CX and AI Implementation Experts Worth Following (2026)

The 10 at a glance

Ranks and scores come from our partner rankings, which score each firm out of 10 on public pages only. This post adds no scores of its own, and how often someone posts never changes their firm's score.

How we picked

We started from every person a B2B people database returned for the firms in our six platform rankings, then took the 33 of them whose LinkedIn activity we'd checked, plus the founders of two AI support vendors, 35 people in all. That check, read-only and run on October 29, 2026, counted each person's own posts over the previous 90 days (reshares of other people's posts don't count) and kept a handful of samples.

Then we applied two tests.

1. Four or more own posts in 90 days. 18 of the 35 passed. Someone who posts once a quarter isn't a feed you can learn from.
2. The samples are about doing the job. We read every sample and kept people whose posts deal with configuration, AI rollout, workflows, migrations, measurement or the daily mechanics of a support queue. We dropped people whose samples were mostly booth announcements, conference sponsorships, hiring, investments or their own product launches. That removed 8 of the 18, including the founders of two AI support vendors and both Intercom partner leaders we checked.

Fame counted for nothing. Two picks are borderline and we've said why in their entries: half of Kurt Pinoy's samples are team news, and Brandon Tidd's recent posts are notes from other people's conference talks, which we kept because they're about running AI in practice and because Brandon Tidd writes practical Zendesk material on 729 Solutions' blog.

Titles are checked on each firm's own site where the site names the person, and marked "per LinkedIn" where it doesn't. Nobody here asked to be included.

Gorgias: the ecommerce operators

Four of the top four firms in our Gorgias partner ranking have a leader who posts about the work. Gorgias partners mostly serve Shopify brands with a hard Q4, and their leaders write about it in October.

Jordan Brown, omnie

omnie ranks 1st on Gorgias with 10/10. Its about page lists Jordan Brown as "Founder & President". omnie runs its own AI on a brand's Gorgias, priced per ticket, alongside outsourced agents.

Jordan Brown has posted 16 times on LinkedIn in the last 90 days, mostly about what ecommerce customers expect over the holiday season. One sample walks through the stages a customer goes through after a Q4 order, from excitement to checking tracking twice a day to suspicion when a label sits at "created" for days, and then contacting the brand on every channel at once. Another points out that big retailers are already accepting returns on October purchases until January 31, and lists what a smaller brand should settle before customers ask for the same. LinkedIn profile.

What you learn: Jordan Brown describes where-is-my-order demand as a sequence you can get ahead of, which is the right frame for deciding what an AI agent answers in November and what it hands to a person.

Jonathan Jamil, Springboard

Springboard ranks 2nd on Gorgias with 9/10, mainly for a weekly review of what the AI says to customers. Jonathan Jamil is founder, per LinkedIn; Springboard's site doesn't name its leadership.

Jonathan Jamil is the most frequent poster here, with 24 posts on LinkedIn in the last 90 days, and most of the samples are jokes. The useful one is about the cost nobody calculates when a brand replaces its CX agency: the current team already knows the products, the customers, the subscription platform and the fulfillment setup, and a new team has to learn all of it. LinkedIn profile.

What you learn: Before you switch outsourced teams or AI vendors, price the knowledge you'd be throwing away, because the incoming vendor has no incentive to mention it.

Jess Cervellon, Open Late Collective

Open Late Collective ranks 3rd on Gorgias with 9/10. Its homepage introduces Jess Cervellon as the founder, and a client testimonial there says she served as their fractional VP of CX.

Jess Cervellon has posted 12 times on LinkedIn in the last 90 days, mostly about where AI belongs in customer experience, plus weekly CX job listings. Her clearest sample says AI shouldn't handle every part of the experience and names the jobs it should take: refunds, order status, ticket analysis, review trends and flagging customers who might churn. Another argues that when AI can work directly with marketing data, the gap between spotting an opportunity and acting on it shrinks. LinkedIn profile.

What you learn: She gives you a concrete first list of AI jobs for an ecommerce queue, and an equally clear line around the rest.

Robert Durrette, Ridgeline Agency

Ridgeline Agency ranks 4th on Gorgias with 9/10 and 2nd in our Front partner ranking with 6/10, where it has the most specific Front AI scope we found. He's President & Founder, per LinkedIn. Ridgeline's about page names him in the firm's story and says he spent years building and running customer service operations before he started it.

Robert Durrette has posted 4 times on LinkedIn in the last 90 days, all about how service teams measure and spend their time. One argues that AI won't replace service teams but will change how agents spend their day once it takes the repeatable, high-volume questions. Another says that counting avoided interactions measures what customers didn't do, while resolution measures what you fixed. A third says customer effort matters more to satisfaction than raw reply speed. LinkedIn profile.

