When Amrutha Suresh talks about AI at Asana, she doesn’t describe it as a feature or another bolt‑on tool. She describes it as an operating layer that’s changing how go‑to‑market teams work and how pipeline is being generated. As Head of Enterprise & GTM AI, she reports to the CIO while enabling business leaders and users across sales and marketing, which puts her at the center of technology and the field.
“AI has stopped becoming an additional feature. It’s changing how we operate. In fact, it’s becoming the core operating layer.”
From GTM Architect to Enterprise AI Leader
Suresh says her path to leading GTM AI at Asana was a very intentional evolution rather than a pivot. She started as a go‑to‑market business architect, then moved into a role heading technology for GTM, mostly focused on sales and marketing, before taking on her current position leading Enterprise and GTM AI. Across those roles, she describes her role as “taking the strategy and translating it into operating models that drive business value.”
“What’s changed now,” she explains, “is the technology inflection point.” With AI, the biggest opportunity she sees is “not optimizing tools,” but “really thinking about redesigning work.” She believes her deep understanding of the GTM stack, credibility with the executive leadership team, and knowledge of the friction points and pains that the field goes through has helped her embrace the transition.
From Back-Office IT to Co-Creator of GTM Strategy
Suresh reports into the CIO, but her customers are the GTM leaders and frontline teams responsible for hitting the number. That organizational setup makes one skill non‑negotiable: influencing without authority.
“Influencing without authority is the survival skill I need to have,” she says. “You can’t just wait for someone to hand you requirements anymore.”
She describes the mindset shift many modern IT and RevTech leaders are making: away from a ticket‑taking, back-office function and toward product management for the business. That means:
Building deep relationships so you are in the room for staff meetings, planning, and trade‑off discussions.
Understanding friction points and process pain at a level that lets you spot patterns before the business even articulates them.
“Selling” internally just as hard as any seller in the field, clearly articulating what’s in it for each team and how an initiative moves the needle on pipeline, retention, or profitability.
“Opportunity recognition or pattern recognition is very important,” Suresh explains. “And in order for you to make that skill come to life, everybody needs to be a salesperson, articulating the value and what’s in it for them.”
It’s poetic that the person responsible for operationalizing AI across GTM does it largely through human relationships. “People are still people,” she says. “That is something that can never be replaced.”
Redefining Success: Why Adoption Is Now a “Vanity Metric”
A year ago, Suresh would have answered the question “What does success look like?” with one word: adoption. Today, she’s blunt about how much that bar has moved.
“Adoption is a vanity metric in my opinion,” she says. “AI should be a business driver.”
Instead of asking how many users logged in, her team now asks how much AI is influencing the metrics that matter most to the business:
Pipeline generation
New meetings booked
Customer retention and expansion
“At the end of the day, the ROI of an AI program comes from influencing these types of metrics,” she says. “Sure, people should use the tool, but adoption is secondary.”
That shift has forced Asana to change how AI is integrated into workflows. AI can’t live as a sidecar tool bolted onto the edge of an existing process; it has to sit inside the flow of work and show up in the same dashboards and reviews leaders already use to run the business. Experiments are still essential, but Suresh is just as focused on something she calls “pilots to program scaling”: building the operating model to move from interesting tests to durable, programmatic impact.
That’s where governance comes in, a word many GTM leaders resist.
“Governance can feel very heavy or restrictive,” she admits. “But AI initiatives should be managed like a program where you’re reporting on ROI, progress, risks, and what you’ve learned.”
At Asana, that rigor shows up in the form of an AI operating committee or council, which Suresh leads, bringing functional AI champions together to align on priorities, share blockers, and prevent fragmentation.
Turning AI Into a Quota-Carrying “Rep”
If AI is going to be treated as a business driver, it also has to be treated like a participant in the revenue engine. That mindset is already changing how Asana thinks about coverage models and quota assignments.
On the inbound side, Asana uses a vendor’s AI SDR on the website. When a visitor lands on the site off hours or when human reps are busy, the AI agent greets them, captures context, and initiates a conversation. The result: incremental pipeline that simply wouldn’t exist without AI.
“We are squeezing additional juice,” Suresh says. “This would not have happened if AI was not there.”
What started as an experiment is now expanding into lower‑value lead outreach and other segments where fully staffing human coverage would be cost‑prohibitive. And the way her team talks about this internally sounds a lot like how they talk about people.
“We’re starting to think about AI as a rep,” she explains, “and how we can set targets and quotas for this AI agent to perform in certain segments.”
The team is now forecasting how much pipeline they have sourced through AI and how much they expect AI to deliver in a given quarter.
The concept is making its way into segmentation language as well. Alongside marketing‑qualified and sales‑accepted leads, Suresh and her peers are talking about “AI qualified leads” (AQLs), opportunities that an AI agent has identified, nurtured, or progressed. In a world where overall website traffic may decline as AI search rises, but visitor intent is higher when they do arrive, that incremental layer matters.
