What should your first CRM and AI investment actually improve?

A sales director is considering AI to help account managers prepare for customer meetings. The demonstration looks promising. A briefing appears quickly, the language is clear and the team can imagine using it.

There is still an investment decision to make. What should become better in the working day, what will the team need to change, and what evidence would justify continuing?

For a first CRM and AI initiative, those questions deserve a place in the brief before the technology is selected. A small, well-defined pilot can help a business make its next decision with more useful evidence.

Start with a decision your business needs to make

“We want to use AI in sales” describes an interest. It leaves the team to decide what the investment is for.

A more useful starting point might be: “We want account managers to prepare for renewal conversations with fewer missing customer commitments, while keeping checking effort manageable.” That gives the sales leader a process to examine and a result to assess.

This is Nemely's proposed planning approach. Microsoft also recommends using business processes to define the scope and outcomes of a Dynamics 365 implementation. Its guidance provides a useful reference for discussing the work across business and technology roles. [1]

Choose a recurring situation with a visible owner. Meeting preparation, reviewing a stalled opportunity or capturing the next customer action can be reasonable candidates. Their suitability depends on your team, information and existing systems. The most impressive demonstration is not automatically the most useful first investment.

Compare a few possible starting points

Before committing, put two or three candidate processes alongside each other. For each one, ask:

• Does the problem occur often enough to observe during a pilot?
• Can the people doing the work describe what goes wrong today?
• Is the necessary information accessible and reasonably dependable?
• Can someone review the result before an important customer commitment is made?
• Is there a business owner who can decide what happens after the test?

These questions are a discussion framework, not a validated scoring model. A process can be important but unsuitable for an early pilot if its dependencies are still unclear. Another may offer a better starting point because its boundaries and review steps are already understood.

Use the comparison to choose one process. Keep the other ideas in a backlog so they do not quietly expand the first pilot.

Write down what happens today

A faster draft does not necessarily mean a faster completed task. An account manager may spend less time collecting information but more time checking whether it is current.

Observe several representative examples of the existing process. For meeting preparation, record the information people gather, the time spent gathering and reviewing it, and the important details they still have to confirm. Include the work done after the meeting, such as recording a commitment or assigning a follow-up.

Agree the measurement definitions with the people doing the work. A sales manager and a CRM owner may interpret “preparation completed” differently. Settle that distinction before comparing results.

Do not turn a small sample into a company-wide savings forecast. The first purpose of the baseline is to make the pilot comparison more useful and expose assumptions that need testing.

Define the boundary of the assistance

For an illustrative renewal pilot, the agreed output could be a short briefing containing the opportunity status, relevant open work, recent commitments and questions for the account manager to check.

The pilot brief should also say what remains a person's responsibility. Who checks conflicting dates? Who decides what can be promised to the customer? Who records the next action?

Check the selected Microsoft experience carefully. Microsoft's documentation distinguishes the Sales agent in Microsoft 365 Copilot from Copilot within Dynamics 365 Sales; their availability, capabilities and integrations differ. A pilot should name the exact experience being assessed and confirm its licensing, access and setup for the organisation. [2]

If the first test only needs assistance with preparation, adding autonomous record changes can introduce a separate design decision. Make that decision explicitly, with its own review and exception arrangements.

Include the work around the software

The cost of a pilot extends beyond the tool itself. Build a short preparation list covering configuration, information cleanup, required connections, representative testing, training and ongoing ownership.

Ask who will maintain the information after the first group starts using the process. Ask how a recurring error reaches the person who can correct its source. Give users a practical route to report a problem.

The business sponsor should understand these dependencies before approving a wider rollout. A promising result in one carefully prepared account does not establish that the same approach is ready for every account team.

Agree the next decision before running the test

The following is an illustrative decision sheet. The team supplies its own thresholds and records exceptions; these are not promised outcomes.

The possible next steps should be clear. Continue with the same scope, improve a specific weakness, extend to another group, or stop and choose a different process. A pilot that reveals an unsuitable starting point can still support a useful investment decision.

Give leadership a short recommendation

At the review, explain the original problem, what was tested, what changed, what remains uncertain and what decision you recommend. Include the effort required to operate and improve the process, alongside any benefit observed.

Keep the recommendation proportional to the evidence. “This works for the tested renewal process with these prerequisites” is a more useful conclusion than a broad claim about transforming sales.

Your first CRM and AI investment should help the business learn where a practical improvement is possible and what it takes to sustain it. Start with one useful process, give it an owner and make the next decision clear.

Find a practical starting point with Nemely

If you are weighing up where AI could improve your customer work, bring the process and the questions behind it. Nemely can help assess the readiness, dependencies and scope of a useful first pilot.