What you learn: He's the one person here who covers both Gorgias and Front, and every post he wrote this quarter is about measurement you can apply to either.

HubSpot Service Hub

Pratik Thakker, INSIDEA

INSIDEA ranks 2nd in our HubSpot Service Hub partner ranking with 8/10, for fixed fees on both onboarding and AI work. Its about page lists Pratik Thakker as "Founder & CEO".

Pratik Thakker has posted 14 times on LinkedIn in the last 90 days, mostly about how business owners adopt AI and where follow-up work stalls. One sample treats AI adoption as a loop: once one task runs on AI and the impact is measured, the next task goes back to the start. Another is about fixes that get booked for a quieter month while the customers who wrote in this week wait. A third describes INSIDEA's own AI worker listing the cost of each deliverable it produced and cutting its own running costs. LinkedIn profile.

What you learn: He writes about AI rollout as ongoing operations with a cost per task, which is how you should be judging an agent three months after launch.

Zendesk

Five of the ten work at firms in our Zendesk partner ranking. The Zendesk feeds lean toward measurement and plumbing: integrations, email ownership, employee service and migrations.

Dave Hobbs, CX Experts

CX Experts ranks 3rd on Zendesk with 9/10. Dave Hobbs is Head of Customer Success and Optimisation, per LinkedIn; the CX Experts site doesn't name its leadership.

Dave Hobbs has posted 17 times on LinkedIn in the last 90 days, mostly about Zendesk projects and automation in South African customer service. The most practical sample puts a per-ticket time cost on agents toggling between the CRM and other tools, and argues the real integration work starts once that interface barrier is gone. Another opens on teams that measure customer wait times in days and describes a Zendesk rollout for a South African client. LinkedIn profile.

What you learn: Count the tabs your agents open per ticket. If an AI agent can't read the same systems, it will hit the same wall.

Jonathan Keane, CustomerHD

CustomerHD ranks 7th on Zendesk with 8/10. Its Our People page lists "Jonathan" as "CEO & Co-Founder"; his surname comes from LinkedIn.

Jonathan Keane has posted 18 times on LinkedIn in the last 90 days, mostly about Zendesk for employee service and why deflection isn't resolution. One sample asks whether customers are now getting faster answers than employees, since many organizations put years of AI and automation work into the customer side first. Another says most service teams are organized around handling requests, not completing them, which is why first-contact resolution slips. A third draws the deflection line plainly: a deflected ticket stayed out of the human queue, and a resolved one got fixed. LinkedIn profile.

What you learn: If you run IT or HR support on the same help desk, his feed is the one that treats employee service as its own queue with its own AI case. Our deflection rate explainer covers the measurement side.

Kurt Pinoy, Premium Plus

Premium Plus ranks 10th on Zendesk with 7/10 and works only on Zendesk. Its Meet our experts page lists Kurt Pinoy as "CEO".

Kurt Pinoy has posted 19 times on LinkedIn in the last 90 days, mostly about client projects, new hires and Zendesk events. Kurt Pinoy is a borderline pick because half the samples are team and family news. The one that earns the place is a migration story: Qobee, a procurement software company, left an in-house support tool it had run for 15 years, and a software release that used to mean 80 ticket replies now means one. LinkedIn profile.

What you learn: Kurt Pinoy posts client results as before-and-after numbers, which is the format to ask any partner for when they quote you a migration.

Lumi Fagbure, advoCX

advoCX ranks 20th on Zendesk with 6/10; its site doesn't describe an AI service yet. Lumi Fagbure is founder, per LinkedIn; advoCX's site doesn't name its leadership.

Lumi Fagbure has posted 5 times on LinkedIn in the last 90 days, mostly short pieces about how support teams handle customer email. One is a bingo card of inbox failures, starting with two people replying to the same customer while nobody answers the next one. The first of a "Friday Fwd" series describes teams whose only system for managing customer email is one person's memory and a color code. LinkedIn profile.

What you learn: Fix ownership before automation. An AI agent dropped into an inbox with no written routing rules repeats the double replies and the missed ones, only faster.

Brandon Tidd, 729 Solutions

729 Solutions ranks 28th on Zendesk with 4/10; it does integrations and custom development and names little AI work on its Zendesk page. 729's own blog calls Brandon Tidd "our Lead Zendesk Architect" in its Zendesk Relate 2024 recap.

