Let's be honest - if you're leading a GTM team right now, you're probably drowning in AI point solutions and pressure to "do AI."

You're not alone.

While everyone's talking about AI transformation, few companies have cracked the code on how to actually implement it effectively. The more common situation I see is that revenue leaders are so overwhelmed with the possibilities of AI, that they don’t know where to start.

So, they end up doing nothing.

That's why I'm on a mission to demystify how AI is transforming GTM. I started a few weeks ago with part one of this series, where I dove into the different use cases for AI across sales and RevOps, as well as a framework for how to assess AI fit.

Today, in part two, I'm going to dive deeper into a few of those use cases, talk about real-world examples of tool implementation, and give you a guide on how to get started AI.

Getting started with AI in sales

To get started with AI, you need a clear framework guiding how to launch, evaluate, and scale. Encouraging everyone to find ways to use AI in their workflows is great, but you’ll get utter chaos without a top-down strategy. Conversely, if you only focus on the top-down strategy, you won’t get enough bottom-up buy-in or excitement to use new processes or tools.

The best way to launch AI is a combination of deliberation (top-down strategy) and doing (testing use cases). Recently, Kieran Flanagan shared with us his framework for rolling out AI tools to his team at Hubspot:

Step 1: Deliberate - Set top-down strategic priorities

Aligning on a top-down strategy first helps create guardrails for your test and prevent thrash. Don't worry if you don't have all the answers - start with:

• Define focus areas: Identify your top 5 areas where AI could be transformational
• Set hypotheses, goals, and metrics: Get clear on what you are testing for each focus area. Create goals for each area and how you’ll measure success, i.e. which metrics are you trying to move? Typically the goal is to improve productivity, reduce cost, increase revenue, or improve quality.
• Assign owners: Not groundbreaking but things won’t get done if too many people own a workstream. Instead, opt for DRIs (directly responsible individuals). These owners will become your "tiger team" for AI experimentation.
• Create accountability: Establish the operating rhythm of weekly and monthly check-ins on progress.
• Monitor effectiveness of the tiger team: Regularly check in to ensure the right team is running with each initiative. Ask yourself if this test could see more impact in the hands of a different person or team.

💡 Pro Tip: Kieran emphasizes running this like a growth project, not an IT project. "If you run it like an IT project, it's not going to be successful.”

Step 2: Do - Enable and empower your team

Once you have a strategy set, it’s time to roll out and empower the team to experiment with AI. Key elements of roll-out should include:

• Access & enablement: Give your team access to the necessary tools and provide training on the process. Don’t skip enablement!
• Make it fun & engaging: Organize internal hackathons, create healthy competition with leaderboards, and offer incentives like a bonus or SPIF (especially for sales!).

💡 Pro Tip: Kieran highlights the importance of cross-pollination between teams and encouraging collaboration when testing AI.

🔮 Pocus Tip: Create a weekly sync or async ritual where the team shares how they’ve used AI to streamline a workflow or improve metrics that week.

The top 5 AI tests to run ASAP

#1 Prospecting & research

Your reps have a list of accounts to go after…then what?

Outbound is best when you can reach the right person, at the right time, with the right message. When you have hundreds of accounts in your book, it’s tempting to take a shortcut and send a generic message to everyone or worse a mildly “personalized” message like “Hey, saw you raised a new round of funding.”

AI can help reps prospect faster and more effectively through deeper research done at scale. If reps had the time to create personalized micro-campaigns for every account, they probably would. The best reps are often doing this via deep research like reading 10Ks, blog posts, listening to podcasts, and more to prepare for prospecting.

Large Language Models (LLMs) like ChatGPT and Claude can cut this research from hours to minutes. Reps can synthesize research and get answers to their account questions, making it easier to do customized account planning.

DIY with an LLM using these account research prompts:

• Here’s [Account Name]’s 10K report. Summarize the business overview section of the report and pull out any interesting opportunities for [Your Company]. Here’s some information on [Your Company]’s value props: [2-3 value props]
• What products does [Account Name] sell and what is their business model?
• Extract risks highlighted in [Account Name]’s 10K report.
• Extract a financial summary about [Account Name]. Highlight revenue, recent M&A activity, and any risks.
• Describe [Account Name]’s strategic priorities and how they fit in with [Your Company] products. Here’s a brief description of the products: [2-3 value props]
• Here are a few customers of [Your Company]: [3-4 current customers] How does [Account Name] compare? Are there similarities in the business model or target persona I can use in my pitch?

Don’t stop there! Once your reps have research from an LLM it’s time to figure out how your company fits in by creating a strategic POV.

Your account POV is your hypothesis on why you will deliver the prospect value. It should lie at the sweet spot between what your potential customer cares about and what your product does well. It’s an art and science that balances what you know and your hypothesis on what will resonate.

