BNP Paribas highlighted key takeaways, including agentic AI, growth outlook, and pricing, from its meetings with investor relations of Salesforce (CRM), Workday (WDAY), and Snowflake (SNOW) at its BNP Paribas Strategic Access & Research Conference in Indian Wells, California.
Salesforce (CRM)
BNP Paribas reiterated its Outperform rating on Salesforce.
“Salesforce is doing all of the right things, but the share price will not likely react until Salesforce can 1) demonstrate stability in its core business and 2) demonstrate continued success and adoption of Agentforce along with Data Cloud. The company is signaling we should see revenue growth acceleration in H2 (and we believe it could be as early as Q3). In the meantime, Salesforce continues to manage its cost base, conduct prudent/targeted M&A, and has accelerated significant buybacks to further lower the share count,” said analyst Stefan Slowinski.
The analyst noted that while most investors do not see companies ripping and replacing their Salesforce CRM systems, there is clearly anxiety around the medium-to-long-term growth impact from agentic AI systems increasingly managing and executing customer-related actions.
Slowinski said Salesforce believes it is well-positioned to be a winner in the new agentic layer, as it benefits from the following things.
Context: Salesforce having the customer data and the metadata to provide customer context results in better and more accurate outcomes.
Work: Salesforce has the systems of work in its various apps that customers do not want to rebuild and then be responsible for maintaining, and even the Large Language Model, or LLM, providers are using Salesforce’s apps.
Agency: Salesforce’s Atlas Reasoning Engine can decide which workflows can be executed within Salesforce, and which workflows are best going to use an external LLM, and which LLM would be best for each task.
Engagement: Salesforce has the front-end interaction layer with Slack, but also enables users to leverage Agentforce via Teams, WhatsApp, Voice or other systems.
In terms of AI competition, the analysts said Salesforce sees companies coming back to Salesforce after testing solutions from start-up AI pure plays like Sierra AI and Decagon AI, and having tried do-it-yourself artificial intelligence, or DIY AI, using tools, including LLMs from OpenAI (OPENAI) and Anthropic (ANTHRO). Salesforce sees its 27 years of operational CRM data embedded within its systems of record as an advantage versus both subsets of new competitors, the analyst added.
Slowinski noted that from a pricing standpoint, investors continue to question Salesforce’s evolving AI pricing models.
The analyst said that while they do not see Salesforce looking to ramp monetization of API calls to its data, they do believe that customers will need to continue to be paying Salesforce customers (for premium tier seats) to use their data and metadata in other systems. To date, the company has not seen any mass piping of data out of Salesforce.
In terms of margins, Slowinski said Salesforce expects to continue to see the benefit of moving to third-party clouds from the hyperscalers (namely Amazon, but Google is also ramping up this year). LLM costs will weigh on gross margins, but Salesforce can secure volume discounts and also expects costs per token to come down over time.
Salesforce has started to report Agentic Work Units, or AWU, so that customers and investors can see the growth in actions performed by Salesforce Agents and not just token consumption, the analyst added.
Slowinski noted Salesforce remains confident that its marketing business will stabilize with new product functionality ramping, but also continues to see Commerce taking longer to demonstrate improvements as competitive pressures continue and as the product competes for R&D resources with other AI areas.
Workday (WDAY)
“Workday is returning to an organic product focus with a goal of delivering revenue growth acceleration. This may result in improved Bookings growth in H2 FY’27, and hopefully revenue growth acceleration in FY’28, but could come at the expense of some of the longer-term margin target. With a low attrition rate, a single data model positioning the company well for AI, and ramping of new products, services and business models in H2 and beyond, along with an attractive valuation, we reiterate our Outperform,” said Slowinski.
The analyst noted that Workday believes that its integrated data model across Human Capital Management, or HCM, and Finance provides the company with an advantage as it increasingly sells the full suite.
Slowinski said that while Workday’s co-founder has returned as CEO, the initiatives of the past three years remain relevant, including the push into mid-enterprise, the investments in international and channel selling, including Workday Go, and the push on full suite sales.
The analyst added that Workday used to generate 60% of growth through its existing customer base and then 40% through new customers. Currently, new customer growth is slower as some industries have slowed, such as state and local government and education, or SLED, Healthcare, and some parts of retail. While 60% of growth historically has come from selling into the installed base, Workday is now pushing in this more with AI capabilities and new products to its existing customer base, according to BNP Paribas.
