for structure and value arguments, and check every draft against the source before it reaches the customer.

Do I need consent to record conversations?

Yes. In Switzerland, recording conversations without the consent of all participants is not allowed. Obtain consent at the start, or work from manual notes that AI then structures.

What does it cost an SME to get started?

Prices change constantly, so compare providers directly. A small pilot with one clear use case makes sense: measure the time saved in hours before deciding in CHF on a broader rollout.

Key terms in the glossary

← Back to overview

Practical AI for your business

From idea to implementation – we show you what is concretely possible in your case.

Request a demo