Why AI Keeps Recommending PiggyVest for Savings in Nigeria

I had asked ChatGPT, Gemini and Perplexity five different questions about saving and investing in Nigeria. The competitors kept changing. Kuda became stronger around safety. Grey and Raenest suddenly appeared when I mentioned dollars. Risevest and Bamboo moved in and out depending on the question.

PiggyVest just kept showing up.

Fourteen times out of fifteen.

And almost every time, the AI tools told me some version of the same story: PiggyVest helps Nigerians save consistently and resist touching their money.

I wanted to know whether PiggyVest had engineered this. So I asked Ifeoluwa Adekoya, who works on SEO at PiggyVest.

His answer started in 2016.

The story started with Piggybank.ng

“Since the Piggybank.ng days in 2016, the goal has been for PiggyVest to be both the platform and the resource for money and money management in Nigeria,” Ifeoluwa told me.

That brief, he said, runs through the company’s wider marketing and individual pieces of content.

There is something almost boringly consistent about the language PiggyVest has built around its products.

AutoSave tells you the behaviour being automated. SafeLock tells you what happens to the money. Target Savings connects saving to a specific goal.

Those names also happen to be the words the AI tools repeated during my experiment.

ChatGPT described PiggyVest as best for “automated, disciplined saving.” Gemini associated the company with AutoSave, SafeLock and Target Savings when I asked which Nigerian apps help people develop financial discipline. Perplexity also selected PiggyVest as its best overall option for disciplined saving.

Ifeoluwa sees that repetition as the accumulated effect of a story PiggyVest has been telling for years.

“When an LLM ties us to savings discipline, it's reading back a decade of consistency. Our product names describe the exact behaviour, and our content has told the same story in the same words for years.”

There is emerging research suggesting that consistency across related questions deserves more attention than performance on an isolated prompt.

In July 2026, Semrush and Kevin Indig studied more than 50,000 brands across 1,094 subject areas in ChatGPT. Each category contained five related buyer questions. Only 15.2% of those categories had what the researchers defined as a clear owner: a brand appearing in at least four of the five prompts while maintaining a meaningful lead over competitors. More than half had no brand appearing consistently enough to qualify even as an emerging leader.

The study covered US categories and uses its own definition of ownership, so I wouldn’t transplant the benchmark directly onto Nigerian fintech. Its underlying measurement choice is useful here, though. The researchers looked across a group of related buyer questions because a company that appears for one prompt can disappear when the wording or intent changes.

That is what happened in my Nigerian fintech experiment.

PiggyVest’s 14 appearances covered general saving, safety, trust, saving discipline and dollar saving. Its only absence came when ChatGPT interpreted dollar saving around USD accounts and foreign income, leading it towards Grey and Raenest.

The story attached to PiggyVest remained unusually stable everywhere else.

Ifeoluwa says GEO has now given that old brand ambition a new destination.

“The aim now is to be recommended, cited too, ideally as the everyday Nigerian's way to build discipline and grow money.”

GEO changed the finish line

I expected Ifeoluwa to tell me about a completely new playbook PiggyVest had developed for AI search.

His answer was much closer to the work SEO teams already know.

They still build a strategy, create content, optimise the website and watch performance. PiggyVest still cares about search intent and whether an individual article can rank.

“The spine of the work is the same as ever,” he said. But the measurement around that work has expanded.

PiggyVest used to concentrate heavily on Google outcomes such as rankings, featured snippets, traffic and clicks. The team now also monitors which prompts trigger a PiggyVest recommendation, how the company gets described, where it gets cited and how those surfaces connect with sessions, direct traffic and conversions. The content brief has expanded alongside it.

“An individual article still has to rank and satisfy search intent, but every piece now also plays a part in larger, more intentional narrative engineering.”

“Narrative engineering” initially sounded grander than what I had seen in my test. Then I went back through the responses.

The models rarely returned a naked list of company names. They gave each company a small story.

PiggyVest helped people build discipline.

Cowrywise was associated with structured saving and investing.

Kuda belonged closer to everyday banking and saving.

Risevest and Bamboo became much more relevant when the conversation moved towards dollar-denominated assets.

The description beside the recommendation matters because that is part of what a buyer receives.

Semrush has found a similar distinction between being present in the source material and having your brand explicitly carried into the answer. In a 2026 study of thousands of appearances across AI search engines, 61.7% were what it called “ghost citations”: the AI cited a company's domain without naming the brand in the response. Only 13.2% of the appearances included both a citation and explicit brand mention.

So PiggyVest has at least two things to watch: whether its information gets used and whether the company itself survives the compression into the final answer.

That helps explain why Ifeoluwa describes recommendation and citation as parts of the new finish line.

Then I showed him the Renmoney finding

One part of my original experiment bothered me more than the rankings.

I had started opening the sources.

When Gemini answered my question about fintech apps that help Nigerians automate saving and develop financial discipline, Renmoney’s Best Savings Apps in Nigeria article kept appearing. Gemini used it while discussing PiggyVest. It appeared around Kuda.

It also supplied information around FairMoney.

Renmoney had published a comparison of its market, and that page was now part of an AI answer explaining several companies in that market.

I asked Ifeoluwa how much PiggyVest thinks about pages it does not own.

“We obsess over third-party sources,” he said.

That obsession existed during classic SEO too, although links carried much of the attention. He thinks AI search has made the surrounding mention and its context considerably more valuable.

“A tweet, a Google Play review, a news feature, a random blog post, even Instagram comments can all feed the picture these models hold of PiggyVest.”

The Gemini example gave him an easy illustration.

“One well-placed third-party page can end up describing a whole category.”

