2023 was a break-out year for generative AI technology, as tools such as ChatGPT graduated from lab curiosity to household name. But CIOs are still cautiously evaluating how to safely deploy generative AI in the enterprise, and what guard-rails to put around it. Sometimes, though, it sneaks in through the back door as a result of ad-hoc individual or departmental initiatives — or even through the front door, bundled by the vendors of enterprise applications already in widespread use.
With major enterprise software vendors adding new generative AI features and support for new large language models (LLM) almost monthly, CIO.com offers this round-up of the latest announcements to help IT leaders keep tabs on their exposure to this new technology.
December 2023
Salesforce adds vector capabilities to support search in Einstein Copilot
Salesforce has updated the Einstein 1 platform it introduced in September to add new search capabilities in Einstein Copilot, and a new vector database to support them. Vector databases store mathematical representations of different attributes of data, making it easier to find related items even when they’re described in different terms.
IBM leads alliance to build open-source generative AI
IBM, Meta, AMD, and around 50 other organizations have created the AI Alliance to open up innovation in generative AI. Their goal is to head off the creation of a monoculture of GPT-on-Nvidia-at-hyperscaler applications by encouraging collaboration, information sharing, and benchmark creation around development of alternative LLMs, hardware, and software. For enterprises and their software suppliers, it could mean a greater choice of generative AI building blocks to incorporate into the applications they build.
Microsoft Copilot will soon get a turbo boost
Microsoft’s Copilot virtual assistants are set to get a boost from OpenAI’s GPT-4 Turbo model in early 2024. This model has been trained on a more recent corpus of information and can accept much longer prompts — the equivalent of 300 pages of text — to give its responses more context.
Q takes Amazon into enterprise software territory
Q, Amazon’s generative AI chatbot platform, could turn the company into a vendor of productivity apps. That, at least, is some analysts’ take on what AWS demonstrated at its re:Invent conference in November. AWS says that Q’s responses can be fine-tuned by connecting it to data held in enterprise systems. However, it may not be ready for prime time: reports are already surfacing of Q hallucinating and allegedly leaking confidential data.
November 2023
Snow covers ITAM with gen-AI
Snow Software has applied generative AI to IT asset management, building a digital assistant inevitably called Snow Copilot to help enterprises manage their data through an interactive chat. The company expects the tool to reduce the need for exporting data into PowerBI or Excel in order to gain insights.
SAP unveils tools to help enterprises build their own gen-AI apps
SAP has built a “large business model” — like an LLM, but trained on business transactions — and introduced a suite of new and existing developer tools, SAP Build Code, to help enterprises build AI into business applications running on its software platform. SAP Build Code will use the company’s Joule AI assistant, introduced in September, as a coding copilot.
Microsoft launches Copilot Studio for building apps on Microsoft 365
Copilot Studio, now available in a free trial, is a SaaS tool for building and publishing plugins for the Microsoft 365 Copilot, enabling developers to customize the online productivity suite for their own enterprise’s needs. There are connectors to enterprise applications from other vendors, including SAP, Workday and ServiceNow.
October 2023
NetSuite adds autocomplete function to its ERP suite
Oracle has added a generative AI autocomplete function to text fields across its NetSuite ERP suite. Starting with a few words to describe the user’s intent, NetSuite Text Enhance will generate content for any text area before inviting the user to review, edit and approve it.
iCIMS applies gen-AI to hiring
Talent management software vendor iCIMS has used AI to filter resumes and match applicants with jobs for a while now, but with generative AI it’s tackling the next step: drawing on descriptions of job openings to create interview questions, with follow-up questions based on candidates’ resumes.
Boomi integrates GPT with its integration platform
Boomi is using GPT to add a conversational interface to its multi-cloud app integration platform. It says this will enable users to build outlines of integrations, APIs, and master data models using natural language, and then accept or modify the software’s proposals.
September 2023
Workday promises generative AI will help with finance and HR busywork
Generating job descriptions, analyzing and correcting contracts, and writing knowledge management articles are among the tasks Workday is targeting with its generative AI move. It expects to simplify and speed up these, and a host of other functions, with the new technology in the next year, but isn’t ready to release anything yet.
