If you run marketing operations on Salesforce, budget season usually brings two questions at once.
- Leadership wants to know your AI plan.
- Finance wants to know why last year’s tools aren’t showing more return.
Both questions tend to have the same answer.
AI tools and the stack you already pay for rely on the same groundwork underneath them. When that groundwork is thin, new spend tends to inheri the same gaps.
Your team is probably stretched already – almost every marketing team is – so every new line item also needs to factor in someone to set it up and keep it running.
This guide covers five budget decisions Salesforce marketing teams commonly face for 2027. Each starts with the question you’re likely to hear in the review, then covers what to fund and how to show it worked.
This guide draws on the 2026 AI Advantage Assessment from MarketingOps.com, the Marketing Ops community founded by Mike Rizzo and the team behind MOps-Apalooza, run with research sponsor Octave. It also draws on conversations with Salesforce teams we work with. The survey retained 111 validated responses, with different bases for individual questions. Its authors describe the findings as directional.
1. Tie AI spend to a task you can measure
The question you’ll hear: “What’s our AI plan for next year?”
Most teams already use AI, just not where it touches pipeline. In the MarketingOps survey, 82% used AI increasingly or all the time for idea generation. For lead routing or notification, that figure was 20%, among respondents where each use case applied.
The report reads this as a gap between individual tasks and connected workflows. Individual tasks are easy to start. Connected workflows need clean inputs, clear owners and a way to check the output.
The savings aren’t automatic either. 43% agreed that editing AI output cost more than it saved them.
What to fund: a pilot built around one task, with a defined input, output and owner. For example, an AI sales assistant drafting a follow-up from verified Salesforce data for a rep to review. Write down what the draft may use and what a good draft contains.
How to show it worked: record the current process before you start, including time taken and how often the output needs correcting. Afterwards, measure the whole task, checking and corrections included. If a simple rule or journey already handles the task reliably, compare against that too.
Give the pilot a review date and expand it only when the results justify it.
2. Fund the context people and AI both rely on
The question you’ll hear: “Why does our outreach still feel generic?”
In the same survey, 73% agreed that context was the biggest limitation. The report means business context, such as ICP, personas, positioning and objections. Those definitions matter whether a rep or an AI tool makes the call.
Buyer context matters just as much. A Salesforce record shows where someone works and that they filled in a form. It rarely shows what they read or what they came back to, which is what first-party intent data adds.
Without that, reps start calls cold and AI agents work from thin records. Our guide to intent data for Agentforce shows what that context looks like for an agent.
Capital Economics saw this with every report it published. Each release and conference appearance brought a wave of inbound interest, and reps couldn’t tell a warm prospect from a curious reader.
With each reader’s behavioral history in Salesforce, reps could focus on people actually in market and prepare more relevant discovery questions. Read the Capital Economics case study.
What to fund: start from one workflow, such as prioritizing inbound leads. Agree what makes a good fit, which activity counts and how recent it needs to be.
How to show it worked: run real records through it. Does activity land on the right lead, contact or account? Can a rep see why a lead was prioritized? Fund the tracking or mapping needed to answer yes to both.
3. Budget for what happens after a lead shows interest
The question you’ll hear: “Sales says our leads aren’t worth chasing. Why?”
A signal only helps if it changes what someone does next. In our customer research, one team had useful engagement data but still sent a generic follow-up.
A score or sales alert doesn’t decide who acts or what they say. That needs Sales and Marketing Ops to agree, and the agreement is usually where budgets come up short.
What to fund: time to agree what qualifies for follow-up, who owns it and what happens if the first owner doesn’t act. Write those rules into your lead routing. Show the reason behind each alert where reps already work, in Salesforce or Slack, so they don’t research across systems. Pilot with one team or inquiry type before rolling out.
How to show it worked: track lead acceptance, time to useful follow-up and progression to opportunity. A simple rule that everyone uses consistently can beat a complex model nobody trusts.
Our EdTech lead management case study shows how one team turned scores into follow-up decisions.
4. See campaign quality before you move money
The question you’ll hear: “Which campaigns should we cut?”
Form completions tell you what generated inquiries. Budget decisions also need to know which sources bring demand that Sales can work with.
Decoded, an education company for AI, ran into this when inbound volume rose about 150% and some reps carried more than 150 open leads. Campaign data sat in ad platforms while leads lived in Salesforce.
Behavioral scoring and fit grading inside Salesforce let the team triage inbound and judge campaign quality within days. It could also trace digital activity through to closed-won revenue. Read the Decoded case study.
National Fostering Group judged campaigns on figures from an external agency, with no independent view of what worked. Once it could see how each inquiry behaved, it moved budget to the activity that converted. Inquiry cost per acquisition fell by more than 30%. Read the National Fostering Group case study.
What to fund: consistent campaign tagging, source data on CRM records, agreed lead statuses and reports people actually open. Use the same definitions throughout, so differences in data entry don’t look like differences in performance.
How to show it worked: compare sources on accepted leads, opportunities and revenue. Allow for your sales cycle. Early qualification feedback can flag weak demand quickly, but pipeline and revenue comparisons need time to mature.
Attribution shows how sources are associated with outcomes. On its own, it doesn’t prove a campaign caused incremental revenue. Our guide to marketing campaign reporting covers the setup, and B2B attribution in Salesforce shows it in practice.
5. Count your team’s time as part of the cost
The question you’ll hear: “Can’t we do this with what we already have?”
Often you can, and it’s a reasonable first option. Capacity still belongs in the comparison, because building inside an existing platform uses people even when it adds no license cost.
An integrated product doesn’t have to mean a long project either. National Fostering Group, mentioned above, was live within a month.
What to fund: compare each route on delivering the same workflow, using the people you actually have. Include delivery, testing, training and upkeep. Our SalesWings vs custom-built comparison breaks down where that effort goes.
How to show it worked: name an owner, track time to a usable capability and check adoption. Schedule time to review tracking and reporting as campaigns and the website change.
For total cost and the business case, see our MarTech budget planning guide.
What to bring to the budget review
For each line item, state the workflow it supports, who owns it and what should change. Ask for proof on representative Salesforce records.
| Decision | Evidence to request | What to measure |
|---|---|---|
| AI pilot | The complete task, including review and corrections | Net time saved and output quality |
| Buyer context | Relevant activity on the correct CRM record | Coverage, freshness and use in conversations |
| Follow-up and ownership | An inquiry reaching an owner with usable context | Acceptance, follow-up time and opportunity rate |
| Campaign reporting | A source connected to agreed CRM stages | Cost per accepted lead, pipeline and revenue |
| Team time | An owner, delivery time and maintenance plan | Time to usable capability and adoption |
Where SalesWings fits
If your 2027 plan includes making buyer interest usable in Salesforce, that’s the part we help with.
SalesWings connects first-party buyer activity to Salesforce records. Sales, Marketing and AI agents can then see each lead’s and account’s areas of interest and decide the next-best action. It works inside the systems your team already uses, and our customer success team helps you set it up around your own workflow.
See how 13 teams have applied it, or test it on your own data with a pilot first.