This is where GEO becomes much harder to contain inside an SEO team. PiggyVest can edit its homepage tomorrow. It can rewrite a help article or publish a new guide.

However, it cannot directly rewrite what a journalist says about the company or control the wording of every review, comparison page, Reddit conversation or competitor article that discusses the market.

Ifeoluwa describes all of that as the brand’s public “surface area.”

And his view of SEO has widened with it.

“SEO, which people used to treat as the ‘chill and easy’ work, now runs through product marketing and comms as much as it runs through content. What reviewers, journalists, and customers say in public is part of the surface area we work on.”

That feels like one of the more consequential changes here.

For years, marketers could divide these jobs fairly neatly on an org chart. SEO worked on rankings. PR worked on media. Product marketing worked on positioning. Community teams worried about public conversation.

AI recommendations can pull pieces from all of those environments into one answer. The person trying to influence that answer suddenly has reasons to care about all of them.

PiggyVest tracks the story AI tells about it

Ifeoluwa also treats LLMs as a strange new source of brand research.

Marketers have always tried to understand what people associate with a company. Search queries, surveys, interviews, social listening and brand studies each give you part of that picture.

Now there is another imperfect observer sitting in the middle of the internet.

“Instead of inferring how the public sees the brand from scattered signals, you can now ask the model directly, and what it tells you is close to what it tells users.”

He is careful about that interpretation. PiggyVest treats the answers as a baseline and runs enough varieties of questions across relevant tools to make the exercise useful.

Their measurement works in layers.

Piggyvest's traditional SEO dashboard still covers clicks, impressions, traffic and conversions. For AI, it checks visibility data through Semrush. The team then maintains rotating sets of TOFU, MOFU and BOFU prompts and records how different models describe the brand. Citations and mentions get monitored through Semrush and Google Alerts as well.

The company is therefore interested in more than a yes/no question about whether PiggyVest appeared.

If someone asks about saving for rent, what story comes back?

What happens around financial discipline?

Which companies appear beside PiggyVest when the buyer talks about investment?

Which sources accompany the answer?

Does the description still match what PiggyVest wants to represent?

That sounds considerably more laborious than checking ten blue links on Google.

“Analytics has become a lot of work,” Ifeoluwa said. “Important work, but a lot of it.”

Then he mentioned one of the models they monitor.

Meta AI.

Why Meta AI made the monitoring list

Ifeoluwa’s explanation was very Nigerian:

“Meta AI matters a lot here, because so much of Nigeria lives on WhatsApp.”

Meta expanded Meta AI to Nigeria in April 2024 and made the assistant available inside WhatsApp, Facebook, Instagram and Messenger. Users can ask questions or get recommendations without leaving those apps.

For a consumer financial company serving Nigerians, that makes the choice to monitor Meta AI easier to understand.

It also points to something marketers should consider when copying GEO playbooks from abroad. The most important AI surface for a company depends partly on where its customers already spend their time.

A US software company and a Nigerian consumer fintech do not automatically need the same monitoring stack.

PiggyVest’s approach starts with the audience and then decides which models deserve attention.

Measuring what happens afterwards is messy

There is another difficulty hiding underneath all this monitoring.

AI can influence someone without sending a neat referral visit.

A person could ask an assistant for recommendations, see PiggyVest, close the conversation and search the company on Google five minutes later. They might type the URL directly. They could open Instagram. Conventional attribution can easily credit the final step while losing the recommendation that introduced or reinforced the brand.

A June 2026 study tried to examine this behaviour by combining opt-in browsing data with conversations from ChatGPT, Claude and Gemini. Among users with no recent observed engagement with a recommended brand, an AI recommendation was associated with a 4.3 percentage-point increase in subsequent same-name Google searches and a 2.4-point increase in visits to the brand's own website. The study is observational and does not track purchases, so those figures should not be read as proof that an AI recommendation causes a sale.

That gives more context to the range of metrics PiggyVest is watching. Referral traffic can only tell one part of the story.

“The ultimate word-of-mouth merchants”

I ended by asking Ifeoluwa what he thinks marketers misunderstand about GEO.

“LLMs are sentiment aggregators and the ultimate word-of-mouth merchants, because they compress everything the internet says, or at least what they believe the internet says, about your brand and repeat it, with full confidence, to people who take them at their word.”

I like the metaphor, with one caveat.

There is no single internet consensus that every AI model faithfully reproduces. The tools retrieve and construct answers differently, and the results can move between prompts and between runs. Semrush’s research finds substantial differences in mention and citation behaviour between AI platforms.

The compression part still feels important.

A buyer used to have to do much of the synthesis themselves. Search for a category. Open several results. Read product pages. Check reviews. Look for comparisons. Ask someone they trust.

An AI assistant can now perform part of that synthesis inside a conversation and hand back a shortlist with descriptions attached.

My original experiment gave me a snapshot of the output. Speaking with Ifeoluwa helped explain some of the work sitting behind PiggyVest’s unusually consistent appearance within it.

He traces the story through years of product naming, content, search work and a repeated association with money management. PiggyVest now layers prompt monitoring, citation tracking, third-party mentions and more deliberate narrative work on top of that foundation.

There was no secret page template in his explanation. There was a decade of accumulated public evidence that the team is now learning to measure through another interface.

His final description of the work was simpler:

“You earn it, with strong strategy, public proof and analytics that measure the right things.”

For PiggyVest, GEO seems to have made all three easier to see.

Would AI Recommend Your Brand?

This Saturday, I’m hosting a free session with Treford on how ChatGPT, Gemini and other AI tools discover and recommend companies.

We’ll look at what makes brands like PiggyVest show up consistently, why competitors can appear for different buyer questions, and what this means for your own company.

Join the free session