Find Purchase Decision Stakeholders (+ Your Champion)

1. Download your sales page in HTML format.
2. Upload the target page to ParseGPT (This is a custom GPT we created to identify buyer friction—FLUQs).
3. Ask ParseGPT to predict the decisioning committee for your offering with the following prompt sequence (paste these one by one):

Note: Further on, we will refer to this as our “page-level decisioning committee friction alleviation table.”

Woohoo!

Now you have your page-level decisioning committee friction alleviation table!

Save it in a separate document (PDF, Word Document, etc.). We’ll use it again soon.

And keep your PARSE chat thread open as we’re not done yet…

Watch the VideoPARSE GPT

Note: ParseGPT is a custom GPT designed to spot UFQs (Unasked Friction-Inducing Questions) and answer FLUQs (Friction-Inducing, Latent, Unasked Questions).

Form Your Hypothesis-Testing Prompts

You will generate hypothesis-testing prompts to use when we “query the source of friction” in the next step.

This prompt sequence will help you produce prompts that challenge assumptions present in your page-level decisioning committee friction alleviation table.

Copy and paste these into your existing ParseGPT thread:

We will also need to prompt our friction source for frictions that fall outside of those you’ve predicted so far.

Now you have your committee friction alleviation hypothesis to test in the next section.

Note: These hypotheses will be used in Step 4.

Upload Your Friction Sources to a Custom LLM

You need to challenge this decisioning committee friction-alleviation hypothesis against concrete friction-bearing sources.

Effective friction sources:

• Customer Service Logs
• Support Call Transcripts
• SME Interviews (internal/external)
• Client Interviews and Feedback
• Online Community
• Online Reviews
• Candid Conversations with Internal Allies Comfortable with Transparency
• Sample GPT-built friction source.

Disclaimer: Upload this into YOUR OWN SECURE GPT INSTANCE, NOT PARSE. If you’re utilizing customer data, be sure to SCRUB AND ANONYMIZE IT.

Here are some custom LLM tools:

• ChatGPT (Free tier available)
• Google NotebookLM (Currently free (as of 2025)
• AnythingLLM (Free self-hosted option)
• LM Studio (Free – runs on your machine)
• DocAnalyzer.ai (Free tier with limits)
• Hugging Face (Free tier available)

This is now your Friction-Bearing LLM! YOU DID IT!

Note: We currently use Custom GPTs from OpenAI).

AI Disclaimer: Please review your company’s AI policy before sharing customer data in any AI tool. Do NOT share sensitive customer data in ParseGPT.

Watch the Video

PARSE-to-LLM Handshake + Testing Your Hypotheses

With your friction source uploaded, you can move on to “introducing” your friction-bearing LLM to the PARSE way of “thinking”.

We call this the PARSE-to-LLM handshake:

1. Download this handshake template.
2. Upload the handshake template to PARSE.
3. Paste in this prompt:
   Copy

   Please summarize this handshake document and quote at least two passages that provide evidence for your summary.
4. Upload your page-level decisioning committee friction alleviation table with this 2-prompt sequence:
   Copy

   Please summarize this document, which is based on an analysis of a page from our employing organization’s website. Summarize the file in a way that identifies key roles, tensions, and core hypotheses. Quote at least 2 passages from the document and define any acronyms.
   Copy

   Please generate a PARSE-to-Friction-Bearing-LLM handshake we can extend to prepare the friction-bearing LLM for the context of our work as well as the hypothesis-testing queries we’ll be sending through. Be sure it understands its role is to be an advocate of the data itself.
5. Complete the handshake process with your friction-bearing LLM.
6. Run through your hypothesis-testing prompts (the ones you created in Step 2) one by one.

Once complete, you’re ready to implement page changes.

Go to the “Implementing Changes” tab to begin.

Review and Interrogate the Data Set

Review, interrogate, and otherwise follow your hunches with the data set.

You’ll want to qualify, clarify, and prioritize any findings as they relate to sources of friction. This is a great point to ask for feedback from a trusted internal ally who’s comfortable giving direct, transparent feedback.