Manually build a POV for outbound:

1. Come up with 3-4 hypotheses for why your product can help the prospect
2. Write out a few bullet points for each hypothesis and any supporting evidence for why it will be true
3. Search for contacts at the account who would care about each of these hypotheses
4. Craft a tailored email to each of them

🔮 Pocus Tip: Where possible, give the LLM a source to work with and ask it to cite specific content in that source to avoid the hallucination trap.

🔮 Pocus Tip: Change the level of detail or jargon by asking the LLM to explain something to a 7th grader or 3rd grader.  

🔮 Pocus Tip: We don’t recommend you ask an LLM to suggest the right personas to reach out to; the data is often stale. Your reps will spend more time validating the data than if they searched on their own from the start.

Tools to try instead:  

• Pocus generates richer insights compared to prompting your standard LLM. The reason? Insights are powered by our custom AI model, fine-tuned to your business and data. We use our proprietary capabilities to scrape source data like 10K reports, podcasts, YouTube videos, news articles, and transcripts, then merge that with your existing internal data, including sales enablement content.

#2 Generate personalized email sequences or collateral

AI can be a great writing partner - just don’t expect to use the results without editing first! When your team is ready to draft emails or create custom collateral, they can use the same chat in their LLM tool of choice where they started their research. This will help the LLM use that information as context when writing the content.

DIY with an LLM to create content using the following prompts:

• Email: Pretend you are a salesperson at [Your Company] and you are writing an email to [Persona at Account Name]. Based on the strategic priorities of the account [summarize them], write a short email that is no more than 2 paragraphs long.
• Customize collateral: Upload a piece of collateral you want to customize, then prompt the LLM with “Use the strategic priorities of [Account Name] to personalize this content. Use the existing style and voice of the content in this transcript and this transcript alone.”

Tools to try instead:

• Lavender for writing sales emails that convert. Lavender's scoring is a helpful assistant that can coach you in real time to improve your content.
• Tofu for creating personalized collateral at scale, including 1 pagers, landing pages, and more based on research you provide (or they can generate it for you).

🔮 Pocus Tip: Create a “style guide” per piece of content. For example, upload 3 or 4 of your best-converting emails and ask the LLM to create a prompt that describes the style and voice of those emails. Ask the LLM to write another prompt that describes why it is a good email. Use these 2 prompts as your “style guide” moving forward.

🔮 Pocus Tip: Using an LLM for content creation is a multi-step process. For best results, break down the content into components and ask the LLM for assistance in writing each section versus writing an entire piece.

#3 Meeting notes --> business plan

Call recording tools do a great job of helping reps summarize content and actions from individual calls, but often reps are stitching together context from across multiple meetings. Aggregating meeting notes to generate content that helps build value or a business case down the funnel can be a huge unlock for AEs.  

Use an LLM to summarize your calls:

1. Upload the audio of your call recording, then prompt the LLM with: “Transcribe this call recording between [Your Role at Your Company] and [Contact’s Role at Account Name].”
2. To generate a summary and action items, follow up with this prompt: “Summarize the content of this call transcript into 1 paragraph. List out any action items. Extract any objections mentioned by [Account Name] as a direct quote”

Tools to try instead:

• Toolflow can create custom AI workflows to manage tasks like transcribing meeting notes, pulling out key themes, and listing action items.

• Gong call recording transcribes your conversations, creating auto-generated summaries and highlighting keywords, like mentions of pain points or use cases).

#4 Call prep

Call prep is one of the easiest and most effective ways for your sales team to incorporate AI into their daily workflows. GenAI is great at summarizing and synthesizing information, which can save your team hours. This use case gets even more powerful when you layer in context and structure around what type of call you’re joining and what questions you need answered.  

DIY with an LLM ahead of your next call:

1. Upload your notes about the customer or any other relevant information
2. Prompt your LLM of choice with this: “Using these notes and these notes alone give me a summary of this account. Include a list of their priorities, if available, and any timing mentioned, and extract key personas mentioned.”
3. Ask the LLM for any publicly available information relevant to your target account. You might want more generic insights or something specific.
4. Use this follow-up prompt “Find me information on [Account Name]. I want to understand the business's strategic priorities and how they fit in with my company, [Your Company]’s products. Here’s a brief description of our value proposition: [Include 2-3 points about what value you offer]”

Tools to try instead:

• Pocus: Streamline the workflow by just using Pocus. Try Pocus AI Strategy to quickly prep for any meeting in minutes instead of hours. Pull up AI strategy directly from your Google calendar or in the Chrome extension and get robust insights ahead of your meeting.

Getting started with AI doesn’t have to be overwhelming. Whether your team is diving into prospecting, automating outbound emails, or just trying to make sense of all their customer conversations, now is a great time to start experimenting with tools and prompts.

Remember, the goal here is to make reps’ lives easier and help them crush their numbers. Whether you're starting with simple ChatGPT prompts or jumping into specialized tools, start small, see what moves the needle, and scale from there. Your team will thank you for helping them spend less time on busy work and more time closing deals.