“You’re not securing additional headcount to do this grunt work,” she notes. “We already have the AI infrastructure. It’s just one aspect of it that’s going to capture additional incremental numbers.”
With new prospecting capabilities that have been built in‑house, the goal is to increase meeting bookings, and she notes that “the way we set targets, books, all that” is changing. She is careful to say that she would not claim they are “100% there” or have AI “working autonomously” yet, but she sees behavioral change in revenue and marketing strategy beginning to emerge.
Inside Asana’s Layered GTM AI Architecture
To keep all of this from devolving into “shadow AI” and fragmented experiments, Suresh anchors GTM initiatives in a clear view of the AI architecture supporting them. She borrows NVIDIA CEO Jensen Huang’s metaphor of a layered cake and adapts it to Asana’s world.
Her stack has several key layers:
An application layer where auxiliary tools and GTM systems like CRM and outreach platforms live.
A data and context layer, including the enterprise AI platform standard, access controls, and how context is segmented for each function.
A models and infrastructure layer, covering LLMs, vector databases, and underlying data platforms such as Snowflake or Databricks.
When someone proposes a new AI capability, the first question is where it sits in this architecture and how it will interact with existing components. That architecture lens also guides the classic build versus buy decision.
If a capability is narrow, vendor‑contained, and doesn’t need to deeply integrate across the ecosystem, Suresh is comfortable turning on a vendor’s AI feature, “obviously in a responsible way,” she notes. That’s how tools like an AI SDR agent enter the stack.
But as soon as a capability needs to interact with multiple parts of the GTM system and meet enterprise‑grade standards, her team steps in to design and integrate it as part of a broader workflow.
She points to Asana’s homegrown revenue intelligence engine as an example. The system aggregates signals from intent tools, enrichment providers like ZoomInfo and 6sense, public data such as 10‑Ks and M&A news, and Asana’s own usage patterns. A custom data science model scores accounts and gives sellers a clear, prioritized list of where to focus.
“As a rep, maybe I have 100 accounts,” she says. “Here are your top 20 accounts. This is what you as a human need to go after and here is why.”
To maintain consistency, the hand‑off from an AI website agent to outreach flows through that same revenue intelligence layer. The goal is simple: the customer should experience one coherent narrative, whether they’re interacting with marketing, sales, or an AI teammate.
She notes that a pilot of this system in November for about six weeks showed “substantial meeting increase influence through the system,” and that while she would not yet attribute revenue directly, it has clearly made reps’ lives easier by reducing the time spent prospecting across multiple tools.
Democratizing AI While Managing “Shadow Tooling”
One of Suresh’s recurring worries is “shadow AI,” a proliferation of rogue tools and agents that may help a few power users but undermine security, compliance, and customer experience. Her solution is not to clamp down on experimentation; it’s to create best practices and guardrails.
“Shadow tooling occurs when demand is not met or people don’t know what exists,” she explains. So the first move is listening: understanding what frontline teams actually need, where current capabilities fall short, and when they simply don’t know where to go.
From there, her team sets clear guidelines:
When and how people can pilot an AI tool, including free trials.
What approvals are needed if they’re using their own functional budgets.
The terms and conditions around data, security, and what it takes to move a pilot into an enterprise‑supported program.
“We shouldn’t be stopping experimentation,” she says. “It has to be very democratized, but with clear guardrails and processes so people understand where to go for what.” Her AI council plays a key role in socializing those norms and pulling promising experiments into the core roadmap.
What Boards and CEOs Should Understand
If Suresh could wave a wand and change one thing about how boards and CEOs talk about AI, it would be this: stop treating it as a tools conversation.
She doesn’t dismiss the curiosity. But she wants the conversation to move up a level, from “Are we using vendor X?” to “What are we trying to do?” and “How will this change the way that we work?”
“I really hope that people understand this more from a holistic lens centered on what we are trying to do and what the outcome is, as opposed to a tooling conversation,” she says.
For boards, that means asking:
How is AI changing quota models, coverage, and segmentation?
How is it reshaping the operating model across marketing, sales, and CS?
How are we measuring AI’s impact on core business KPIs, not just usage?
Advice to the Next Generation of GTM Operators
To operators just entering the field, Suresh doesn’t promise a stable roadmap. In fact, she warns them to expect the opposite.
“People ask me in the technology org, ‘What’s your roadmap?’” she says. “Directionally, I can tell you what we’re going to do. But every day there’s a new release. Don’t plan out your entire life. That’s just not the way AI is going to work.”
Instead, her advice centers on flexing three muscles:
Intellectual curiosity: a genuine desire to understand both the technology and the business context it lives in.
Adaptability: comfort with changing plans as new capabilities emerge.
Commercial fluency: the ability to tie any AI idea back to pipeline, retention, productivity, or profitability.
Her parting words, “If you can combine deep understanding of the business with pattern recognition and the ability to sell your ideas internally, people will listen.”