AI in sales and marketing: what it changes for your team

AI helps your sales and marketing team with four things: working through scenarios, prioritising leads and opportunities, preparing quotations and drafting content. It is only as good as the customer data your team records. Have a person review every proposal before it reaches a customer.

In previous blogs, Erry de Boer, CEO of Eqeep, shared his view on the impact of AI in finance and operations. This time it is sales and marketing, and Erry starts with the quarterly target.

Do you recognise these challenges?

Sometimes sales lag behind at the end of a quarter, or stock has to move faster. You may then give up some margin for higher volume.

Beyond Erry’s example, these may sound familiar: account managers set each deal’s probability by gut feel. The same customer appears as lead, prospect and customer. You do not know why you lose quotations, or to whom. Marketers paste customer lists into ChatGPT.

How can AI help you still hit your quarterly target?

By working through scenarios. Take the end of March. Your team is still a million euros behind the quarterly target, and that revenue has to come in.

You know the traditional approach. You gather the team, cover the wall in sticky notes and brainstorm. It takes a lot of time and energy, with an uncertain outcome.

An AI system calculates the scenarios instead. What happens if you offer a 10% discount on product A, or 12% on product X? Perhaps a bundle works better. You get an advance estimate of what each action delivers, without weeks of testing first.

Can AI also choose the best channel for your campaign?

AI can weigh up the channel and estimate the effect in advance. Perhaps a LinkedIn campaign works best for your audience, because that is where most responses come from. In other cases, Facebook might be more effective. This action could then bring in, for example, an extra 1.2 million euros in revenue.

AI also puts its own ideas on the table, complete with calculations and expected results.

We add one caveat here. The two sales assistants IFS describes for release 26R1 do not calculate discount or channel scenarios. Many teams choose channels and write copy in other tools. This is the least IFS-specific use of AI in this series.

What AI for sales is in IFS Cloud?

Two assistants that answer from your own CRM and quotation history. According to the IFS documentation for release 26R1, IFS.ai Copilot for Sales Representatives works on the Business Opportunity, Business Activity and Account pages. Your account managers can ask it for, among other things:

• customers they have not contacted for a long time;
• deals in the final stages that can be closed;
• the top five opportunities this month, quarter or year, weighted by value and probability;
• the competitors they most often lose similar opportunities to;
• opportunities that contain inconsistencies, with the reason for each.

The second assistant works inside the sales quotation. For a customer, it shows the top five reasons quotation lines were lost in the last six months. For a sales part, it shows which competitor products appear most often on lost lines. It also shows the sales parts with the highest win rate and the parts customers most often order together.

That is the bundle from Erry’s example, based on your own order history. The sales coordinator can add a suggested part to the quotation straight from the Copilot panel. We check with you which functions are available in your IFS version. What needs to be in place before you use AI is covered in our article on becoming AI-ready.

Why is the AI only as good as your CRM?

Because it counts what your team records when a deal is won or lost. Reasons for lost quotations come from quotation lines marked Lost. When an opportunity is lost, the sales rep can record which competitor won and why. If those fields stay empty, the assistant has nothing to count.

The same goes for each deal’s probability. In IFS Cloud, each stage of your CRM process can carry a default probability, which the sales rep can change or you can force. If probability simply follows the stage, Copilot’s top five is mostly a list of your biggest late-stage deals. The ranking is only as good as your CRM template and your team’s discipline.

You decide which steps are mandatory. With validation type Stop, a rep cannot set a stage to done while mandatory actions are open. The duplicate check shows whether a new lead already exists as a lead or account, and you merge the duplicates. A qualified lead converts into a prospect or customer, so your team works from one customer record.

Data lost between departments is the subject of our article on inefficient processes.

Can you put customer data into an AI chatbot?

Not without rules, because even in B2B much of it is personal data. According to the Dutch Data Protection Authority (AP), data about organisations fall outside the General Data Protection Regulation (GDPR). A business email address used by one specific employee, however, is personal data. So is information about what someone does or buys online, such as the click behaviour you use to score leads.

In August 2024 the AP warned about AI chatbots after several breach notifications. In one, a telecoms employee had entered a file with customer addresses. If staff enter personal data against company agreements, it is a data breach. If it is company policy, it is often not permitted by law.

Most companies behind chatbots store everything that is entered, the AP adds. So set out which data your marketers and sales reps may and may not enter. The AP also suggests agreeing with the chatbot provider that it does not store the data. If something does leak, notifying the AP and the people affected is mandatory in many cases.

According to IFS, IFS.ai Copilot only gives insights from data the user is authorised to access. With each question, the conversation history and up to five selected records per list are shared with the Machine Learning Service. Function calls can also add data from the database that is not visible on the page. Discuss that with your data protection lead first.

One rule affects marketing directly. If a contact objects to direct marketing, you may no longer use their data for it, and you may not ask why. Make sure the objection also reaches the lists AI uses to build audiences.

Who reviews what AI proposes?

Always a person, and you decide in advance who. IFS itself says the quotation assistant’s responses should be verified before you act, as they can occasionally be incorrect. So the sales coordinator checks the quotation before it goes out. The commercial director decides on discount and margin, even when AI has run the scenario.

The same applies to content. If your marketers use AI to write adverts or emails, they need to know that AI sometimes makes things up. Under the EU AI Act, making sure they know this is a duty for your organisation.

Where do you start with AI in sales and marketing?

With your data, just as in finance and operations. Erry’s advice is to record sales results, customer behaviour and campaign effectiveness carefully.

In IFS Cloud, start with a reason and competitor on every lost quotation line, a deliberate probability per CRM stage and one record per customer. Next, pick one question your team asks every week, such as why you lose quotations. Act on the answer, then measure the win rate for those parts or customers over the next quarter. Our guide to KPIs and metrics covers the measures that help.

Questions about AI in sales and marketing

What AI functions for sales does IFS Cloud have?

In release 26R1, IFS documents IFS.ai Copilot for Sales Representatives for opportunities, activities and accounts, and IFS.ai Copilot in Sales Quotation for quotations. Among other things, they show the top opportunities, reasons for lost quotations, win rates per sales part and parts that customers order together.

Is a business contact’s email address personal data?

Yes, if one specific employee uses that address. According to the AP, information about an organisation only falls under the GDPR when it also says something about a person. For a small company, business information can also say something about the owner.

Who approves a discount that AI proposes?

A person you name in advance. The commercial director decides on discount and margin, even when AI ran the scenario.

Which sales question would you answer with AI first?

Tell us where you stand with IFS. After one conversation, you have a clear next step. Bring the question that takes your sales team the most time today. Discuss your IFS question.