Brandon Tidd has posted 11 times on LinkedIn in the last 90 days, mostly about putting AI to work, drawn from the MAICON 2026 conference. The samples pass on one practical idea from a speaker there, Rachel Woods: if you want AI you can rely on, decide what you expect from it the way you would for a new person on your team, and write that down as an AI playbook. LinkedIn profile. Brandon Tidd's longer Zendesk admin posts are on 729 Solutions' blog.

What you learn: Write the expectations for an AI agent before you switch it on, the same way you'd write a job description, so you have something to check its replies against.

Which person to follow for your situation

This list fits teams that run the help desk in-house and want practical reading from people who implement it for a living. If you're shortlisting a partner, use the rankings instead.

• You sell online and dread Q4. Jordan Brown for peak-season demand, Jess Cervellon for which jobs to give AI first.
• You're deciding how to measure an AI agent. Robert Durrette and Jonathan Keane on resolution versus deflection, Pratik Thakker on cost per task.
• Your queue is email-heavy and ownership is loose. Lumi Fagbure.
• You run employee or IT service on Zendesk. Jonathan Keane.
• You're planning a migration or replacing a vendor. Kurt Pinoy for the result to aim at, Jonathan Jamil for the switching cost.
• Your agents live in five tabs. Dave Hobbs on integrations.
• You're writing your first AI agent instructions. Brandon Tidd's playbook idea.

What we check before an AI agent goes live on a partner-built help desk

Macha builds AI agents that run inside the help desk a team already uses, and the Macha team sets them up and monitors them. Many of those help desks were configured by a partner like the ones above, so the first week is mostly reading what they built.

• Ticket fields. We read the field definitions and the values on real tickets. On Zendesk our agent can write dropdown, text and checkbox values, so we need to know which custom fields drive routing before it touches any.
• Triggers. We add a unique tag when our trigger fires and exclude that tag from the trigger's own conditions, so it can't loop. Partner-built triggers that rewrite tags can break that, so we check them first.
• Past tickets. The AI Knowledge Builder turns historical tickets into a knowledge base, and the categories in those tickets tell us which one to start with. Then we replay past tickets through the agent in simulation, as our post on testing agents on real data shows.
• Internal note or public reply. We confirm which one the agent writes before any autonomous mode. In safe mode, drafts land as internal notes and a person approves each one.

If your help desk is already set up and you want an agent working inside it, start a trial; the Macha team sets it up and monitors it.

FAQs

Who are the best CX and AI implementation people to follow on LinkedIn?

From partners in our rankings, the ten here: Jordan Brown, Jonathan Jamil, Jess Cervellon, Robert Durrette, Pratik Thakker, Dave Hobbs, Jonathan Keane, Kurt Pinoy, Lumi Fagbure and Brandon Tidd. Each posted four or more times in 90 days, mostly about doing the work.

Why is nobody from an Intercom or Freshdesk partner listed?

We checked the leaders of firms in our Intercom and Freshdesk rankings whose activity we had. Those with enough posts wrote mostly about trade-show stands, conference sponsorships, expansion and hiring, so they didn't pass the substance test. We'll re-check at the next review.

Does posting on LinkedIn affect a partner's rank?

No. Ranks come from public pages scored on each post's rubric. Posting counts appear here only to show who shares their thinking often enough to learn from.

Are any of these people Macha partners?

No. None of them or their firms has a commercial relationship with Macha, and nobody was asked to take part.

<div class="docs-callout"><div><p class="docs-callout-title">Are you listed here?</p><p class="docs-callout-body">Partners and app makers can send corrections with a public source to [email protected]. We fix facts within a week. Scores follow the rubric above, not requests.</p></div></div>

How we researched this

On October 29, 2026 we read our six live partner rankings for Zendesk, Gorgias, Intercom, HubSpot Service Hub, Front and Freshdesk and took every rank and score from them. Names and LinkedIn URLs came from MoltSets and QuickEnrich, two B2B people databases. Posting counts and the samples described here come from a read-only check of each person's LinkedIn activity over the 90 days to that date; we paraphrase the samples and quote no more than a short phrase.

We confirmed titles on the firms' own pages that day: omnie's about page, Open Late Collective's homepage, INSIDEA's about page, CustomerHD's Our People page, Premium Plus's Meet our experts page and 729 Solutions' Zendesk Relate 2024 recap. Springboard, CX Experts and advoCX don't name these people on their sites, and Ridgeline's about page names Robert Durrette without a title, so those titles are per LinkedIn. We didn't contact anyone or use emails, phone numbers, locations or photos.

Disclosure: Macha makes AI agents for help desks. Nobody paid to be listed, and Macha isn't ranked.