On the risk to the seat-based model, Slowinski said Workday prices its initial contracts based largely on the size of the overall customer employee base and their needs, with a three-year minimum commitment. The analyst added that renewals then reflect the growing size of the customer’s business, and inflation (consumer price index, or CPI), and any new functionality purchased. While customers may be able to reduce spending if the size of their institution has reduced, the ‘floor’ is typically the initial year 1 base price. In the near term, if customers are no longer growing headcount, and as CPI inflation has come back, then that impacts revenue growth, the analyst noted.
Slowinski said Workday’s Data Cloud offering is expected to ramp up in the second half, which will be charged on a consumption basis based on the amount of data. In addition, Flex Credit revenues are expected to start ramping up in fiscal year 2027 for new agentic offerings, which will be outcome-based, helping ensure that customers understand the value from the new services.
The analyst noted that Workday has mentioned potential monetization through charging more for API calls. The analyst believes this is likely to be limited to heavy users and to third parties who attempt to come ‘over the top’ to access Workday data.
Workday is being less vocal about seeing revenue growth acceleration potential this year (unlike Salesforce and ServiceNow), but Workday believes that should it be successful with new initiatives, second-half bookings should show improvement. The analyst said that if this is the case, then they believe Workday could deliver an inflection in revenue growth in fiscal year 2027.
Workday is committed to continuing to improve margins over time, but as the company re-enters a period of greater focus on organic product development, the analyst believes margin targets may be lowered, with the company focusing on accelerating top-line growth and ensuring the product portfolio is relevant in the upcoming Agentic AI world.
In addition, the analyst said that Workday is committed to continuing to improve margins over time. However, as the company re-enters a period of greater focus on organic product development, they believe margin targets may be lowered, with the company focusing on accelerating top-line growth and ensuring the product portfolio is relevant in the upcoming Agentic AI world.
Snowflake (SNOW)
BNP hosted Snowflake’s CFO Brian Robins and Investor Relations at the conference.
“We see Snowflake as a key beneficiary of AI. Not only can AI help drive acceleration in demand for Snowflake’s core platform (migration projects), but it also provides opportunities for Snowflake to extend its use (more use cases) and extend its reach (more users) through development simplification. We reiterate our Outperform,” said Slowinski.
The analyst noted that investor excitement around the company’s AI coding agent Cortex Code, or CoCo, is clear. Snowflake sees AI as an opportunity to increase the accessibility of its technologies to more business users, rather than just business and data analysts, by putting tools such as CoCo in people’s hands.
The analyst added that while Snowflake partners with the main LLM providers, including Anthropic, making Claude available to run on Snowflake, the company believes customers are finding value in using Snowflake’s own coding tool for data operations. This is because it is better at identifying patterns and context within Snowflake data, while also benefiting from Snowflake security and governance controls
In terms of the LLM providers themselves, Snowflake has collaborations with both Anthropic and OpenAI. While Snowflake does not rule out those companies potentially looking to enter the data platform market, the LLM companies appear more focused on disrupting the applications market, according to the analyst.
On AI more broadly, Slowinski said Snowflake is seeing some customers already replace some third-party point solution application functionality by running workflows directly in Snowflake.
The analyst noted that overall, AI native is likely less than 2% of revenues but may have already helped drive the beat in the fourth quarter. The analyst would expect AI Annual Recurring Revenue, or ARR, to be provided periodically. While AI products will likely create some gross margin headwind, Snowflake is committed to maintaining product gross margins at 75% this year, the analyst added.
AI technologies can be used to accelerate system and data migration (automate code conversion, but also testing), while reducing risk, time, and cost. Leaning into this is a top priority for Snowflake, and the analyst expects to hear more about this in the year in terms of tools to accelerate migration, and also pricing, contracting, and packaging.
In addition, the analyst noted that Snowflake signing a $400M deal in the fourth quarter, having many other large deals, and growing Remaining Performance Obligations, or RPO, 42% last quarter, has raised eyebrows about how Snowflake still has room to scale significantly with existing large customers.
On Databricks, Slowinski said that competition from Databricks is not new, and competitive fears appear to have died down from their peak a couple of years ago.