Bobola Ojo-Ami, known to the audience as Bob, delivered one of the more practically focused sessions of Day 1. As Head of Global Marketing, Affiliates and Channel Partnerships at Ultrahost (the hosting sponsor powering the Web Agency Summit platform itself), Bob brought both strategic depth and a vendor's on-the-ground perspective to the central question facing the agency world right now: how do you continue to grow when AI has entered your market as a direct competitor?
His argument is structured around three core rules. First, agencies must stop selling deliverables and start selling outcomes. The work itself, code, copy, design, has been commoditised by AI. What has not been commoditised is the business judgment that determines what work matters and why. Second, agencies must reinvent how they acquire attention by transforming the entire organisation into a distributed thought network, where every employee becomes a public-facing thought leader, not just founders or senior staff. Third, and most critically for long-term revenue, agencies must identify and embed themselves into the "gray areas" of their clients' operations: the messy, unstructured, operational problems that AI cannot solve without deep contextual knowledge. Solving those problems and owning the infrastructure that supports the solution is how agencies build switching costs, retain clients, and grow recurring revenue.
The session closes with a brief Q&A covering Ultrahost's product stack for agencies and a candid breakdown of which LLMs are best suited to different use cases.
Key takeaways
- 01AI is not just a tool, it is a competitor. It has lowered the barrier to entry for producing agency-quality output and is driving the cost of delivery toward zero. Agencies that do not acknowledge this are not positioned to respond to it.
- 02Selling deliverables is a losing strategy. Clients can now access comparable deliverables from AI for a fraction of the price. The agency's value must shift entirely to the outcomes those deliverables are meant to produce.
- 03Speed and output quality are no longer premium services. They are the new baseline.
- 04Billing models tied to hours worked are structurally vulnerable.
- 05Reframe every service in terms of business impact. A logo is not a logo, it is market authority. A CRM implementation is not a CRM implementation, it is a shorter sales cycle. This reframing is not just messaging, it changes what the agency is competing on.
- 06Positioned attention is the new acquisition mode. In a market where anyone can demonstrate capability, the differentiator is being known and trusted. That requires intentional, scalable attention-building.
- 07Transform the entire organisation into a thought network. Every employee should be creating public-facing content from their own professional perspective. Ten employees with 1,000 followers each outperform one corporate account with 50,000 followers.
- 08Thought leadership is individual and authentic, not corporate. Employees reposting company content is not thought leadership. Each person should be sharing their own expertise, from their own context, in their own voice.
- 09Gray areas are the highest-value territory for agencies. These are the operational problems clients live with but rarely name. Solving them creates recurring revenue and deeply embeds the agency in the client's business.
- 10Own the infrastructure, not just the project. When an agency builds and maintains the tools, dashboards, and workflows that solve a client's gray area problems, it becomes very difficult to remove. This is how agencies create switching costs and true client ownership.
- 11AI is a tool for delivering on gray areas, not a competitor in them. AI cannot identify a client's operational gaps without niche context. It cannot embed a solution into a specific client environment. These require human judgment and domain expertise. Agencies that use AI to deliver gray area solutions are combining AI's execution speed with their own irreplaceable contextual knowledge.
The Core Problem: AI as a Direct Competitor
Bob did not ease into the topic. His opening framing was blunt: the competitive landscape has fundamentally changed, and AI is now a legitimate competitor to agencies, not just a tool agencies use.
The Lowered Barrier to Entry The central disruption is that the barrier to entry for producing agency-quality output has collapsed. Anyone with enough technical literacy, whether a product manager, a solo entrepreneur, or an internal team at a mid-sized business, can access Claude, ChatGPT, or similar tools and achieve outputs that would previously have required an agency engagement. The cost differential is stark: Bob estimated that a client might produce comparable output for less than 10% of the cost of hiring an agency.
This is not a future concern. It is the current reality. Companies are already equipping their internal teams with LLM subscriptions and instructing them to become more productive.
The example Bob gave was vivid: a UI/UX designer, who historically would have handed off to a development team, can now use AI to write the entire website themselves. The one-person agency is becoming increasingly viable, and this is eating into the traditional agency model.
Speed Is No Longer a Premium Service Agencies have long been able to charge a premium for fast turnaround. Bob's point is that this lever has been removed. Whereas a 48-hour turnaround was once a service 10
differentiator, AI can produce comparable output in minutes. Speed has moved from being a value-added service to being the baseline expectation. Billing models based on time and hours are structurally exposed to this shift because they cannot compete on the cost dimension.
Output Has Become a Commodity "Output is the new commodity." This phrase was one of the session's most quotable lines.
