Eric Williamson Is Teaching Enterprise AI to Start Small

Eric Williamson likes a small victory. This is mildly unfashionable in enterprise artificial intelligence, where the preferred unit of promise is transformation and the usual timescale is yesterday. Yet when Williamson talks about putting AI into customer service, his advice begins with the sort of assignment that rarely earns a keynote: find a repetitive task with low compliance risk. Automate it. See whether life improves. Then take another step.

“Little wins. Start small,” he told an interviewer in 2025. The phrase is compact enough for a sticky note and sturdy enough for a strategy. It also reveals the marketer behind it. Williamson has spent more than two decades moving between brands, agencies and software companies, learning how grand language behaves when it meets an actual customer. The customer usually wins.

Today he is Chief Marketing Officer at CallMiner, the Waltham, Massachusetts company that analyzes voice and text conversations. He oversees the global marketing operation: brand, demand generation, events, product marketing and the business development team. It is a broad jurisdiction, spanning the promise a company makes, the people it hopes to reach and the first human exchange that follows. In Williamson's telling, those are parts of one system.

The education of a translator

Before conversation intelligence, there were pizzas, insurance commercials, hotel rooms and cruise ships. On the agency side, Williamson rose through digital strategy at The Martin Agency and digital and social leadership at MullenLowe. The client roll call crossed Google, Microsoft and Intel, then GEICO, Walmart, P&G, Pizza Hut, Acura, Royal Caribbean and Hyatt. It is hard to imagine a single audience common to all of them, apart from the rather important fact that each contains people.

Agency work trained him to enter an unfamiliar business, locate its pulse and explain it in language someone else might care about. It also exposed him to the discipline of a brief. A brand may want to discuss itself indefinitely. A customer has other plans. Digital strategy, especially in the years when every brand was discovering social media at once, required choosing which new behavior mattered and which novelty merely glittered.

Williamson had seen the revenue machinery from closer range, too. He has described his first job as essentially a business development representative role. That early vantage point stayed with him. Marketing can generate an immaculate campaign, but someone eventually has to pick up the conversation. At CallMiner, the BDR team sits within his remit, while marketing works closely with sales and channel partners. Alignment, in this arrangement, is not a quarterly ceremony. It is an organizational fact.

He later crossed from agency to vendor. At Acquia, the digital-experience company built around Drupal, he led brand development, creative services, web operations, editorial and demand generation. In July 2020, CallMiner appointed him CMO. The move placed him inside a business whose raw material is unusually revealing: the things customers say when they need help, want to buy, plan to leave or cannot understand the bill.

“Find those low compliance risk, mundane tasks that can be automated and start there. Little wins. Start small.”Eric Williamson on adopting AI in the contact center

The customer is already talking

Companies have long asked customers what they think. The survey is polite, structured and wonderfully easy to turn into a chart. It is also a narrow aperture. A support call or chat contains something less groomed: urgency, hesitation, repeated questions, the moment an explanation fails and the improvised workaround that follows. Those conversations are feedback that nobody had to request.

CallMiner's proposition is that artificial intelligence can examine those interactions across channels and at a scale human reviewers cannot. Williamson describes a view that can move from patterns across millions of exchanges down to a single conversation. Scale spots the recurring problem; detail shows what the problem feels like. Between them sits the possibility of action.

The organizational implication is more interesting than the software category. A problem voiced to a contact-center agent may belong to product, operations or marketing. Perhaps the feature is broken. Perhaps the policy is obscure. Perhaps the company promised the wrong thing. Williamson argues that the insight should rise beyond the service floor to a CMO, product leader or operations executive who can address the root cause.

That makes the contact center less like a complaint department and more like a listening post. The distinction matters because most companies do not suffer from a shortage of data. They suffer from data that stops at a departmental border. In CallMiner's 2025 customer-experience research, 96 percent of surveyed leaders called AI key to their strategy, yet 42 percent still relied on manual processes to analyze customer-experience data and 62 percent said they often failed to use that data to its best advantage.

Brand before wardrobe

In 2021, CallMiner refreshed its brand and language as its platform expanded beyond the older label of speech analytics. Customers were no longer speaking only by phone; the company was capturing and analyzing an omnichannel world. Williamson's explanation was refreshingly practical: what the platform could do had outrun the way the company described it. A rebrand was needed because the old story had become inaccurate.

His order of operations for such work begins far from color palettes. Start with the “North Star elements”: mission, values, the long elevator pitch and the version that fits in a tweet. Reduce those decisions to three or four messaging pillars. The visual system comes later. This saves a brand from dressing handsomely for the wrong occasion.

The philosophy also clarifies Williamson's complaint about B2B branding. Technical companies are fluent in machinery. They list integrations, features and architecture because those things are concrete and defensible. Buyers, however, still need a reason to care. A product story has to connect capability to consequence without turning every sentence into confetti. His years around consumer brands supply a useful corrective: clarity and memorability are not frivolous merely because the purchase requires a committee.

He applies a similar economy to content. Make fewer pieces of higher value, he has advised, then find multiple uses for the good work. The point is not austerity for its own sake. It is concentration. A useful idea can become a conversation, an article, a sales aid and a point of view. A forgettable idea merely becomes four forgettable assets.

A measured appetite for AI

Williamson is not skeptical of AI's reach. He has called generative AI transformative and speaks plainly about analyzing every aspect of millions of customer interactions. His caution concerns the route from possibility to deployment. After ChatGPT arrived, he watched organizations hurry features into the market without enough strategic consideration. By late 2023, he saw leaders paying more attention to the balance between agility and responsibility.

Customer-facing systems make that balance immediate. A model that invents an answer is not committing an abstract technical error; it is speaking to somebody on the company's behalf. Security, governance and business value therefore belong in the same conversation as speed. The useful question is not whether a company possesses an AI feature. It is whether the feature improves a real interaction, under conditions the company understands.

Here the small win returns. Summarize the call so an agent can spend less time typing. Retrieve relevant context while the customer is still present. Identify a routine step that machines can handle and people need not. Each modest application produces evidence, and evidence earns trust. The grand strategy becomes a staircase rather than a leap over an inconvenient canyon.

The future of customer experience is not hidden in a slogan. It is scattered across millions of ordinary conversations, waiting for a company patient enough to listen.

This is where Williamson's varied career resolves into a consistent occupation. At the agencies, he translated brands for consumers. At Acquia, he translated a technical platform for enterprise buyers. At CallMiner, he markets a system built to translate customers for the companies serving them. The categories change; the work keeps asking for attention, judgment and a sentence clear enough to travel.

In an industry fond of claiming that everything has changed, his method preserves a few stubborn continuities. Begin with the mission. Know the audience. Keep sales close. Make valuable things. Listen to what people actually say. Use new tools where they improve the work, and prove it before enlarging the promise. None of this produces the theatrical rush of an AI moonshot. It produces something more useful: the next credible step.