# GP Abhishek

B2B marketer sharing work on brand building, demand generation, executive content, and AI for go-to-market teams, including marketing at Everstage.

- Profile: https://b2binfluencer.marketing/creators/gp-abhishek/
- LinkedIn: https://www.linkedin.com/in/abhishekgpmarketing
- Median of last 9 posts: 36 reactions, 3 comments (checked 2026-11-02)
- Topics: artificial-intelligence, b2b-marketing, revenue-operations, demand-generation, positioning
- Based on 9 dated posts, latest 2026-09-30

## Summary

Most useful for concrete B2B brand and executive content tests, with candid evidence about what did and did not work.

- Best for: Brands selling sales, revenue operations, or marketing software into a sales-led motion, planning multi-quarter brand work.
- Not for: A paid sponsored-post package with guaranteed impressions, published demographics, fixed rates, and campaign flight reporting.
- Check first: Confirm whether they offer sponsored placements and can provide audience demographics and comparable results.

### Takeaways

- Only one of three executives moved branded search, and six of 21 high-performing posts did so.
- One executive's 1.5 million views left brand search flat because audience lacked budget or deal influence.
- Before copying a neighboring category's feature, ask whether another product owns the buyer's word or habit.

## Main ideas

- **Brand is a demand lever, and the bill comes before the proof**: The argument is that buyers form opinions about a company before it ever appears in a CRM, so a team can optimise the demo form and still lose because it never made the shortlist. GP Abhishek says that before joining Everstage they agreed with Siva that brand would be treated as a demand lever, and that the hard part is not the agreement but keeping the funding when someone asks which campaign created the opportunity. The post also states plainly that brand still has to be proven, and that a quarterly reporting cycle cannot capture everything that influences a buying decision.
- **Judge executive content by branded search, not by reach**: Rather than count likes or impressions, GP Abhishek describes building a scoreboard from the company's branded Google ad, which appears on roughly 98% of US searches for the name and filters out bots, so daily impressions approximate how many people actually typed the company name that day. Ad spend was held flat and bot spikes discarded so the post was the only variable, with the week after a post compared against the two weeks before. Over 8–9 months, three executives and 80-plus posts that reached at least four times that executive's usual impressions, only one executive moved brand search at all. GP Abhishek calls the setup not the most statistically rigorous holdout group but says it does not get much cleaner.
- **Aim at the people who can buy, not the people who will applaud**: GP Abhishek argues that most LinkedIn executive playbooks assume the people reading the posts are the people buying the product, and that in B2B this almost never holds. The CFO signing the contract is not hanging out in growth communities, but the moment a sales rep emails them they will search the company name, and that search is where the deal is won or lost. The practical conclusion is unfashionable: treat top of funnel as the target accounts the sales team is calling next quarter, and skip the Telegram channels, paid groups and growth hacks.
- **AI moved the leverage from drafting to editing**: Responding to Ryan Law's writing on AI in content marketing, including the failures, GP Abhishek argues that AI drafts are finally good enough and the real leverage has moved downstream, to the edit — the metric GP Abhishek tracks as a proxy for progress. The method: run AI on first drafts primed with skill files loaded with audience, brand voice and best examples before the model writes; have every editor log each change with a one-line reason for 30 days; group the reasons; push anything appearing in more than 30% of drafts into the core instructions; then re-measure the following month. The stated read on the numbers is blunt: zero edits means nobody read it closely, full rewrites mean the prompt is broken rather than the writer, and a drop in repeat edit patterns means the system is learning. In GP Abhishek's experience about 90% of bad AI output is a bad instruction problem.
- **AI sees the problem, but teams still do the homework**: GP Abhishek reports a survey of 195 revenue operators for a Revenue Execution 2026 report: more than three out of four SaaS and AI-native revenue teams are nowhere near the agentic, ontology and context-layer conversation happening on X, every team surveyed runs AI, but only 23% do anything about what happens after the AI tells them something. The first half is treated as solved — intelligence arriving in real time — while everything then lands in Slack as a to-do for a human, which GP Abhishek calls a firehose of homework. The suggested starting point is deliberately small: take your three loudest AI alerts and ask who acts on it, and how many manual steps sit between the alert and the action.
- **Category lines were build cost, so check who already owns the habit**: Watching two AI products he pays for both claim the fn key on his Mac, GP Abhishek argues that most category lines were build cost dressed up as strategy, and that this cost is now heading to zero because it became cheap to build a neighbour's product. The fn key becomes the physical form of a borrowed framework, the Law of Exclusivity from Ries and Trout's The 22 Immutable Laws of Marketing: two companies cannot own the same word in the prospect's mind. From that comes the fn key test before anything from a neighbouring category reaches the roadmap — does the new thing need a word or a habit that someone else already owns in your buyer's head? If yes, the roadmap is fine but the positioning is in trouble; if no, go build.
- **One specific stunt per event beats a booth and a brochure**: Across several Dreamforce and event posts, the pattern is one defined job or one physical object instead of bought reach: give a single person a single task and let the obstacle become the content, or mail something to a narrow segment with the ask folded into the object.
- **The 90-day demand gen plan no longer tests the candidate**: GP Abhishek calls "build me a 90-day demand gen plan" the worst take-home you can give a growth marketer in 2026, because every model writes a passable one in four minutes and the interviewer ends up grading the model rather than the candidate. Instead of a replacement exercise of his own, GP Abhishek points to the playbook Andy Mowat crowdsourced this week on testing AI fluency in RevOps hires, and says he is using an abridged version of it for growth marketers in a sales-led motion. The post ends by asking readers what they would plant in the brief.

