Preparing Marketing and Sales for AI in 2026

AI is already changing how marketing and sales teams work.

Marketing teams are using AI for:

• research;
• campaign planning;
• content;
• personalisation;
• reporting;
• audience analysis;
• and automation.

Sales teams are using it for:

• prospect research;
• meeting preparation;
• CRM summaries;
• follow-up;
• pipeline administration;
• account insights;
• and prospecting.

The next stage is more significant.

AI is moving from helping employees complete individual tasks towards becoming part of connected marketing and sales workflows.

That means businesses need to prepare differently.

Giving teams another AI tool is not enough.

Before scaling AI across marketing and sales, organisations need to understand whether their People, Process, Technology, Data and Governance can support the next level of AI adoption. These same foundations sit within Lupo AI, connecting readiness and capability with practical AI implementation across the business.

What should a company do before implementing AI across marketing and sales?

Start by defining the business outcome you want AI to improve.

Then assess whether your marketing and sales teams are ready across five areas: People → Process → Technology → Data → Governance

The objective is not to make every part of the organisation perfect.

It is to identify the gaps that could prevent the next AI use case from delivering value.

Why marketing and sales need preparation before more AI

Marketing and sales are often among the first business functions to experiment with AI.

That makes sense.

Both teams work with:

• large amounts of information;
• repetitive administrative tasks;
• content;
• customer data;
• research;
• communication;
• reporting;
• and repeatable workflows.

But early experimentation usually happens at an individual level.

One marketer uses ChatGPT for campaign ideas.

Another uses AI to repurpose content.

A salesperson uses AI to prepare for meetings.

Someone else uses a CRM's built-in AI features.

These can all be useful.

But scaling AI is different.

Once AI becomes connected to:

• CRM records;
• lifecycle stages;
• lead qualification;
• marketing automation;
• sales pipelines;
• campaign activity;
• customer communication;
• reporting;
• or AI agents,

the quality of the underlying operating model matters far more.

Start with the business outcome

The first question should not be: “Which AI tool should Marketing and Sales use?”

Start with: “What business outcome are we trying to improve?”

Marketing objectives might include:

• producing campaigns more efficiently;
• improving lead quality;
• increasing personalisation;
• reducing repetitive campaign administration;
• accelerating content production;
• improving reporting;
• or identifying opportunities faster.

Sales objectives might include:

• reducing CRM administration;
• improving meeting preparation;
• increasing sales capacity;
• improving lead response;
• making follow-up more consistent;
• identifying stronger opportunities;
• or improving pipeline visibility.

Once the outcome is clear, you can determine whether AI is the right solution.

Technology selection comes later.

Avoid treating every AI opportunity as a priority

Marketing and sales teams can quickly generate dozens of AI ideas.

For example:

Marketing

• AI content generation;
• campaign agents;
• audience research;
• personalisation;
• lead scoring;
• performance analysis;
• reporting;
• campaign automation.

Sales

• prospecting agents;
• account research;
• meeting preparation;
• follow-up;
• CRM updates;
• call summaries;
• pipeline analysis;
• forecasting support.

Trying to implement everything simultaneously creates fragmentation.

Instead, prioritise use cases based on:

• business value;
• time saved;
• frequency;
• data readiness;
• technical feasibility;
• adoption readiness;
• risk;
• and measurement clarity.

The strongest first use case may not be the most advanced.

It is usually the one that solves a meaningful problem while remaining practical enough to pilot.

1. Prepare your people

AI adoption depends on employee capability.

Marketing and sales teams need more than access to AI tools.

They need to understand:

• where AI belongs in their role;
• how to provide useful context;
• how to assess AI output;
• when to verify information;
• what data is appropriate to use;
• when human judgement is required;
• and how AI changes the wider workflow.

For example, generating a sales email is easy.

Knowing whether that email:

• reflects the account correctly;
• uses reliable information;
• fits the stage of the relationship;
• follows brand standards;
• and should actually be sent,

requires human judgement.

AI capability therefore includes critical thinking, not only prompting.

Different roles need different AI skills

Generic AI training has limits.

Marketing and sales need role-specific application.

Marketing teams may need capability around:

• research;
• campaign planning;
• content development;
• brand consistency;
• customer segmentation;
• performance interpretation;
• automation;
• and AI-assisted reporting.

