Revops best practice

10 RevOps Best Practices to Follow in 2027

Executive Summary

Agents are predicted to execute 75% of RevOps tasks in workflow management, data stewardship, revenue analytics, and RevTech administration by 2028. That changes what good RevOps looks like. The best practices that matter now are less about building reports and more about building foundations: complete customer data, shared definitions, a consolidated stack, governance for agents, and a team that runs systems like products. Here are ten practices for 2027, with insights from RevOps and AI leaders on The Revenue Lounge podcast.

Most RevOps teams are already experimenting. An IDC study found 68% of organizations are scaling or optimizing AI across revenue functions. The foundations lag behind. In the April 2026 LXA and LeanData survey of enterprise revenue leaders, 47% said manual processes that cannot scale are their top lead management failure, and only 50% were confident they could govern AI safely at scale. The practices below close that gap.

Get our latest insights into your inbox

10 RevOps Best Practices for 2027

Here are expert-curated best practices for you to follow as a RevOps practitioner in 2027:

1. Measure outcomes with leading and lagging indicators

Most RevOps scorecards still look backward: win rate, quota attainment, pipeline coverage. Those numbers matter, but by the time they move it is too late to act. Pair them with leading indicators that show where the quarter is heading: first meetings booked, deals advancing from discovery to evaluation, senior buyers engaged, and stakeholders going quiet.

Apply the same discipline to AI. Amrutha Suresh, Head of GTM AI & Innovation at Asana, separates AI metrics into three tiers: adoption (useful diagnostics, not ROI), efficiency (time saved), and business outcomes (pipeline, conversion, win rate, churn), and insists measurement is designed in from the start:

When you think about your design, definitely factor in attribution, tracking, and reporting. That way you're able to measure business outcomes, not just activity.

At Mimecast, Tim Seamans’s team works the same way: “We have a lot of leading and lagging indicators that help us understand our impact.”

2. Design for the full revenue journey, not just new logos

The earlier guide to RevOps was to break silos across the “lead to cash to renewal and expansion” process. In 2027, the post-sale half of that journey carries more weight. Christine Royston, former CMO at Wrike, frames even marketing this way:

I really consider marketing a revenue function. Marketing should be looking at how close they can get to revenue.

She adds that it is much easier to retain customers than acquire new ones, so marketing, sales, and customer success need shared goals and shared data after the contract is signed. Mimecast shows the payoff: by unifying engagement data across previously siloed systems, it launched an internal Expansion AI tool that drove $2M in attributed expansion revenue in its first 80 days. Read the full story here.

Do this: Make sure the same account, contact, and engagement data powers pipeline, renewal, and expansion reporting, so every team sees one customer.

3. Build the data foundation before the agents

Every AI project in RevOps inherits the quality of the data underneath it. 76% of CRM users say less than half of their organization’s CRM data is accurate and complete. Gartner predicts that through 2026, organizations will abandon 60% of AI projects unsupported by AI-ready data.

Pino Soro, Chief Business Transformation Officer at Quickbase, defines AI-ready data with four properties: clean, connected, contextual (defined), and governed (owned). He also argues for building the foundation once rather than chasing every new tool:

Figure out what your AI operating system is, the foundations you need in place around data, systems, and governance. Because what you'll find is the business is gonna come at you with: Claude just released this new connector, I want it tomorrow. If you don't have those foundations in place, you're gonna be the office of no, because you're just trying to protect the business.

4. Capture customer activity automatically

The biggest gap in most CRMs is data that was never entered: emails, meetings, calls, and the stakeholders who joined them. Reps will not log all of it, and training does not change that for long. Automate capture so every interaction lands in the CRM, matched to the right contact, account, and opportunity, with zero rep effort.

Christine Royston, former CMO at Wrike describes why this matters for AI: data from disparate systems, including sales conversations that rarely make it into the CRM, has to come together in one place before you put agents on top of it.

Teams can make decisions based on the full picture, not just, well, this is the tool that I log into every day.

This is the layer Nektar provides: automatic capture across email, calendar, meetings, and calls, written natively into Salesforce, with missing contacts and contact roles created and associations kept correct as records change.

Learn more here.

5. Treat documentation as context for AI

Documentation is now the context AI uses to answer questions correctly. Adrian Rosenkranz, CRO at Webflow, built a shared handbook that loads on every prompt across his organization:

The thing I was really struggling with is how do I make sure that when anyone in my org asks a question, we actually make sure we go through a shared handbook first? Because there's only going to be a handful of people who know exactly how all the contracts are stored and what the definition is.

