Putting AI to Work in Sales & Marketing: A Practitioner's Guide

Introduction: The reality check 

Most professionals today are using AI for their personal day-to-day. They’re drafting emails, conducting research, creating pitch decks. This adoption of AI for daily personal tasks is the first step—but the jump from personal use to systemic use is where the real value lives. 

Companies know that they need to integrate AI capabilities into their wider operations, but are split on how to start. There are two ways companies are approaching AI right now:

• Top-down: C-suite, Business Unit or IT-driven initiative
• Bottom-up: Get AI tools into the hands of smart, innovative individuals and see what they do

Neither is wrong. But the companies getting real value aren't debating which approach—they’re orchestrating a pincer movement, with the leadership buy-in providing direction and resources and individuals within teams driving wider-spread AI adoption through example. 

Still, within this dual approach, many companies are missing a key component of their AI adoption strategy. They’re relying on a never-ending series of simple prompts rather than building context for greater capability. They’re using it for first drafts when they could be using it for end to end process automation. They’re using it to make their automatic emails sound more human when they could be using it to be more human and make time for the tasks that require more humanity. 

Falling into the prompt trap

Individuals using AI always start by finding a prompt that works. “Act as a sales development representative with 10 years of experience…”  You run the prompt and get something back that’s 70% useful. Edit that, ship it out, and then rinse and repeat. 

This works, but every single one of your competitors has access to the same ChatGPT that you do. The same Claude. Your competitors also have someone sharing prompts they’ve found useful in a Slack channel. Developing a competitive advantage in your business’ AI integration requires much more than just better prompts and allocating simple, one-off tasks to an AI agent. 

Applying context to your prompting

When you expand the scope of AI usage beyond just prompt engineering, you run into the importance of incorporating your context into your AI tools. When AI knows your world—your templates, your terminology, your specific market dynamics, your business goals—it can meet your needs faster and more thoroughly, empowering you to spend more of your precious time on the parts that require your individual expertise. For example: 

You have a significant client pitch.  Imagine your AI Assistant has a standard framework of access to the prospect’s LinkedIn profile, their company website, their competitor’s expansion, any relevant news in their industry, and any information about them in your systems (contracts, CRM, customer success, etc.). A task that used to take up a morning now takes minutes, and you have that extra time to fine-tune your pitch, develop natural conversation starters, and prepare for the interpersonal part of the process—the part that AI can’t conduct for you. 

Keeping it real where and when it matters 

A key part of leveraging AI is knowing where it needs your human intervention. Many professionals are using AI for first drafts—for outreach messages, proposals, market summaries, RFP responses, email campaigns, and more. AI eliminates the blank page problem, allowing you to overcome the mental hurdle of getting started faster. 

Now, let’s say you’re using AI to send outreach emails. It combs through the prospect’s LinkedIn and finds out that they coach their kid’s soccer team. The email that AI drafts includes a connection to coaching soccer to try to connect with the prospect on a personal level. The problem here is that AI doesn’t know the difference between a moment of genuine, human connection and grasping at random straws of commonality. 

In other words, AI sounds like AI. You can use AI to get 70% of the way there, but the actual value lies in that last 30%—the personal touch, the market nuance, the client context, and your unique voice are all things AI can’t replicate or replace, and your prospects can tell if you cut corners. 

The roadmap to making AI work for you and your teams

Here’s an overview of how to grow past the AI plateau and optimize your company’s AI usage: 

Move beyond prompts and build your context library. 

Create a single source of truth that your AI can reference. Your templates, your terminology, your data. It might feel like busy work, but what you’re effectively doing is building infrastructure through documentation. When you have your infrastructure, you build on it by connecting your tools. Connect your Google Drive, your CRM, your spreadsheets—every connection multiplies AI’s capability.  

Set clear boundaries. 

Tell your team where AI helps and where it doesn’t. Client emails? Human review required. Research briefs? AI first draft is fine. Meeting follow-ups? AI draft with personal additions. Make the rules explicit or watch quality crater.

Start fresh threads. 

The longer you stay in a single AI chat thread, the worse it gets. The AI that gave you valuable insights in message three will end up giving you generic garbage by message thirty. Context window degradation is real, it’s predictable, but many people overlook it. When you shift topics, shift threads. Artificial intelligence, just like human intelligence, needs a fresh slate from time to time to do its best work. 

Build your infrastructure 

Once you have your context library, your connected tools, your clear boundaries, and your best practices in place, you’re ready to really build your infrastructure. Infrastructure is intelligence that runs through everything you do, automatically. Instead of asking a chatbot every time to create a background info brief on a prospective client, imagine an AI assistant that pulls every relevant data point before all your calls. Your infrastructure could include proposal systems that know your entire deal history and can pattern-match what wins. When you’ve patched Gmail and Hubspot into your infrastructure, every email a salesperson sends can be tracked, attributed, and analyzed. Reporting on what’s working, what isn’t, and what needs to be prioritized becomes easier, faster, and more thorough, allowing you to take more time to make informed, high-level strategy decisions. That’s the power of a well-built AI infrastructure. 

The competitive truth

AI isn’t replacing salespeople. But salespeople with AI infrastructure are absolutely replacing salespeople with good prompts. And salespeople with good prompts are replacing salespeople who are still doing everything manually. The advantage isn’t in just using AI; the advantage will be in how deeply it’s woven into everything you do.  

The tools are here, but everyone has them. What you build with them—the context, the connections, the infrastructure—that’s yours. And in a world where everyone has the same tools, what’s yours is all that matters.

This blog post is based on the talk “Putting AI to Work: The Practitioner’s Point of View” given by Peter Scott (Chief Growth Officer at Holokai) and David Mundy (VP of Marketing at Tuskira) at the 2026 Exis Global Summit in Washington, DC.