Sales teams may need capability around:

• account research;
• prospect preparation;
• CRM context;
• follow-up;
• meeting summaries;
• qualification;
• pipeline administration;
• and AI-assisted prospecting.

Managers need capability around:

• identifying strong use cases;
• evaluating outputs;
• supporting adoption;
• monitoring performance;
• and deciding where human oversight remains necessary.

The objective is not for everyone to become an AI specialist.

It is for employees to become capable AI users within the work they already own.

2. Prepare your marketing and sales processes

AI can accelerate a process.

It cannot automatically make that process good.

Before automating a marketing or sales workflow, map what happens today.

For marketing, that may be: Brief → Research → Create → Review → Approve → Publish → Measure → Optimise

For sales: Lead → Qualify → Assign → Research → Engage → Meet → Follow up → Update CRM → Progress Opportunity

Then identify:

• duplicated work;
• manual hand-offs;
• unclear ownership;
• inconsistent steps;
• repetitive administration;
• bottlenecks;
• approval delays;
• and steps that require judgement.

A broken workflow should usually be improved before AI is added.

Otherwise, you may simply automate the dysfunction.

3. Prepare your CRM data

Marketing and sales AI increasingly depends on CRM context.

That may include:

• contact information;
• company records;
• lifecycle stages;
• lead status;
• deal stages;
• campaign interactions;
• sales activities;
• customer history;
• ownership;
• and previous communications.

If that information is unreliable, AI inherits the problem.

Common warning signs include:

• duplicate contacts;
• inconsistent lifecycle stages;
• incomplete company information;
• unclear deal stages;
• inaccurate ownership;
• unused properties;
• outdated records;
• and important information stored outside the CRM.

Before giving AI more responsibility, ask: Can Marketing and Sales currently trust the CRM themselves?

If they cannot, the AI layer needs to be designed carefully.

You do not need perfect CRM data

This is where businesses can become stuck.

Your entire CRM does not need to be perfect before you can pilot AI.

Data readiness should be assessed against the specific use case.

For example, if the objective is to create automated sales meeting briefs, perhaps the AI needs:

• reliable company details;
• contact roles;
• opportunity information;
• previous meetings;
• recent activities;
• and selected account history.

Problems elsewhere in the CRM may not stop that particular use case.

Focus on: Does the AI have sufficiently reliable information for the task we want it to perform?

4. Prepare your technology and integrations

Marketing and sales work across multiple systems.

Your environment may include:

• CRM;
• marketing automation;
• website;
• advertising platforms;
• email;
• sales engagement;
• meeting tools;
• customer service;
• reporting;
• enrichment tools;
• data platforms;
• and other business applications.

AI becomes much more valuable when it can work with the right context.

But that often requires systems to connect properly.

Before implementation, ask:

• Where does the required information live?
• Which platform is the source of truth?
• Are integrations reliable?
• Is information duplicated?
• Can AI access what it needs?
• Should it have that access?
• Which system should execute the final action?

Sometimes the fastest route to better AI is actually a better integration or CRM architecture.

5. Prepare your governance

Marketing and sales teams deal directly with:

• customer data;
• prospect information;
• communication;
• company intelligence;
• brand content;
• pricing;
• commercial information;
• and customer-facing decisions.

That makes clear governance essential.

Teams need practical answers to questions such as:

Which AI tools can employees use?

Approved tools should be clear.

What data can employees enter?

Sensitive, customer and commercial information needs defined boundaries.

What information can AI access?

Access should match the use case.

What actions can AI take?

Reading a CRM record is very different from updating it or communicating with a prospect.

What requires human review?

External communications and higher-risk decisions may require approval.

Who is accountable?

An AI recommendation does not remove business ownership.

What happens if something goes wrong?

Escalation and monitoring should already be defined.

Good governance should make AI easier to use confidently, not simply create more documentation.

Map human judgement into the workflow

Marketing and sales contain decisions that should not automatically be handed to AI.

For example:

Marketing

AI may help:

• draft;
• analyse;
• research;
• categorise;
• or recommend.

But people may still need to own:

• brand decisions;
• strategic positioning;
• sensitive communications;
• campaign approval;
• and final quality.

Sales

AI may help:

• research;
• prepare;
• summarise;
• recommend;
• and automate administration.