SAP readies generative AI chatbot Joule
SAP is taking a slow but steady approach with its generative AI roll-out. It’s working on a new chatbot, Joule, that will respond to natural-language requests with relevant enterprise information, summaries of interactions, and guidance on using SAP apps. Although SAP intends to add it to most of its applications in time, only the company’s SaaS HCM tool SuccessFactors and its SAP Start homepage for cloud apps will get Joule this year. The public cloud edition of its S/4HANA application suite is likely to follow early in 2024.
ServiceNow’s new Vancouver platform release goes live with gen-AI
ServiceNow has developed its own domain-specific LLMs to support the three major workflows that run on its Now Platform: IT Service Management, Customer Service Management, and HR Service Delivery. The generative AI features that the models enable are now available to users of the Vancouver release who are willing to pay for the “Professional Plus” or “Enterprise Plus” add-on packs.
Automation Anywhere expands its automation platform with gen-AI
Automation Anywhere is using generative AI to help with process discovery, thereby speeding up the end-to-end development of automations. It’s tuned the LLM it uses with anonymized metadata from millions of automations developed on its platform. It’s also used generative AI to build Automation Co-Pilot, which it says can help automate use cases, such as email triage or anti-money-laundering that couldn’t be automated previously.
Planview will add Copilot to its project portfolio management tool by year-end
Planview Copilot, planned for release before year-end, is a chatbot for interrogating project planning data held in the Planview platform. Planview’s idea is that it’ll identify bottlenecks and offer solutions to project planning problems in natural language.
Torii applies generative AI to SaaS management
To enhance its SaaS management platform, Torii is turning to generative AI to speed up the ingestion of SaaS contracts and automate data entry to identify SaaS applications. It’s also adding a natural-language interface to interrogate the data it holds, accessible through Slack.
Oracle will add gen-AI to its customer experience cloud apps
Oracle’s Advertising and Customer Experience Cloud (Fusion Cloud CX) is scheduled for a generative AI update sometime in 2024. Oracle said it’ll use the technology to help customer service agents find relevant knowledge articles, respond to service requests, and summarize their exchanges with customers, among other features.
Pega uses gen-AI to deliver next-best-action recommendations to its low-code platform
Pega Infinity 23, the latest release of Pega’s low-code platform, uses generative AI in two areas: enhancement to its training and documentation, enabling it to answer product support questions on the fly; and its creation of next-best-action recommendations, where it uses the technology to adapt the tone and the perspective of the recommendations.
Salesforce’s Einstein Copilot is almost ready for take-off
A generative AI chatbot on every page is what Salesforce is planning as it prepares to roll out Einstein Copilot to some customers before the end of the year. All its enterprise applications will soon incorporate an open-ended digital assistant able to summarize customer information and trigger Salesforce workflows in response to natural-language requests. Underlying these capabilities is Einstein 1, an update to Salesforce’s Data Cloud with an improved metadata framework.
August 2023
Salesforce sets out to train generative AI models with Einstein Studio
With Einstein Studio, Salesforce offers to help enterprises train and run generative AI models on their customer data stored in its Data Cloud. It can connect the data to a variety of LLMs, open source, or proprietary, the company said.
July 2023
Microsoft adds Copilot abilities to Dynamics 365 suite
Microsoft will roll out its Copilot generative AI assistant across more of its products. Already it can help analyze Excel data, create PowerPoint presentations, and write code. Now it’s added a Sales Copilot to its Dynamics 365 to help sales staff create email pitches based on customer data held in the ERP suite. Microsoft hopes this and other generative AI features will encourage the 90% of Dynamics customers that still run it on premises to move to the cloud.
Splunk adopts generative AI to help with observability
Splunk already uses AI in its suite of tools to analyze machine-generated data. Now it’s adding a generative AI assistant built with Google’s text-to-text transfer transformer model, T5. Splunk says it may take a little work to get good answers from the preview version as it’s looking for help from customers to improve the model’s training.
ServiceNow expands Now Assist virtual assistant to help coders
Now Assist will soon gain new generative AI features to help customers make greater use of the Now Platform. A case summarization tool will spare people in IT, HR, and customer service the hassle of writing up their interactions with clients, while a text-to-code function will make it easier to automate new processes using natural language to describe how they should work. Both additions use LLMs developed in-house by ServiceNow.
