Sanjjna Parulekar

Sanjjna Parulekar grew up in San Jose, surrounded by parents who both worked in tech, so she treated software less like an abstract industry and more like the family business. That shows up in how she talks about AI: not as futuristic magic but as infrastructure that needs to work inside the messy reality of CRM data, sales teams, and enterprise processes that have been in place for years.

By the time Salesforce Einstein launched its AI for CRM vision, she was already close to the product and customer narrative, later stepping into Director and then VP roles focused on Einstein and broader Salesforce AI. Instead of chasing shiny models, she kept a simple constraint in place: every AI feature had to move a core CRM metric, from lead conversion to case resolution time, or it did not deserve space in the release notes.

That discipline made her a natural fit to lead the go to market story around Einstein Copilot and then Agentforce, Salesforce’s push into AI agents that can take action, not just generate text. Her background in product marketing meant she spoke both API and business outcome, so she could translate research and engineering work into offers that sales teams could pitch and customers could test in narrow, high value workflows.

On podcasts like Mission’s IT Visionaries and Shift AI, Sanjjna hammers the same point: AI at Salesforce is about human plus AI augmentation that lives inside CRM, not about replacing entire teams with generic chatbots. She repeatedly draws a line between vendors that ship raw APIs and dashboards and Salesforce’s approach, which is to wire models into existing objects like accounts, opportunities, and cases so that recommendations and agents feel native instead of bolted on.

When Einstein Copilot and Data Cloud were promoted as the way companies would get more out of generative AI, she fronted the narrative, explaining how Copilot sits alongside sales reps and service agents as a coworker. That framing matters, because it lowers buyer anxiety: customers are not told to rebuild their stack, they are shown how AI can incrementally upgrade familiar workflows like drafting emails, summarizing calls, or generating knowledge articles inside Salesforce.

With Agentforce and the Agent Exchange, she pushed a different but related lever: turning AI agents into a partner friendly marketplace rather than a closed Salesforce only feature. In public conversations she highlights that partners can build on top of Agentforce, then list agents that tap Salesforce data and actions so they can grow their own revenue without owning foundation models or infrastructure. That converts Salesforce’s massive partner ecosystem into a distribution engine for Agentforce in a way that independent AI tools struggle to match.

1. Attach AI to an existing high frequency workflow instead of selling it as a blank canvas. Sanjjna’s team consistently ships Einstein, Copilot, and now Agentforce inside objects and flows that sales and service teams already use, which shortens adoption time and lets Salesforce measure impact on metrics buyers actually care about, like pipeline velocity or case deflection.

2. Sell coherence rather than raw openness. In interviews she calls out the trap of platforms that brag about being open but leave customers stitching together brittle APIs. Salesforce flips that by doing the integration work in house, so customers get curated actions, governed data access, and guardrails that feel opinionated, which lets risk averse enterprises say yes to AI without hiring a platoon of ML engineers.

3. Turn the ecosystem into the growth loop. By launching Agentforce with Agent Builder and an Agent Exchange, she gives partners a concrete economic reason to implement Salesforce’s AI stack instead of rolling their own tools. Every partner that ships an agent effectively markets Agentforce to its own customers, so distribution compounds through the existing Salesforce partner and customer network rather than relying on net new channels.