The Industry Kills Innovation, So Stop Selling It Tools and Just Do the Work

Ali Safavi opened the conversation with a dream rather than a product pitch: insurance should be holistic risk management. Understand someone’s risk, then manage it for them so they do not have to worry about it.

That sounds idealistic until he grounds it. His argument runs through Henry Ford’s faster horses — ask people what they want and they describe a better version of what already exists. What people actually want is safety and peace of mind, and for a few hundred years the mechanism offered to them has been risk pooling. The question worth asking is whether there is now a better way to deliver the same underlying thing.

Insurance already has the harder half of the problem solved. It is one of the most data-driven industries in the world and genuinely expert at understanding risk. What it has mostly done with that expertise, though, is price risk rather than manage it.

In Episode 135 of InsurTechTalk, Ali and I covered what holistic risk management looks like in practice, why COVU chose to be an AI-native services business rather than a software vendor, how acquisitions fit a company that insists it is not a rollup, and what he has learned about hiring for mission rather than money.

About Ali Safavi

Ali Safavi is the co-founder and CEO of COVU, an AI-native services company that manages customers on behalf of insurance agencies — servicing their books under the agency’s own name, method, and relationships. COVU launched in August 2022 and now employs more than 100 people, split across Los Angeles and San Francisco. It holds an agency license but is not structured as a conventional agency. The company has announced roughly $35 million in equity funding, including an oversubscribed Series A extension closed in Q1, plus a separate debt facility supporting acquisitions. Before COVU, Ali was at Plug and Play, where he first encountered the insurance industry.

What Holistic Risk Management Actually Means

Ali uses cyber insurance as his model, because the discipline is already visible there:

• Understand the risk
• Explain it to the customer through risk reports
• Help them finance it — self-insurance, risk pooling, or transfer
• Mitigate it, through actual security work that reduces the exposure
• When a claim happens, minimize it by helping the customer recover

That full loop is standard practice in cyber and almost absent for individuals and small businesses, because doing it at that scale without technology is prohibitive.

His example is close to home: the LA wildfires. What could have been done to help someone understand their wildfire risk beforehand? To drive mitigation — clearing brush, cutting back trees, roof sprinklers? To influence those decisions rather than merely price the outcome? And once the fires came, to play a genuinely better role in making people whole?

Ali was explicit that he never set out to build a faster quoting engine. Valuable, but not what motivates him.

Why an AI-Native Services Company

The strategic decision underneath COVU is the one Ali returned to as his single biggest lesson, and it is a pointed critique of how InsurTech has historically approached the market.

• The insurance industry, in his words, sometimes kills innovation — the model of building a tool and convincing carriers and agencies to adopt it has been tried for many years with limited success
• His alternative: do the work yourself, using the tools, and hand over the outcome
• Rather than selling an agency software they must integrate, adopt, and train staff on, COVU absorbs all of that and delivers the result — more efficiency and better customer experience without the adoption burden
• The company is building vertical infrastructure for AI and automation in insurance, then sitting on top of it as the service layer

The pitch to an agency principal has three parts: give them their time back, give them a scalable foundation, and cost less than what they were spending — while maintaining the same customer satisfaction.

He is candid that the main obstacle is novelty rather than competition. Agencies rarely decline because they found a better alternative; they decline because they do not yet trust anyone else with their customers and need more proof points before they get comfortable.

Acquisitions Without Being a Rollup

COVU added acquisitions as a function in 2025, which invites an obvious comparison to the consolidators. Ali draws the distinction carefully.

• COVU is reactive, not proactive, in the market — it is not out hunting for agencies to buy
• The trigger is usually an existing partner: a principal who has spent 30 years building an agency and now wants an exit
• Most of the business remains helping agencies that intend to stay and grow
• On competitors in the consolidation game, his position is that they are partners rather than rivals: COVU works with some of the largest rollup players, providing the scalable foundation and absorbing much of the integration, consolidation, and technology implementation that drives their EBITDA
• If another buyer wants to make a fair offer on an agency, COVU stays out of it

His framing: COVU is not trying to be the best rollup company, it is trying to be the best service company.

Hiring: Fear-Based Fuel Only Takes You So Far

With the company past 100 people, Ali’s biggest personal focus is talent — and he was unusually reflective about motivation.

• Fear-based motivations, greed included, are among the strongest drivers we have, and he is not dismissive of them — but they are, in his phrase, a different type of fuel that only takes you so far
• He was making good money at Plug and Play; the question that mattered was what the money was actually for
• That introspection is why the mission had to be something serving the community, and why he now looks for “missionaries” rather than mercenaries

On recruiting patterns, he sees an asymmetry:

• Hiring from insurance has not been especially hard — entrepreneurial insurance people, particularly those who have tried to change a carrier or agency from within, immediately recognize COVU’s proposition, because instead of persuading an insurance company to change, COVU performs the change on their behalf
• The harder and more current challenge is hiring from outside insurance and teaching them the industry. He notes the gap between a strong new computer science graduate and someone with real production experience — the “workhorse mentality” is what he screens for
• Early on the company could not afford that ramp and had to hire people who already understood both insurance and technology. With more funding and headcount, teaching insurance to outside talent has become viable

On the AI Hype Cycle

Ali has seen this movie before and said so plainly.

• In 2016, half his conversations with insurers were about blockchain and how to build risk tools on it. Blockchain is a fine technology; many of the applications being chased then — UBI on-chain among them — were hype
• He sees the same pattern now: a great deal of attention going to applications that are not where the value is
• His sharper structural point: the industry and its investors may create many tools that never get adopted, and those are two very different things. The next five to ten years could look radically different, or not, depending entirely on how much change is actually permitted to happen

He also flagged a specific gap in how the sector learns:

• Roughly 98% of startups fail, and the accumulated lessons from those failures never get properly translated into shared knowledge
• Lessons from four or five years ago about how agencies, brokers, and carriers actually adopt technology — which tools matter and which do not — have largely been forgotten
• Many VCs’ understanding of insurance is necessarily superficial, which makes distinguishing real problems from ones that only look good on paper a genuine edge

Key Takeaways

• Insurance is expert at understanding risk but has mostly used that expertise to price it rather than to manage or reduce it
• The full holistic loop — understand, explain, finance, mitigate, recover — already exists in cyber and is largely absent for individuals and small businesses
• Selling tools into an industry that resists adoption has a poor track record; delivering the outcome as a service removes the integration and training burden entirely
• COVU’s acquisitions are reactive exit options for existing partners, not a rollup strategy, and it treats consolidators as customers rather than competitors
• Greed is real fuel but limited fuel; building a company and a following requires a motivation that survives the hard parts
• The insurance sector systematically fails to metabolize the lessons from its own failures, and forgets adoption lessons within a few years