Great copy, clean code, and sharp design are no longer scarce. They are now accessible, cheap, and abundant. This does not mean agencies cannot produce them. It means that producing them is no longer sufficient justification for the fees agencies have historically charged. The technical deliverable, in isolation, no longer puts the agency in what Bob called "the strategy seat."
The Real Scaling Problem Many agencies, Bob observed, still rely predominantly on one-off project revenue supplemented by support and maintenance contracts or hosting subscriptions. That model is already fragile and AI has made it more so. Cheaper production tools increase volume capacity but compress margins. Without a structural change to the business model, profitability does not grow with volume. He was direct: "The model has to change or profits will not change."
Rule One: Sell Outcomes, Not Deliverables
This was the first of three rules Bob presented as the framework for scaling in an AI-dominated market.
The Fundamental Reframe A website is a deliverable. The purpose of that website is an outcome. Bob's instruction to agencies was simple: stop leading with the deliverable and start leading with the outcome the client actually wants. The mechanism by which the outcome is achieved is largely irrelevant to the client. How the work gets done does not belong in the pitch. The result does.
His example: a client comes to you wanting to launch an e-commerce store. The real goals are to launch cheaply and to sell more. The agency that wins is the one that positions itself around helping the client achieve those goals, not the one that sells a Shopify build. Can you also help them with customer acquisition? Can you help them close last-mile sales? That is where the value lies.
The Value Translation Layer Bob walked through several concrete examples of how to reframe deliverables as outcomes: Logo and brand kit reframed as: market authority and trust. The pitch is not "we design logos," it is "we help you build the trust that shortens your sales cycle."
CRM implementation reframed as: shorten the sales cycle. The outcome is a bottom-line impact on the client's P&L.
Email templates reframed as: automated retention systems. The deliverable becomes an ongoing mechanism, not a one-time asset.
Development work reframed as: revenue infrastructure.
Conversion-focused design reframed as: the conversion experience itself.
The pattern across all of these is consistent: move from describing the work to describing the business result of the work. Bob characterised this as moving the agency into the client's P&L conversation rather than the procurement conversation.
The Inversion of Effort Bob introduced what he called the "inversion of effort." In the old model, agencies were paid for labor-intensive work: lines of code, hours worked, pages designed. These were visible inputs that justified invoices. AI has rendered this model obsolete. Work that took a week can now be completed in hours with a $20 subscription.
In the new model, agencies should be paid for business transformation, revenue growth, and strategic outcomes. The question the agency answers is not "how do we build this?" but "why does this need to be built, and what business result should it produce?" AI can execute.
The agency provides judgment.
Language Shift Bob gave a pointed example of how this plays out in agency positioning. Saying "we build responsive React applications" or "we build beautiful websites on WordPress" is language that anyone can take directly to an AI and replicate. The value is gone from that statement.
The new language sounds like: "we build high-performance infrastructure that handles your peak traffic without dropping sales." The shift is from describing technical capability to describing business impact.
He used Ultrahost's own sponsorship of the Web Agency Summit as a live example of outcome-oriented positioning. Rather than simply sponsoring a banner or a slide, Ultrahost provided the entire infrastructure powering the event. Every attendee interacting with the summit platform was interacting with Ultrahost's product. That is an outcome-led play, not a logo placement.
Rule Two: Positioned Attention as the New Acquisition Mode
The second rule addressed how agencies get found and remembered in a market where capability alone is no longer a differentiator.
Why Attention Is Now a Bottleneck Bob's argument here starts with an observation: AI has lowered the barrier to entry for new agencies. More agencies can now be formed more quickly and more cheaply than ever before. This increases the noise in the market dramatically. At the same time, it makes capability a poor differentiator because almost anyone can demonstrate competence. In this environment, he argued, attention is the genuine bottleneck. The question is not whether you can do the work. It is whether the right clients are aware you exist.
The Problem with the Black Box Agency He described the traditional agency positioning model as a "black box." The agency presents polished deliverables and case studies but keeps its people hidden. Clients interact with a brand, not with individuals. Bob's contention is that this model cannot generate enough attention to compete in the current environment.
The Decentralised Thought Network Bob's solution is structural. The agency must transform itself from a company with one or two public-facing people, typically the founder or CEO, into what he called a "thought network." Every member of the organisation becomes a "thought agent," his term for a thought leader contributing publicly from their own professional perspective.
The logic is network-based. Each employee is a node. Each node has a follower base, a social presence, and a niche perspective. Together, those nodes generate a surface area for the agency that no single founder's LinkedIn profile can match.