## Work to read

- [Treating brand as a demand lever, and paying for it before the pipeline proof exists](https://www.linkedin.com/posts/abhishekgpmarketing_b2b-marketing-has-a-strange-expectation-activity-7511080339260235776-dr-s)
- [Testing whether executive LinkedIn posts actually move brand search](https://www.linkedin.com/posts/abhishekgpmarketing_everyone-posts-their-exec-led-linkedin-wins-activity-7518319920686899201-yytY)
- [Choosing the people who can buy over the people who will applaud](https://www.linkedin.com/posts/abhishekgpmarketing_the-hard-part-about-linkedin-exec-presence-activity-7516140316391837696-6jd7)
- [The last mile of AI: teams that see more without doing more](https://www.linkedin.com/posts/abhishekgpmarketing_if-youve-been-on-x-over-the-last-12-weeks-activity-7519406062433624065-VZVl)
- [A direct-mail stunt built on a spreadsheet's obituary](https://www.linkedin.com/posts/abhishekgpmarketing_revops-n-value-activity-7508178627855925248-hh27)
- [Judging AI-assisted content by its edit rate, not its draft quality](https://www.linkedin.com/posts/abhishekgpmarketing_ryan-law-over-at-ahrefs-has-been-publishing-activity-7516476654597115904-3Q4x)
- [The "fn key test" for AI products colliding in the same category](https://www.linkedin.com/posts/abhishekgpmarketing_two-ai-products-i-like-competing-for-one-activity-7521826412404809728-srsf)

## Brand relationships

- **Everstage** (employment)
- **Brandbeat by billups** (mention)

## Partnerships

The material supplied for GP Abhishek documents marketing campaigns they built for their employer, not a published sponsorship offer. Across the ten posts reviewed there is no rate card, media kit, audience or deliverable specification, disclosure policy, or booking route for sponsored placements. A brand that wants to work with them has no documented self-serve path from this evidence, so any arrangement would have to be negotiated directly — most plausibly through the LinkedIn profile listed in the identity section. The posts do show the brand-side work they plan and ship, which is useful when proposing a collaboration: a full rebrand, an owned summit where an outside host got more stage time than the company CEO, a Times Square billboard reacting to a LinkedIn comment, trade-show merchandise, a 3D-printed direct-mail campaign aimed at revenue-operations teams, a creator-shot video at Dreamforce, an exec-led LinkedIn program measured against branded search, and AI-assisted first drafts for survey reports and thought leadership. Those are examples of campaigns GP Abhishek has run, not formats they have published for sale. One distinction matters for planning. The reviewed material shows no sponsorship offer, but it does not show that GP Abhishek never accepts sponsorships; the question is simply not addressed. Before a paid placement is budgeted, collect primary evidence such as a profile or media-kit description of partnership options, a disclosed sponsored post, or a direct written answer about availability, rates, and expected deliverables. Treat that as an open research gap, not a settled "no."

## Alternatives

- Olivia Fuller: Choose this alternative when the campaign's audience is one-person businesses and very small service firms, and the job is patient, plain-language education that makes an unfamiliar tool feel approachable. Two posts carry the whole argument. One walks a small operator through why a CRM is really just one place to keep track of people who might become clients — who they are, where the conversation is up to, what happens next — and why running more than one enquiry at a time is the real trigger for needing one. The other reassures a client who apologises before opening a messy CRM, reframes the mess as useful information about where a business has outgrown its process, and keeps the tone on the client's side throughout. For the shared brief — a B2B SaaS company building a sustained brand presence that drives pipeline — this is the better pick only if the buyer definition is genuinely small-business and the desired output is teaching content, not a brand program.
- Sahil Aggarwal: Choose this alternative when the brand moment is an event and the strategy is hospitality plus product transparency. One post describes renting a space a few blocks from Dreamforce and turning it into a three-day offsite with Counter Culture coffee, pastries, puppies, lunch, a vinyl DJ and drinks, a CRO roundtable run with Fullcast, a RevOps happy hour, and open time for one-to-one meetings — an explicit answer to the conference floor being too much. The other explains making a product's underlying knowledge graph visible so revenue-operations teams can see what the AI actually knows and where it came from, with customer graphs and custom data applications planned over the following three months. For the shared brief, this is the better choice when the budget should go into a place buyers actually want to sit and into a product story about trust in AI, rather than into a publishing rhythm.
- Othmane Khadri: Choose this alternative when the brief wants an operator who builds and ships in public. One post describes wiring a Yalc.ai agent through Claude Code to pull every deal marked closed-lost in Attio over twelve months, read the lost reason and the last three activities, cluster objections against ICP attributes and the stage each deal died at, score open deals sitting on a losing pattern, and draft a personalised re-engagement ask to the lost prospects — a design a brand could copy. The second post describes releasing an open-source agent that replies to AI-generated comment spam with fake buying interest and an endless loop of questions, and it reports ten of eleven recipients answered. For the shared brief, this is the better pick when shareable, self-built tools and a sharp point of view matter more than a documented brand-measurement method.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/