Published on the 15/03/2023 | Written by Heather Wright

Are you ready to let AI write your sales proposals or respond to customer complaints?

A wave of big name tech vendors are banking on it, with Salesforce, Microsoft and Hubspot all jumping on the generative AI bandwagon. And while ChatGPT may have taken the world by storm largely as a bit of fun at least in its early days, these vendors are counting on you being willing to add generative AI to your CRM – and highly sensitive customer data.

Last week, Microsoft – which was partly behind the current generative AI hype with its work with and funding of OpenAI – launched generative AI capabilities across its CRM and ERP systems with Microsoft Dynamics 365 Copilot.

“The world is experiencing one of the most profound technological shifts.”

Microsoft is claiming it’s ‘the world’s first copilot in both CRM and ERP that brings next-generation AI to every line of business’.

The promise, according to Microsoft, is that the AI will help write email responses to customers and can create an email summary of a Teams meeting in Outlook, pulling in details from the company’s CRM, such as product and pricing information, along with insights from the recorded Teams call.

On the customer service front Microsoft says Copilot in Dynamics 365 Customer Service will draft contextual answers to queries in both chat and email, as well as providing an interactive chat experience over knowledge bases and case history.

A day later, Salesforce – which was an early adopter of AI with Einstein –  launched what it claimed was the first generative AI CRM technology, Einstein GPT. It uses generative AI technology and OpenAI’s advanced AI models which power ChatGPT to handle tasks including generating personalised emails for sales teams to send to customers or generating specific responses for customer service professionals to quickly answer customer questions.

Creating marketing campaigns targeted at specific customers, and auto-generating code for developers is also possible.  

Customers will be able to connect real-time data from Salesforce language models from Open AI out of the box, or choose their own external mode, and use natural language prompts within Salesforce CRM to generate the content, which adapts to changing customer information and needs in real-time.

The company intends to open up to a wide ecosystem of AI partners, with OpenAI just the first cab off the rank.

ChatGPT is also being integrated into the Slack instant messaging platform which Salesforce acquired in 2021. It will provide conversation summaries (sound familiar?), research tools and writing assistance for drafting messages.

Einstein GPT for CRM applications will be launched within the Slack interface, Salesforce says.

It hasn’t revealed when the features will be live.

The company’s venture capital arm also unveiled a US$250 million generative AI fund, saying it is an important area, given a Salesforce survey shows 67 percent of senior IT leaders will be prioritising generative AI over the next 18 months, with a third naming it as a top priority.

One of its first investments is in OpenAI rival Anthropic.

Hubspot, too is getting in mix, announcing a private beta release of Content Assistant, along with the alpha release of ChatSpot.ai – both powered by OpenAi’s ChatGPT.

ChatSpot.ai will provide a natural language chat-based user experience for Hubspot customers, enabling them to add contacts and companies to the CRM platform, create reports and draft emails.

Content Assistant is designed to help marketing and sales teams create content such as emails, websites, blog posts and landing page, along with streamlining marketing workflows in one place.

While Microsoft and Salesforce are both claiming firsts they, and Hubspot, won’t be the last: Pegasystems has already revealed plans for generative AI-supported tools in its low-code CRM platform, Pega Infinity, and reply.io, a platform which automates business communications, is also  introducing an AI-powered B2B ‘conversational assistant’, in Jason AI – again, built on ChatGPT.

Marc Benioff, Salesforce CEO, says “The world is experiencing one of the most profound technological shifts with the rise of real-time technologies and generative AI. This comes at a pivotal moment as every company is focused on connecting with their customers in more intelligent, automated and personalised ways.”

AI software spending is forecast to nearly double to US$64 billion between 2021 and 2025 according to Forrester. OpenAI, which reportedly made less than $10 million last year, has projected revenues of $200 million for this year and $1 billion by 2024. In January, just two months after launch, OpenAI’s ChatGPT was estimated to have reached 100 million active users, making it the fastest-growing consumer application in history, according to a study from Swiss bank UBS.

But behind the hype – and the potential for real benefits – lie plenty of risks still to be navigated.

Analysts have warned that as vendors move generative AI into areas which are home to customer data – something businesses are rightly very sensitive about – it will be crucial to be transparent on the data used to train the AI and on what is being done to make it secure and compliant.

ChatGPT, in its consumer use, has been the source of plenty of humour with its ‘hallucinations’ flooding social media. But while they raise a chuckle on social media, they’re less likely to do so in business scenarios, and close supervision of the technology will be required to avoid embarrassing racist, sexist and insulting language or inaccuracies, will be a necessity.

The use of AI tools to create content has also been criticised. CNet paused publishing of AI-written stories earlier this year after it came under fire for using the tools to write stories without clearly disclosing the use of the tools to readers. It was also forced to issue corrections – some of which were described as ‘substantial – on a number of the AI-written articles.

In a blog post, IDC noted of generative AI in general for business that, while providing lower-cost, higher value solutions, there are ‘significant ethical and perhaps legal implications’ for businesses using the technology.

“Businesses should choose models where techniques such as adversarial input (training against bad or manipulated data), benchmark dataset training (checking for biases via label tests) and explainable AI are used.”

Both Salesforce and Microsoft have said their systems were built on ‘ethical’ best practice.