He gave a specific example: a developer working on a project shares their experience of solving a particular technical problem, from their own perspective, in their own voice. A strategist shares their thinking on a client challenge. A project manager writes about how they manage delivery. None of this is about reposting the agency's corporate content. It is original, first-person, professional thought leadership that happens to be associated with the agency's brand.
The Surface Area Argument Bob made the math explicit: ten employees with 1,000 followers each generate more inbound leads than one corporate account with 50,000 followers. The reason is trust. People follow people, not logos. The humanised voice of an individual team member creates a level of trust that a faceless agency brand cannot replicate.
He made an interesting parallel to consumer brands: Red Bull does not expose its CEO to the world, but the product is deeply humanised through its positioning and narrative. For service businesses, the humanisation must come through people. The people in the agency are the vehicle for that humanisation.
Employee Incentive Alignment Bob noted that this approach also serves employees directly. Building a public professional profile makes individuals more valuable career-wise. They are not just doing work for the agency. They are building their own reputation and expertise in public. This is a retention-adjacent benefit: people value organisations that help them grow professionally. A culture of thought leadership creates that growth in a visible, trackable way.
A Critical Clarification Bob was emphatic that this is not about asking employees to share the company's posts or to promote the agency's services. That is not thought leadership. The expectation is genuine, individual, public-facing expertise content. Each person speaks from their own context, their own work, their own craft. The brand association follows naturally.
The outcome: the agency becomes a network of trusted voices rather than a single entity.
When a potential client encounters one of those voices on LinkedIn or Discord or YouTube, they connect with a person before they connect with a brand. That person becomes "the gateway into the organisation."
Rule Three: Win in the Gray Areas
This was Bob's third rule, and arguably the most commercially important for agencies looking to build durable, recurring revenue.
What AI Is Good At Bob acknowledged AI's genuine capability clearly and without hedging. AI is excellent at clearly defined tasks. Give it a well-scoped objective, from writing copy to building websites to sending outreach emails to nurturing leads through a pipeline, and it will execute consistently and at scale. Agentic AI amplifies this further: give it a goal, and it will reason through how to achieve it.
This matters because it establishes what territory agencies should not fight for. Competing with AI on clearly defined, well-scoped deliverables is a losing position.
Defining the Gray Areas Gray areas are the operational realities of a client's business that exist outside the scope of any formal project brief. They are the messy, ongoing, unstructured problems that clients
live with, often without flagging them as solvable. They are not what clients ask for. They are what clients quietly struggle with.
Bob's example was an e-commerce brand managing refunds. The client hires the agency to build a Shopify store. The store gets built. But the client is losing money through a poorly managed refund and returns process. Nobody is routing complaints consistently. Social media messages, emails, and review platform feedback are all handled ad hoc. This is a gray area.
The agency that identifies this problem and builds a solution, even something as simple as a Slack bot that automates the claims routing process, has created new recurring value. The client pays $50, $200, or $500 a month for that workflow. The agency has embedded itself in the client's operations in a way that is difficult to unplug.
Why Gray Areas Matter Strategically Bob's argument for the strategic importance of gray areas is multi-layered: Recurring revenue. Solving a gray area almost always produces a subscription or retainer rather than a one-off fee. The client is not paying for a deliverable. They are paying for an ongoing operational function.
Ownership of the customer. When an agency embeds itself in the operational infrastructure of a client's business, the switching cost rises dramatically. Removing the agency means dismantling a system the client relies on. This is the opposite of the project-delivery model, where the agency exits as soon as the site is live.
AI as an implementation tool, not a competitor. In gray areas, the agency's niche knowledge and contextual understanding are irreplaceable. AI cannot identify that a specific e-commerce client has a refund problem, understand the business context of that problem, design a solution for that specific environment, and embed it into the client's operational flow. The agency can. AI becomes a tool the agency uses to deliver the solution, not a competitor threatening to replace the agency.
Examples of Gray Area Interventions Bob walked through several concrete examples: Post-purchase logistics. A Shopify store is live, but the post-purchase refund and returns workflow is leaking money. The agency fixes the workflow and owns the dashboard that monitors it.
Customer support automation. Complaints are arriving across social media, email, and review platforms with no consistent process. The agency builds an automated routing and response system and charges a monthly fee to maintain it.
Internal knowledge base. A client asks for SEO and AEO content. The agency realises the client's internal knowledge base is too thin to support effective AI-assisted content
generation. The agency builds and maintains the knowledge base, embedding itself in the content infrastructure.