June 2023
Oracle’s generative AI investment surfaces in HCM tools
Oracle is taking a three-pronged approach to generative AI adoption. At the hardware level, it’s adding new instances to its Oracle Cloud Infrastructure service tailored for AI workloads. OCI Supercluster will combine thousands of Nvidia GPUs and high-speed memory to train LLMs. On top of that, it’ll offer cloud services that it, other SaaS providers, and enterprises can use to build generative AI into their own applications. The first of those applications in which the capabilities will show up is Oracle Fusion Cloud Human Capital Management (HCM), where it’ll help users by writing, summarizing, or recommending content.
ServiceNow puts gen AI in Now Assist for Virtual Agent
Now Assist for Virtual Agent will eventually allow ServiceNow customers choose from a range of generative AI tools (initially, only OpenAI and Microsoft Azure tools are supported) to help them design self-service business processes by accessing relevant pieces of enterprise content.
Salesforce bundles generative AI tools in one AI Cloud blanket
For $360,000 a year, Salesforce’s AI Cloud Starter Pack bundles all its generative AI services in a single offering. It includes Slack GPT, Tableau GPT, Apex GPT, MuleSoft GPT, Flow GPT, Service GPT, Marketing GPT, Commerce GPT, and a couple of other tools to wrangle data and train models. Analysts don’t expect AI Cloud to attract many customers, though, as there’s no unified interface for all the disparate tools, and many customers will already have access to the ones they most need as part of other Salesforce offerings.
Salesforce’s Einstein GPT learns about marketing and e-commerce
Salesforce said its Einstein GPT generative AI toolkit will soon be able to personalize marketing emails by automatically segmenting audiences, resolving customer identities, and creating relevant content for them — although most of the new features in Marketing GPT aren’t slated for release until October. It’ll also lend a hand with e-commerce and deliver a multi-channel “concierge” experience from February 2024.
May 2023
Alteryx generates data visualizations with AI
Analytics automation vendor Alteryx uses generative AI to add three new functions to its data visualization platform. Magic Documents, a feature of Alteryx Auto Insights, uses generative AI behind the scenes to deliver data visualization summaries and automatically generate targeted PowerPoints and emails based on them. The other two additions are in Alteryx Designer: Workflow Summary enables users to automatically document processes and related metadata in natural language, while the OpenAI connector lets them incorporate ChatGPT interactions in their data workflows.
SAP plans home-grown generative AI assistant
Hot on the heels of its announcement of a generative AI collaboration with IBM, SAP said it’s also developing its own generative AI technology. SAP Digital Assistant won’t be a fully-fledged chatbot like ChatGPT, but rather a contextual prompt in apps within SAP’s Sales, Service, Marketing, and Commerce Cloud products. Be patient, though: it won’t even enter beta testing until Q4 2023.
Tableau GPT is latest target of Salesforce’s generative AI initiative
Tableau is getting the Einstein GPT treatment. Despite the GPT name, Salesforce is deploying a variety of LLMs from OpenAI, Cohere, or its own development team in efforts to add generative AI to its product range. Tableau GPT will use generative AI to make data analysis more like a conversation, the company said.
ServiceNow develops generative AI applications with Nvidia
ServiceNow sees plenty of scope for using generative AI to help enterprises automate more of their workflows, and is working with Nvidia to develop AI models dedicated to those tasks. Nvidia has refocused its business on AI over the last few years, working to deliver computing power and software optimized to train AI models such as the LLMs behind generative AI, and make them available as a service through its DGX Cloud offering. That’s what ServiceNow will use to build the models it’s developing.
Informatica will use generative AI to manage data
Intelligent Data Management Cloud (IDMC), Informatica’s tool to manage enterprise databases, is getting a generative AI makeover. One of its components, data management automation engine Claire, will soon become Claire GPT, which will enable users to ask questions of their data in natural language without having to learn SQL. Claire GPT is still in private preview, but the company plans to release it before year-end.
SAP aims to automate application management with generative AI
SAP intends to build IBM’s Watson AI engine into much of its software portfolio, beginning with the SAP Start AP, where it’ll enable users to interact with SAP applications using natural language. Watson isn’t itself a generative AI tool, but SAP sees its natural language processing capabilities as a first step on the way to use generative AI more widely. The two companies are jointly developing LLMs and generative AI capabilities, they said.
