Where AI fits in your sales process

Where Can AI Fit in Your Sales Process?

Start with what happens now.

You do not need to know whether you need an AI agent. You do not need to choose an AI platform. And you definitely do not need to automate everything.

Start with your sales process. Follow a lead from the moment it appears to the moment it becomes a customer, goes quiet or is closed.

At each stage, ask: What has to happen here? Who does it? What information do they need? What gets delayed, repeated or forgotten?

That is how you find where AI might actually be useful.

Find your starting point See how we work
Context moving down a sales pipeline through stage gates
Follow the lead

Where does the work get stuck?

Most sales processes contain some version of these stages:

Enquiry→ Understand→ Qualify→ Route→ Prepare→ Converse→ Record→ Follow up→ Progress→ Handover

Your process might use different words. It might skip some stages. It might have another ten in between. That is fine. The useful part is looking at each stage separately instead of asking whether you should "use AI in sales". Let's walk through it.

01

A lead or enquiry arrives

Does somebody have to notice it?

Your leads might arrive through: Website forms. Email. A shared inbox. Referrals. Campaigns. Events. Phone calls. CRM forms. Other sales channels.

Ask yourself
  • Does every genuine enquiry reliably reach somebody?
  • Are enquiries sitting alongside unrelated emails?
  • Does somebody manually copy information into another system?
  • Do different lead sources follow completely different processes?
  • Does anybody have to check somewhere regularly to see whether something has arrived?
  • Do enquiries sometimes wait because everyone assumes somebody else has seen them?
AI might help if: The enquiry needs to be understood before the next step can happen.
Automation might be enough if: The information simply needs moving from one known place to another.
Possible opportunity: Enquiry handling
02

Understand what they want

Can you tell what the enquiry is actually about?

Forms are structured. People are not. An enquiry might say: "I saw what you did for another company and wondered whether you could help us with something similar. We currently use X but we're having problems with Y." Somebody has to understand that. What are they asking for? Which service might be relevant? What information have they provided? What is missing? Does something need immediate attention?

Ask yourself
  • Does somebody manually read and categorise every enquiry?
  • Do enquiries need interpreting before they can be routed?
  • Does your team repeatedly extract the same pieces of information from emails?
  • Do you receive long or unstructured requests?
AI might help with: Classification. Information extraction. Summarisation. Identifying missing information. Preparing the next step.
Possible opportunity: AI enquiry understanding
03

Check what you already know

Is the information already somewhere in the business?

Before treating somebody as a completely new lead, you may need to know: Are they already in the CRM? Has somebody spoken to them before? Is their company already a customer? Is there an existing opportunity? Are there relevant previous emails? Is there information another team already knows?

Ask yourself
  • Does somebody search several systems before responding?
  • Do people regularly miss useful existing context?
  • Are duplicate records common?
  • Does the answer depend on who happens to recognise the name?
AI or automation might help with: Finding existing records. Gathering permitted context. Preparing account history. Identifying possible matches. Bringing useful information together.
Possible opportunity: Context gathering
04

Qualify the opportunity

What makes something worth pursuing?

Qualification should come from your business, not from AI. What matters to you? Service fit? Requirements? Location? Timing? Company type? Commercial criteria? Existing relationship? Something particular to your industry?

Ask yourself
  • Does everyone qualify leads in roughly the same way?
  • Are the criteria actually defined?
  • Does qualification depend on information that already exists elsewhere?
  • Does somebody repeatedly ask the same initial questions?
  • Are leads given arbitrary scores nobody really understands?
AI might help with: Comparing information against agreed criteria. Identifying what is missing. Preparing clarification questions. Recognising when a situation needs human judgement.
But first: If your qualification rules are unclear, clarify them. Do not automate a decision nobody has defined.
Possible opportunity: AI lead qualification
05

Decide who or what handles it

Does every lead know where to go next?

Once an enquiry is understood, somebody or something needs to decide what happens. Which salesperson? Which team? Which workflow? Does it need a response? Does it need specialist review? Does it need more information?

Ask yourself
  • Are leads manually forwarded?
  • Do people regularly receive enquiries that should have gone elsewhere?
  • Does routing depend on somebody understanding the content first?
  • Are unusual enquiries easy to identify?
  • Does the person receiving the lead get enough context?
AI might help with: Selecting between approved routes. Preparing a handover. Identifying exceptions. Gathering context for the recipient.
Automation may be enough when: The routing rule is completely predictable.
Possible opportunity: Intelligent lead routing
06

Prepare for the conversation

How much work happens before somebody actually speaks to the customer?