Custom reporting dashboards. A client needs reporting beyond what standard tools like Power BI provide. The agency builds a bespoke dashboard surfacing the KPIs that actually matter for that business, then owns the tool.
In each case, the pattern is the same. The agency starts with a visible deliverable request, looks beyond it to identify an operational gap, proposes a solution to that gap, and embeds the solution in infrastructure the agency controls.
Transitioning from Asset Builder to Operator Bob framed this as a shift in agency identity: from being an "asset builder" to being an "operator." The asset builder delivers a thing and leaves. The operator runs a function inside the client's business. That is a fundamentally different relationship and a fundamentally different commercial model.
He described this as making the agency "unpluggable." Once you have solved the refund workflow problem, built the knowledge base, and architected the customer support routing, removing the agency means removing those operational functions. Most clients will not do that.
The AI Market Scaling System: Bringing It Together
Bob synthesised the three rules into what he called the "AI Market Scaling System" and described how they work together in practice.
The Scaling Loop The loop works as follows: Thought agents (nodes) create public content from their individual perspectives, drawing attention to the agency's expertise in specific areas, including the gray area problems they have solved.
Positioned attention generates inbound interest from the segment of the market that will always prefer a trusted expert over a self-service AI tool.
Outcome-led pitching converts that interest by focusing on business transformation, not technical deliverables.
Gray area identification deepens the client relationship and expands recurring revenue.
Platform ownership embeds the agency in the client's operational environment, creating retention and referral loops.
The Compounding Effect Bob's key point about combining all three rules was that their impact compounds. Selling outcomes positions the agency for higher-value engagements. Positioned attention ensures a steady pipeline of those engagements. Winning in the gray areas turns individual engagements into long-term, growing relationships. Each rule reinforces the others.
He was unambiguous about what this means for the future: the agencies that scale will not be the ones building the most. They will be the ones owning the environment where results happen.
Ultrahost Partner Programme
Toward the end of the session, Bob introduced a specific offering for Web Agency Summit attendees: a dedicated web agency partner programme run by Ultrahost in partnership with Atarim.
He described the programme as designed to give agencies access to consistent recurring revenue, greater control over their clients' environments, access to the Ultrahost ecosystem, dedicated partner support, tools to help maximise profitability, and additional benefits. The sign-up URL mentioned during the session was ultrahost.com/atarim-partner-program (note: this URL was dictated verbally and the spelling was inconsistent in the transcript; see Editor's Notes).
Bob framed the programme as a direct expression of the gray area principle: own the infrastructure and you own the client relationship. Agencies that resell or manage hosting through Ultrahost are embedding themselves in a layer of their clients' operations that is difficult to remove.
Q&A: Recommended Stack for a Two-Person Agency
Andrew Palmer asked Bob what he would recommend as a stack for a freelancer or small agency looking to grow.
Bob's answer centred on Ultrahost's product range: Cloud hosting: dedicated servers, VPS, and VDS for agencies needing full control and flexibility.
Shared and WordPress hosting: for clients with simpler requirements.
AI-focused infrastructure: Bob mentioned "open claw host" plans (almost certainly referring to LLM-optimised hosting configurations, though the terminology in the transcript is unclear), as well as agent-specific plans and specialised gaming hosting.
His emphasis was on Ultrahost's flexibility: the stack can be configured contextually for different client needs within a single environment, which gives agencies the control they need to manage clients at scale.
Q&A: Which LLM Do You Recommend?
A question from attendee Ivana prompted a nuanced answer from Bob. He declined to give a simple "use this one" recommendation and instead broke the decision down by use case: Gemini delivers the best value when the user is already embedded in Google Workspace.
Connecting Gemini to Google Slides, Sheets, Gmail, and other Workspace tools turns it into an operational AI system with deep context across all the user's work. He argued that most people dramatically underutilise Gemini's potential in this integration mode.
Claude is, in his view, particularly strong for technical work and strategic reasoning. He recommended it as a "strategy and technical LLM toolkit" for agencies doing significant development work.
ChatGPT competes with Claude across many dimensions, and he acknowledged that Codex in particular is competing directly with Claude for development tasks.
His overall recommendation: use all three, but use each deliberately. Decide what role each tool plays in your stack and rotate through them accordingly. Do not pick one and ignore the others.
Decide what role each tool plays in your stack and rotate through them accordingly.
About the speaker
Bobola Ojo-Ami
Head of Global Marketing, Affiliates and Channel Partnerships
Bobola Ojo-Ami leads Global Marketing, Affiliates and Channel Partnerships, working closely with agencies on positioning and growth strategy in an increasingly AI-saturated market.