Each team owns its part: RevOps owns the Salesforce definitions and metric logic, legal and RevOps own the contracts knowledge base, and product marketing owns the ICP. Pino adds a practical benefit: codifying business context in documents and systems of record, rather than inside one vendor’s platform, keeps you free to switch AI models as better ones emerge.

Do this: Turn your process documentation into an owned, versioned knowledge base that both people and agents read from.

6. Define shared terms in a semantic layer

The 2023 version recommended a common data dictionary so every team speaks the same language. For AI, that dictionary becomes a semantic layer, and it is no longer optional. Without it, two people ask an agent the same question and get two different answers.

Amrutha Suresh, Head of GTM AI and Innovation at Asana, shares an example that shows the difference. Raw data might say a customer logged into the product 70 times. A semantic layer turns that into a business signal: a high-intent user in a strategic account whose pattern resembles customers who historically expanded. On top of that, a context layer assembles what each decision needs:

For each decision, you can dynamically assemble the context that is relevant for that specific use case or agent you want to build.

Do this: Start with the definitions agents get wrong most: ARR, active customer, qualified opportunity, stage criteria, and engagement. Give each a single owner and a single source.

7. Consolidate the stack around one data spine

The original guide predicted 2024 would be the year of stack consolidation. It was, and it is continuing. Gartner published its first Magic Quadrant for Revenue Action Orchestration in December 2025, a category that merges sales engagement, revenue intelligence, and sales force automation into platforms built on one normalized revenue data model. Salesforce found more than 80% of sales teams without a unified platform plan to consolidate.

Consolidation is about data, not just licenses. Decide which system is the source of truth for each data type, which tools write and which only read, and remove duplicates that send agents conflicting signals.

8. Govern AI agents from day one

Agents in RevOps route leads, update records, and draft customer-facing messages. Governance determines how much latitude they get. Adrian Rosenkranz, CRO at Webflow, argues it speeds teams up rather than slowing them down, because people move faster when they trust the answers:

Local innovation is amazing. But empowering them to innovate is about how they show up for customers better.

His practical rules: give deterministic questions (contract terms, metric definitions, ICP) one governed answer, leave open-ended judgment to people, and keep human ownership on actions that are hard to reverse. Amrutha adds that security, legal, and enablement should be design partners from the start, not reviewers at the end.

Do this: For every agent, document which objects and fields it may change, which actions need approval, and where changes are logged.

9. Run RevOps like a product team

AI systems drift as processes change, and a ticket queue never notices. Quickbase restructured RevOps, business systems, and data engineering into one team that runs on roadmaps, business product managers, and technical owners. Pino Soro’s starting point for prioritizing that roadmap is deliberately narrow:

If we started from where can AI help, we're gonna find a hundred things we could do. And then we're gonna spread ourselves thin. We end up a little bit better at everything but we didn't really transform anything.

Default’s State of AI in Revenue Operations report supports focus: teams running one or two AI workflows often reported stronger time savings per use case than those spread across seven or more, and companies that assigned AI ownership to RevOps reported more workflows in production. Amrutha sees the same shift in the role itself, with RevOps moving from implementing third-party tools toward GTM engineering that builds proprietary capabilities.

10. Build AI fluency across the revenue team

The original guide closed with democratizing data: give non-analysts access and the skills to use it. In 2027 the equivalent is AI fluency. Mimecast built a tiered program, starting with a short certification for everyone and a dedicated AI induction for every new hire. Tim reported close to 95% of employees using generative AI day to day, and he is direct about what is at stake:

AI is not going to replace people, at least not necessarily right away and directly. But what will happen in the very short term is that people using AI will replace your relevance.

Pino Soro’s advice on change management applies here too: celebrate every small win, break big projects into visible weekly progress, and acknowledge that what got the team here was good work, even if it will not get the team where it needs to go.

RevOps Earns its Seat By Owning the Foundation

For RevOps, the seat at the strategy table comes from owning what AI depends on: complete data, shared definitions, a consolidated stack, and the rules agents follow. Tim Seamsn’s advice on how to frame the value is worth borrowing for any budget conversation:

Focus on your business levers... if you focus on your top line, you will also get that, you know, expense line to come down with AI.

Salesforce is the system of record. Nektar is the data foundation underneath it, capturing every customer interaction across email, calendar, meetings, and calls, creating the contacts and contact roles reps never add, matching activity to the right deal, and keeping those connections accurate as the CRM changes. That gives RevOps a complete, trustworthy base for every forecast, workflow, and agent above it.

Enjoyed our content? Follow Nektar on LinkedIn

In this blog

See what your CRM is missing before your agents do

Scroll to Top