Before a sales call, someone may check: The CRM. Previous emails. The company website. Old meeting notes. Open opportunities. Outstanding actions. Documents. Internal information.

Ask yourself
  • How long does meeting preparation take?
  • Does your team repeatedly search the same systems?
  • Is useful information scattered across several places?
  • Do people sometimes go into conversations without context because they did not have time to find it?
AI might help with: Gathering permitted information. Summarising account history. Surfacing previous commitments. Preparing a meeting brief. Identifying unanswered questions.
Possible opportunity: AI meeting preparation
07

Have the conversation

This is often where the person matters most.

Your salesperson understands the customer. Asks questions. Listens. Explores the problem. Uses judgement. Builds the relationship. Responds to something unexpected. Negotiates. Makes commercial decisions.

Not every stage of the sales process is waiting to be automated.

Ask yourself
  • Does AI genuinely improve this conversation?
  • Or would it simply get in the way?

Sometimes the right answer is: Leave this bit alone.

AI can still help before and after the conversation without needing to become the conversation.

08

Record what happened

Where does the useful information go?

After the conversation, somebody may need to: Write notes. Update the CRM. Record requirements. Change an opportunity stage. Create tasks. Record the next action. Tell another person what happened.

Ask yourself
  • How much time does this take?
  • Does it happen consistently?
  • Is important information buried in free text?
  • Do people postpone CRM updates until later?
  • Does useful context disappear because nobody has time to record it properly?
AI might help with: Structuring meeting information. Preparing CRM notes. Extracting agreed actions. Suggesting updates. Creating or preparing tasks. Preparing internal handovers.
Possible opportunity: AI CRM automation
09

Make sure the next action happens

What did everyone agree to do?

This is where small gaps become lost opportunities. Someone promised to send information. A proposal needs preparing. The prospect asked a question. Another meeting needs booking. Somebody needs to check internally. A follow-up should happen next week.

Ask yourself
  • Is every next action recorded?
  • Does it have an owner?
  • Does it have enough context?
  • Does follow-up rely on people remembering?
  • Can somebody easily see what the customer is waiting for?
AI and automation might help with: Identifying commitments. Preparing tasks. Tracking agreed next actions. Preparing follow-up. Surfacing overdue actions.
Possible opportunity: Next action management
10

Follow up

Do you know why you are following up?

There is a big difference between: "It has been seven days. Send email number three." and: "We sent the information they requested last Tuesday and agreed to check back this week." The second has context.

Ask yourself
  • Is follow-up based purely on time?
  • Does somebody need to reconstruct the previous conversation before contacting a lead?
  • Are generic sequences being used where a more relevant response would be better?
  • Do opportunities sometimes go quiet simply because nobody notices?
AI might help with: Gathering previous context. Identifying what is outstanding. Recommending a next step. Preparing appropriate follow-up.
Human review may still matter when: The relationship or situation is sensitive, unusual or commercially important.
Possible opportunity: Contextual follow-up
11

Notice when something has stalled

Who is watching the process?

Not every problem creates an alert. Sometimes nothing happens. That is the problem. A lead has no owner. An opportunity has no next action. A customer is waiting for information. A follow-up is overdue. A meeting happened but the CRM was never updated. An opportunity has stopped moving.

Ask yourself
  • How do you find these situations today?
  • Does somebody check dashboards?
  • Run reports?
  • Look through the CRM?
  • Notice only when the customer contacts you again?
An AI agent might help by: Watching for agreed conditions. Gathering the relevant context. Surfacing exceptions. Recommending an action. Escalating something to the appropriate person.
Sometimes the most useful agent does not sell anything. It notices.
Possible opportunity: Sales attention agent
12

Move the opportunity between people

Does the context move too?

Sales processes often involve handovers. Marketing to sales. Founder to salesperson. Sales development to account executive. Sales to technical team. Sales to operations. Sales to account management.

Ask yourself
  • Does the next person understand what has already happened?
  • Do they have to search through emails and notes?
  • Does the customer have to repeat themselves?
  • Does important context live inside somebody else's head?
AI might help with: Preparing structured handovers. Summarising relevant history. Identifying outstanding commitments. Gathering information from permitted systems. Highlighting what the next person needs to know.
Possible opportunity: AI sales handover
13

Learn from the process

What keeps happening?