But people remain especially important for:

• negotiation;
• complex qualification;
• relationship development;
• commercial judgement;
• and sensitive customer conversations.

The objective is not to remove humans.

It is to use AI where it creates leverage and preserve human input where it creates greater value.

Decide where traditional automation is enough

Not every marketing or sales workflow needs AI.

If the process follows a stable rule: If X happens → do Y traditional automation may be the better choice.

For example: If a contact submits the consultation form → create a sales task.

AI does not add much value there.

AI becomes more useful when the workflow involves:

• language;
• unstructured data;
• interpretation;
• changing context;
• classification;
• or reasoning.

The best operating model may combine: AI + traditional automation + human judgement.

Prepare managers, not only individual users

AI adoption can stall when employees receive training but managers do not change how work is managed.

Marketing and sales leaders need to understand:

• which use cases matter;
• what experimentation is acceptable;
• how success will be measured;
• how employees should use AI;
• which risks to watch;
• and how workflows may need to change.

Managers also need to create room for employees to practise.

If employees are expected to redesign their work while being measured only on maintaining old processes, adoption becomes difficult.

Create shared ways of working

Unstructured AI adoption often produces fragmentation.

One salesperson has their own prompts.

Another uses a different AI platform.

Marketing has separate templates.

Different teams follow different rules.

Useful individual experiments should gradually become shared organisational capability.

That may include:

• approved tools;
• role-based prompt patterns;
• reusable projects;
• workflow templates;
• quality checks;
• verification processes;
• approved use cases;
• governance rules;
• and examples of successful AI-enabled work.

The aim is not to stop experimentation.

It is to avoid every employee repeatedly solving the same problem independently.

Choose a controlled first pilot

Once the foundations are understood, choose a bounded use case.

For example:

Marketing pilot

Use AI to turn an approved campaign brief into structured first-draft content and repurposing suggestions.

Measure:

• production time;
• revision time;
• quality;
• and team adoption.

Sales pilot

Automatically prepare an account summary before sales meetings using approved CRM context.

Measure:

• preparation time;
• information quality;
• usage;
• and salesperson feedback.

RevOps pilot

Use AI to classify or structure selected unstructured CRM information before a controlled update.

Measure:

• administrative effort;
• accuracy;
• exceptions;
• and manual review required.

Start small enough to learn.

Then decide whether to: Scale → Improve → Redesign → Hold → Stop

Measure business outcomes, not AI activity

Avoid treating these as the main signs of success:

• number of AI licences;
• number of prompts;
• number of employees using AI;
• number of AI tools deployed.

Those measures tell you about activity.

Better measures connect directly with the business problem.

For Marketing:

• campaign production time;
• conversion;
• cost;
• output capacity;
• content performance;
• reporting effort;
• or lead quality.

For Sales:

• response speed;
• administrative time;
• meeting preparation time;
• conversion;
• pipeline velocity;
• CRM completeness;
• or seller capacity.

Ask: What became better because AI was introduced?

A practical readiness sequence for Marketing and Sales

Before scaling AI, work through this sequence.

1. Define the outcome

What should improve?

2. Select the use case

Choose a specific workflow.

3. Map the current process

Understand how the work happens today.

4. Review the people capability

Do employees understand how AI fits into the workflow?

5. Assess the technology

Can the systems support the use case?

6. Review the data

Is the required CRM and business information reliable?

7. Define governance

What can AI access and do?

8. Establish human oversight

Where should a person review or intervene?

9. Define measurement

What business outcome proves value?

10. Pilot and learn

Start controlled.

Improve before expanding.

This is a much stronger path than switching on multiple AI features simultaneously and hoping adoption follows.

Signs your Marketing and Sales teams may not be ready to scale AI

Look for these signals:

• employees are experimenting but there is no shared direction;
• CRM information is not trusted;
• Marketing and Sales disagree on lifecycle stages or qualification;
• workflows are inconsistent;
• AI knowledge sits with only a few people;
• employees are unclear about approved tools;
• governance has not been defined;
• systems remain disconnected;
• nobody owns the AI use case;
• or there is no measurement beyond adoption.

These signals do not mean: “Stop using AI.”

They tell you what needs to become stronger before the next level of AI is introduced.

Frequently Asked Questions