The Assistive Agent Optimization strategy roundtable on March 17, 2026 brought together three voices who have been watching the same shift from different angles for years: Jason Barnard of Kalicube®, Beatrice Gamba of WordLift, and Laurence O’Toole of Authoritas.

What made this session especially valuable was not simply that it explored the future of search. It clarified that the next stage is already underway.

For Jason, the central point was simple: Assistive Agent Optimization is no longer a future concept. It arrived in January 2026. Agents are already participating in transactions, already influencing choices, and already compressing parts of the customer journey that used to require multiple human decisions.

That changes the role of digital marketing.

The old model was built around visibility in search. The next model is built around recommendation in AI Assistive Engines. The one now emerging is built around decision-making by agents.

Search still matters. Recommendations still matter. But when agents begin making choices on behalf of users, the brand that gets selected is the brand the machine understands best, trusts most, and can act on with confidence.

That was the real theme of the webinar.

Jason Barnard explained why brand clarity comes first

Jason opened the discussion by mapping the progression from SEO, to Answer Engine Optimization, to AI Assistive Engine Optimization, and now to Assistive Agent Optimization.

His point was not that each phase replaces the last. It was that each new phase contains the previous one. If you optimize for agents properly, you are also building the foundations for visibility in search, in answer engines, and in AI Assistive Engines.

That is where Jason’s brand-first approach becomes particularly powerful.

He focused on the Entity Home, the single URL that gives AI systems a clear source of truth about a brand. For Jason, that page is where a company explains who it is, what it does, who it serves, and why it is the most credible solution in its market.

But he also made a point many brands still miss: saying something clearly on your own website is not enough.

You need corroboration.

Jason returned to the Claim, Frame, Prove model. A brand makes a claim on its own platform, frames that claim strategically, then earns third-party proof around the web. Without that external reinforcement, AI systems hedge. They mention the brand, but cautiously. They lack confidence.

And confidence, as Jason argued, is now king.

That idea fits directly with The Kalicube Process™. Understandability comes first, then Credibility, then Deliverability. Not because it sounds neat, but because machines need that sequence. They cannot recommend what they do not understand, and they cannot deliver confidently what they do not trust.

Beatrice Gamba showed why technical structure is now a business issue

Beatrice’s presentation gave the technical layer a very practical business meaning.

Her core argument was that most organizations are not failing because they lack content. They are failing because AI systems are still forced to interpret too much by inference.

When a page lacks strong structure, clear entity definitions, and visible relationships, the model guesses. Sometimes it guesses correctly. Sometimes it does not.

That is where the three failure modes she highlighted become so useful: doubt, displacement, and absence.

• Doubt appears when the machine is unsure.
• Displacement happens when a competitor is clearer to the machine and gets surfaced instead.
• Absence is the harshest outcome, when a relevant brand is simply not mentioned.

Beatrice explained that structured data still matters, but markup alone is not enough. If JSON-LD sits invisibly in the code while the page itself remains thin or ambiguous, modern retrieval systems may not use that information in the way teams expect.

Her answer was the entity hub.

In practice, that means making the Knowledge Graph visible on the page. Relationships between entities, descriptive attributes, internal links, and canonical references need to appear in ways both humans and AI systems can follow. The graph cannot just exist in theory. It has to become navigable, interpretable, and retrievable.

She also pushed the conversation toward what comes next: preparing data so agents can query it directly, not just crawl it indirectly. That is where MCP readiness, resolvable IDs, and structured endpoints start to matter.

The important takeaway was not that brands need a brand-new discipline.

It was that they need to do the fundamentals far better, and far more intentionally, than most are doing them today.

Laurence O’Toole explained why old measurement habits are no longer enough

Laurence brought the measurement lens, and it was probably the most useful reality check of the session.

Traditional SEO metrics were built for a more stable environment. Rankings were more deterministic. Search volume was directionally useful. Interfaces were less fluid.

That is not the world we are in now.

As Laurence explained, Large Language Models are volatile. Results vary. Interfaces are becoming generative. Users are asking longer, more contextual questions. And increasingly, brands are being judged not through a list of links, but through synthesized responses.

That means the measurement mindset has to change.

Rather than treating AI visibility like a classic rank-tracking problem, Laurence argued for a market research framework. Build prompt families. Understand audience context deeply. Look at use cases, buying situations, languages, locations, and question types. Then measure consistency of appearance over time, not just one-off visibility snapshots.

That distinction matters.

The issue is no longer just whether your brand appears once. It is whether it appears repeatedly, credibly, and in the right commercial contexts. Laurence described this through patterns of consistency, from default choice and leadership positions down to intermittent and sporadic appearances.

It was a useful reminder that AI visibility is not a simple leaderboard. It is a competitive presence pattern.

He also made one of the sharpest points of the webinar: your next customer could be an agent.

That idea shifts everything. It means brands are not only optimizing for discovery by humans. They are increasingly optimizing for evaluation, recommendation, and transaction by machines.

The webinar brought together three complementary pieces of the same strategy

What made this roundtable work so well was that the three presentations were not competing perspectives.

They were three legs of the same stool.

• Jason covered brand clarity and the strategic narrative.
• Beatrice covered the technical and semantic structure that makes that narrative machine-readable and machine-usable.
• Laurence covered the measurement framework needed to know whether any of it is actually working.

Taken together, the message was difficult to ignore.

This shift is already happening. Brands do not need to wait for full agent adoption to justify acting. The work already creates value now: clearer positioning, stronger corroboration, better machine understanding, and more reliable visibility across AI-mediated journeys.

And for teams that delay, the risks are already visible: doubt, displacement, and absence.

This webinar also marked a strategic transition for Kalicube and its partners

There was another layer to the event that matters.

This was not just an educational webinar. It also marked a positioning moment.

Jason has been explicit that his Search Engine Land articles, patent work, academic paper, this webinar, and the Kalicube Summit 2026 all fit into a broader “claim, frame, proof” model designed to establish Kalicube at the center of the AI-era marketing conversation.

This roundtable played the transition role in that strategy.

It publicly connected the present reality of Assistive Engine Optimization with the next step, Assistive Agent Optimization. It positioned Kalicube®, WordLift, and Authoritas not as commentators watching the shift happen, but as companies actively defining how the industry should respond to it.

That matters because the market is still early.

The language is still forming. The frameworks are still being tested. The benchmarks are still being set.

And in moments like that, the brands that explain the shift clearly, repeatedly, and with strong third-party corroboration are often the brands AI systems learn to associate with the category itself.

Three convictions from the Assistive Agent Optimization webinar

By the end of the discussion, three convictions felt particularly clear.

1. First, 2026 is the year to build the baseline. Jason closed on an important point: brands need to benchmark now, while there is still enough visibility to understand what AI systems currently think, say, and cite. Waiting until the environment becomes more opaque will make that much harder.
2. Second, AI readiness is no longer only a content problem. It is a brand clarity problem, a Knowledge Graph problem, a corroboration problem, and a measurement problem all at once.
3. Third, the brands that start now will not simply be better prepared for agent-led discovery later. They will shape the data, structure, and confidence patterns that machines learn first.

That is the deeper opportunity.

The shift to Assistive Agent Optimization is not only about keeping up. It is about teaching the machines early enough that, when the rest of the industry catches up, they are already repeating your narrative with confidence.