Once sales activity is recorded consistently, AI can potentially help analyse it. Not to create another dashboard full of numbers. To answer useful questions. Where are leads getting stuck? Which parts of the process create repeated manual work? Where are exceptions happening? Which information is frequently missing? Where does the team repeatedly override the workflow? Which steps generate the most administration? Where are handovers failing?

Those patterns can help you decide what to improve next.

Possible opportunity: Sales process intelligence
So where should you start?

Look for these five signals.

You probably have a workflow worth examining when something is:

Repeated

Someone does essentially the same piece of work again and again.

Manual

Information is repeatedly copied, reformatted or moved between systems.

Forgotten

The process depends on somebody remembering to check, chase or update something.

Scattered

The information needed for one decision lives in several places.

Interpretive

Somebody repeatedly has to read information and decide which known next step applies.

That does not automatically mean AI is the answer. But it is a good place to investigate.

And watch for these five warning signs

AI may not be the first thing you need.

Nobody agrees on the process.

Clarify it first.

The underlying information cannot be trusted.

Fix the data problem.

A simple automation already solves it.

Use the automation.

The task is rare and highly unusual.

Keeping it human may be easier.

The decision carries consequences that justify human judgement.

Keep the person involved.

The objective is not to find somewhere to put AI. It is to improve the sales process.

Your AI Sales Process Check

Run through these questions with your team.

EnquiriesDo all genuine enquiries reliably reach the right place? Does someone have to manually understand and categorise them?
ContextDoes your team repeatedly search several systems before responding?
QualificationAre your qualification criteria clear and consistently applied?
RoutingDo leads reach the right person with enough information attached?
PreparationDoes meeting preparation involve repetitive research or searching?
CRMAre people manually recreating information that already exists elsewhere?
Next actionsDoes every active opportunity have a clear owner and next step?
Follow-upDoes follow-up depend too heavily on memory?
AttentionCan you easily identify the opportunities that need somebody today?
HandoversDoes useful context survive when work moves between people?

If several of those questions made you think: "Not really." That is where we start.

What happens next?

Bring us the part that is not working.

You do not need to diagnose the technology. Tell us: "Website enquiries are taking too long to sort out." Or: "Our salespeople spend ages updating the CRM." Or: "Follow-up is inconsistent." Or: "We have information everywhere and nobody can see what needs attention." Or simply: "This part of the process is a mess."

We can map what happens, work out what should change and determine whether the right answer is a person, automation, AI, an agent or a combination.

See where AI fits How we work
Frequently asked questions

Where to start with AI in sales.

Where should I start using AI in sales?
Start with a specific sales process rather than an AI tool. Look for repeated work, manual information movement, missed actions, scattered information or tasks where somebody repeatedly has to interpret information before choosing a known next step.
Do I need an AI agent?
Not necessarily. Some sales problems are better solved with conventional automation, changes to the existing process or functionality already available in your software. An agent is useful when AI needs responsibility for a defined piece of work involving understanding, decisions or actions.
Should we automate our whole sales process?
That is rarely a useful starting point. Choose one clearly defined workflow, understand it properly and improve it before deciding what else should change.
What sales tasks can AI help with?
Potential areas include enquiry understanding, lead qualification, routing, meeting preparation, CRM administration, follow-up, handovers and identifying opportunities that need attention. Whether AI is appropriate depends on the specific task and business.
How do we decide what should stay human?
Consider the judgement, relationship, sensitivity and consequences involved. Negotiation, unusual situations, significant commitments and sensitive customer interactions may justify keeping a person directly involved.
What if we already have sales automation?
Keep the automation that works. AI does not need to replace predictable rule based workflows. It can be added where interpretation or contextual understanding is genuinely useful.
Can you look at our existing process and tell us where AI fits?
Yes. That is the point of starting with the process rather than a predetermined AI product. We map what happens, identify friction and determine which parts may benefit from process changes, automation, AI or an agent.
Do not start with AI.

Start with the annoying bit.

The enquiry someone has to sort manually. The CRM update nobody wants to do. The meeting that takes twenty minutes to prepare for. The follow-up somebody has to remember. The lead sitting without a next action. The information spread across five places.

Find that. Then we can work out what should happen next.

Find where AI fits Explore Agent Console HQ