Like every other department, sales is getting an AI makeover, with sales teams using the tech to retrieve information about companies and their decision-makers, identify potential leads and communicate with customers via AI chatbots or, increasingly, voice agents.
Account executives tell me AI has sped up the whole process of handling a sales call and that sellers now have a series of connected apps at their fingertips to do a lot of their work for them.
Gong, for example, transcribes sales conversations and then uploads notes from the call to the company’s CRM automatically. With everything synchronised, every time a seller takes a call, the data on a customer account gets updated in real time. All a seller has to do is check everything is correct.
Startups are also using AI voice agents for customer communication. One tool, called Telli — a Y Combinator alum that raised $3.6m in pre-seed funding in April — helps automate part of the conversation sellers have with clients. Its AI agents (that can ‘speak’ in a variety of accents, such as English with a German twang) field calls from customers, and can direct transfers to a sales advisor, book appointments with them and follow up after a call.
But the company’s cofounder, Philipp Baumanns, is sceptical about whether AI can completely replace the human in the loop. “I do believe that the majority of very long consultations between a customer could and will still be human for a very long time, just because customers prefer that,” he says. “It’s something where the end customer value is the human connection.”
AI is also changing the composition of sales teams, and some experts predict that its automation of routine tasks will mean fewer jobs for junior salespeople. Sales teams, they think, will mostly be made up of experienced, senior people who focus on more complex negotiations, strategy and reeling in bigger deals.
Organisations will also have at least one person tasked with implementing and overseeing a new AI tool brought into a business; a lot of the companies Telli works with already run things this way. “You still need a lot of people to have the information context, to train systems and to also manage them long term,” says Baumanns.
It’s still early days though, and companies are still figuring out what works and what doesn’t, according to Johann Butting, partner at VC firm Visionaries Club and a former Dropbox and Slack executive. “People are still very unclear on best practices and are experimenting with it. There’s also a lot of nuance, depending on what kind of company you are.”
Visionaries hosted a go-to-market (GTM) day in London a few weeks ago focused on the role of AI in sales and customer success. Attendees included leaders from companies such as OpenAI, Anthropic, DeepL and Synthesia.
Butting says he found it surprising that many of the companies in attendance said they’re building their own AI tools on top of large language models (LLMs). One company he’s on the board of is building tools on top of Claude, ChatGPT and Gemini to help its business development representative (BDR) team identify targets to reach out to and prepare pitches.
“They find it easier to write these applications themselves than to pick the right solution, then configure and implement it correctly, and finally, trust it — and also optimise it. I’m hearing that a lot of the optimisation is harder in third party tools,” Butting tells me. “Having said that, over time, third party apps, as we better understand what the problems are, are going to become a lot more important.”
Other companies in attendance said that the primary AI tool they use is an enterprise search tool to find information — such as sales frameworks or rules for contact negotiation — across various documents, databases and applications. They can be built in house, used as part of an enterprise platform like Salesforce or purchased off the shelf (examples include Highspot or Confluence).
Butting says he also knows of companies beginning to experiment with and integrate third party tools like N8n to automate workflows in finance and legal — or Qualified to automate inbound lead flow or create live chat on their website.
So how can startups that aren’t already using AI in their sales process begin? And what should they keep in mind? I asked sales executives for their tips.
Identify the problem
Baumann suggests beginning with a clear understanding of what you’re trying to solve. For example, the top three problems Telli hears from companies are that they are not reaching enough customers, they don’t have the capacity to answer customer requests and they’re spending too much money on things like customer acquisition.
To identify the real bottleneck:
- Analyse your internal data. For example, look at the performance metrics (such as reachability rates or conversion rates) of your sales advisors to pinpoint weaknesses in your sales or support process.
- Benchmark against competitors. Compare the performance of specific aspects of your sales funnel with others in the industry to spot areas for optimisation.
Pilot small and iterate
Once you’ve identified the problem, decide what it is specifically you want to improve with AI — for example turning more leads into meeting bookings.
If your team currently books 30% of leads into appointments, use this as your internal benchmark.
- Run a controlled pilot. Launch a small-scale version of your AI tool targeting the same task (for example, making outbound appointment booking calls).
- Compare outcomes. Then, run it in parallel with your normal processes. If AI performs better, great. If it performs worse, consider a different use case.
Look for immediate value
Karine Martinez, head of sales at payment solutions provider Edenred Payment Solutions, advises companies to solve one clear bottleneck rather than juggling multiple new tools
- Choose tools that are easy to test and that integrate naturally with your existing processes.
- Focus on repetitive, low risk tasks where automation can have “the biggest benefits”.
“Train it on your own data when possible,” she says, “but always keep a human in the loop. We shouldn’t expect perfection right away, but as long as you track what’s working and keep iterating, you’ll find improvements along the way. It doesn’t need to be complicated to make a real difference.”
Educate your team
Butting emphasises that AI fluency across your team is essential, as the technology is going to rapidly change the job of sales people.
- Prioritise hiring candidates with experience with AI tools.
- “Orchestrate a learning process” to teach employees how to use AI. Tacto, an AI-powered supply chain and procurement platform, for example, brings together its entire team fortnightly to share knowledge and teach each other new applications with AI.
- Level up. Having knowledge of how AI works is not just essential for how sales teams work internally, it’s also necessary to understand and sell AI solutions, says Butting. The sellers of the future require technical understanding married with a deep understanding of businesses and their specific needs and challenges.
“As a salesperson, the most valuable thing you can do is understand the business of your customer and understand the person whom you’re dealing with, and to understand these business problems that are now on such a higher level,” he says. “That’s a new task for many sellers.”
🛠️ Useful third party AI sales tools, according to sales executives
Fluint — offers AI-powered content generation from business cases to account plans.
Clay — helps enrich data about leads and accounts, and automates prospecting.
N8n — automates workflows across CRMs, marketing platforms and analytics tools. It can help with things like automating lead management and enhancing customer engagement through personalised email campaigns.
Lavender — an AI email coach that helps personalise emails.
Gong — records, transcribes and analyses sales calls.
Scratchpad — helps with sales admin and keeps Salesforce data clean.
Glean — an enterprise search tool helping users find things in their company’s internal data.
Read the orginal article: https://sifted.eu/articles/how-to-sales-ai/