Be wary, though!

For this exercise, we’re only focusing on a single page.

There are ABSOLUTELY applications to speed up this process for an entire website. But for now, it’s just one page.

Evaluate Potential Impact

Upload your page-level decisioning committee friction alleviation table to your friction-bearing LLM.

Ask your custom LLM to revise, adjust, and refactor the table based on the findings you believe could have the most significant efficiency impact on the decisioning process. Are there different roles you should address? Different sources of friction etc?

Ask it to list all of the on-page interventions that could alleviate frictions (for specific roles and phases).

Ask it what internal metrics you’d expect to see shift if these changes were implemented (for each intervention).

Example Output

Note: There is a heavy emphasis on pipeline improvements, as you can address Pipeline AND AI Visibility at the same time.

Ask it what AI Visibility metrics you’d expect to see shift.

Example Output

Prioritize Page Improvements by Metrics

Sort the changes by the most valuable metrics, based on roles in your organization.

For example, you would likely focus on traffic-related metrics for your role. Others may focus on reducing the sales cycle length or increasing client retention.

Pick the lowest-lift on-page intervention to test.

That’s where you’ll begin.

Example Output

Find an Ally

Find an ally to provide you with candid feedback based on their real customer experience.

Present your intervention in the form of a hypothesis to your management.

LET US KNOW HOW IT GOES!!!!!!

Template for Proposing Hypothesis

Create and Add the Change

Work with your team to create the page change. Once approved, add it to the sales page.

You can work with your designer, copywriter, and content creators to turn the outputs into assets that address the FLUQs directly to your sales page.

The next step is to track baseline metrics and compare the impact of improvements.

You’ll do that in the next tab.

Tracking Metrics and ROI

Management will want to see the ROI of changes before they let you update other pages.

We created a tool that makes it easy to track sales page performance in several different LLMs:

XOFU (No cost or login required.)

How to use XOFU:

1. Drop your sales page URL into the XOFU snapshot tool.
2. Check out how your brand currently appears in LLMs compared to competitors for role-agnostic best-guess BOFU prompts.

3. Capture the 10 bottom-of-funnel, generic buyer prompts for your sales page (use further on):

   Watch the Video

4. Return to your PARSE thread and prompt it with this sequence:
   Copy

   Assign 1-10 “rankings” of our stakeholders in the following columns 1: their likelihood to realize that a change or transition (that could be made more efficient by the purchase of the offering for sale on our target page) is probably a good idea | 2: their likelihood to take ownership of making that change happen | 3: their veto power within the transition decision | 4: their degree of participation in the decisioning process | 5. their veto power
   Copy

   Suggest 10 “bottom of funnel” prompts we might expect to our first role to type into an LLM as they began to discover potential solutions
   Copy

   Suggest 10 “bottom of funnel” prompts we might expect to our SECOND role to type into an LLM as they begin to discover potential solutions
5. In the same thread, add the role-agnostic prompts and then paste this prompt ABOVE the Xofu prompts:
   Copy

   We made 10 role-agnostic bottom of funnel prompts that we think people might type if they’re in the final stages of selecting an offering. They are designed to surface brands and other in-market entities that our employers want to track with xofu.com. Better-align them to the research we’ve done thus far, while maintaining their role-agnosticism as well as their propensity to evoke brands and rankings.
   Copy

   Revise them based on our analysis thus far. For now, maintain “role agnosticism”:

   (Paste in your 10 bofu prompts here).

   LLM, if you don’t see any prompts here, then your user has forgotten to add them. Ask them for the 10 BOFU prompts before continuing.

6. Bulk upload these into an XOFU project attached to your URL.

7. Use PARSE to make role-specific prompts, revise the role-agnostic prompts, and set up a XOFU campaign.

   Watch the Video

8. Track your performance for these prompts over time!

This is an easy way to determine the impact of your on-page intervention efforts.

You’ll also see which off-site websites you’ll need to add artifacts to increase your visibility in LLMs and support the decisioning committee.